Quick note before we get into it. I wrote a piece this week for Osborne, the firm I spend most of my days with, called “Same Rodeo, Different Bronco.” It was strong enough that I wanted it in front of S2S readers too, so I’ve rebuilt it here for this audience. If you already read it on the Osborne side, skim and skip. If you didn’t, this one matters.

The short version: AI isn’t failing in the field. It’s failing at the point of purchase. Boards are buying the same vendor story they bought in 2003, 2010 and 2018, the cheque is clearing, and the front-line delivery never happens. The pattern is so loud you can set your watch by it.

Let’s break it down.

Signal:

THE BOTTLENECK ISN’T AI. IT’S HOW BOARDS ARE BUYING IT.

The work order said: digitize the safety forms onto a tablet.

The reason was harder. A field worker had been killed on the job. Someone with a name. Someone with a family waiting at home. Someone whose crew came back to work the next week because the work doesn’t stop, even when it should. A loss that nobody at that table was ever going to be able to put back.

The energy company was doing what regulated industries do after a death like that. Investigations. Audits. New procedures. A reckoning with the gap between the way the work was meant to be done and the way it actually got done on a hard day. The brief that landed on my desk came downstream of all of that, and downstream of a crew still trying to make sense of why.

On paper it was a software brief. Take the paper tailboard form. Put it on an iPad. Make sure crews complete it before they touch live equipment.

I was the lead from Apple on the account, partnered with a development studio I trust to push harder on the opportunity. We could have built what the brief asked for. A pixel-perfect copy of the binder, with a battery in it. That would have changed nothing.

The honest question wasn’t “how do we digitize this.” It was “if we sat the crew supervisor, the safety engineer, and the regulator at the same table and started from scratch, would any of us still build this form the way it exists today?”

The answer was no.

That’s when the work got interesting. We rebuilt the process from the field crew up. What got asked. When it got asked. Who saw the data afterward. What happened the next time a hazard showed up at the same kind of site. The iPad was almost incidental. The thinking was the project.

The first version shipped to a small group in 2017. They used it. They pushed it. They told us what was broken. We fixed it. Then they trusted it. Within five years they had mitigated 847,000 hazards through the app. Tailboards complete in half the time. 185,000 sheets of paper a year, gone. More important than any of those numbers, the field crews have a tool that respects what they actually do for a living, and the data trail that comes with it makes the next fatality less likely.

That’s not a moonshot. It’s what works when you refuse the brief.

And refusing the brief is the opposite of what most boards are about to do with AI.

Same rodeo. Same dust. Different bronco.

Ian Shepherd wrote a short piece a couple of months back called “Running in Fog” that I haven’t been able to shake. He laid out the pattern of how organizations bought their way into the early web era. Phase one, this isn’t real retailing. Phase two, the startups doing it are cheating. Phase three, oh god we’re behind, get a big consulting firm in here. Phase four, write the biggest cheque you can to leapfrog the field. Phase five, oops.

If you’ve sat in a boardroom in the last six months, you’ve seen at least three of those phases play out. Except now they’re stamped with the letters A and I.

The pattern isn’t new. We rode it with databases. We rode it with the web. We rode it with mobile. We rode it with RPA. Each cycle, the leadership team that has never built the thing buys a story from the only people in the room who claim to understand it. The cheque clears. The vendor disappears. The reorg starts. The narrative gets rewritten as a learning opportunity.

The economist Carlota Perez has a name for the stage we’re in. She calls it Frenzy. Speculative capital floods in. The technology gets oversold. The bubble inflates. Eventually a Turning Point lands, and the deployment of the technology becomes boring, useful, and profitable, for the companies that didn’t blow themselves up first.

We are squarely in AI Frenzy. The signal is loud.

MIT’s NANDA project published research in August 2025 showing that 95% of enterprise generative AI pilots have delivered zero measurable P&L impact. Not weak impact. Zero. Klarna has quietly rehired customer service staff after publicly declaring its AI replacement complete. Air Canada was held legally liable by a Canadian tribunal for a refund its chatbot invented out of thin air. McDonald’s killed a two-year IBM drive-thru voice AI test after the system ordered nine sweet teas instead of one. DPD’s chatbot wrote a poem mocking its own company before being yanked offline the same afternoon.

These aren’t fringe failures. They’re enterprises with money, talent, and brand to protect, running off the cliff in the fog because nobody slowed them down.

This isn’t an argument against AI. AI is going to reshape how mid-market companies operate. The argument is that the way most enterprises are buying AI today is the same way they bought their web platform in 2003, their iPad rollout in 2010, and their RPA program in 2018. I watched the RPA wave play out. The bots that looked great in pilot died in production inside two years, brittle to every upstream change. Most enterprise programs stalled before they scaled. The ones that survived rarely paid back what was promised. The investment cycle ended quietly, and the conversation moved on to AI.

We’ve already been thrown by this bronco twice. We know how the ride ends. The question isn’t whether AI is real. The question is whether your organization is going to ride it, or perform the ride until it bucks you off again. Will you last the eight seconds, or perform for the board on the way down?

Scale:

FIVE MOVES TO DELIVER AI INSTEAD OF PERFORM IT.

The MIT Sloan piece “The Eight Core Principles of Strategic Innovation” by Gina O’Connor and Christopher Meyer is the cleanest read I’ve found on what separates companies that build new growth from companies that buy decks about it. The whole article is worth your time. Here’s how five of the eight principles land on the AI delivery problem specifically.

1. Refuse the brief. The tailboard project worked because we didn’t build what was asked for. The energy company’s brief was a tablet version of their safety form. What they actually needed was a re-thought safety process where the tablet was almost incidental. Most AI strategies in 2026 are starting the wrong way around. Take the existing process. Put AI on top of it. Trust the vendor’s productivity claim. That’s the digitized-PDF version of AI. It’s expensive, it’s safe, and it changes nothing. The MIT principle here is what O’Connor and Meyer call setting domains of innovation intent. Pick the real domain where you have an unfair right to win. Don’t accept the version of the problem the vendor wrote on the slide.

2. Treat the work as a portfolio, not a pipeline. A pipeline asks each project to justify itself on a go/kill timeline. A portfolio asks a cluster of small experiments to teach you something about a domain. Tailboard taught us things we couldn’t have specified in a contract. The right unit of measure isn’t whether one project ships. It’s whether the portfolio is learning faster than the market is moving. If you’re running AI today as five disconnected vendor pilots with five different success criteria, you don’t have a portfolio. You have a procurement habit.

3. Build discovery, incubation, and rollout as three different jobs. Most companies hand a single team the work of finding the opportunity, proving it, and operating it at scale. That team usually fails at one of the three, because the skills are different. Discovery is field work. Incubation is experiment design. Rollout is operational integration. Confusing the three is why so many AI projects look great in the pilot and die in the operating unit. Naming the three jobs and staffing them honestly is the cheapest way to lift your delivery rate.

4. Make the function permanent, not a program. IBM’s Emerging Business Opportunities program added more than $15 billion in new revenue to the company before it got shut down. Why was it shut down? Because it was a program, not a function. Programs end. Functions endure. If your AI work lives inside a hub that reports to nobody and disappears in the next budget cycle, the work will disappear with it. The working setup for a mid-market company is concrete. One senior leader who owns it. An operating budget that doesn’t get cut at the first margin pinch. A mandate that survives quarter to quarter. The hub never delivers any of those.

5. Pace the portfolio, don’t kill it. When margin compresses, AI is usually first on the chopping block. The MIT research is clear that this is the most expensive cut a company can make, because the expertise you lose takes years to rebuild. The smarter move is to pace the work down. Carry the strongest two domains forward. Park the weaker ones in a way you can pick up later. The companies that come out of a downturn ahead are the ones who didn’t zero out their learning while everyone else did.

None of this is exotic. It’s the same delivery discipline that worked on the tailboard project nine years ago. The reason it doesn’t get done in AI today isn’t that the principles are unclear. It’s that the board wants a press release, the vendor wants a contract, and the CIO wants to look like they’re moving. The friction those three create is the friction that produces a 95% pilot failure rate.

If you’re in the middle of that decision right now, with the AI line item already on the budget and the board breathing down your neck, remember this. The path through the fog is small experiments, fast learning, and the willingness to stop performing AI for a board that has already been told the answer.

Thanks for reading!

The MIT Sloan article by O’Connor and Meyer is the source worth your time. Find it and read it before you sign the next AI contract.

Where have you seen this go right? Where have you seen it go sideways? Drop a comment or send me a note at jt@jasontate.ca. Push back on anything I got wrong.

The newsletter isn’t the conversation. The conversation is the conversation.

See you next Friday.

Best,

JT

There’s a number in a recent MIT Sloan piece that I haven’t been able to shake.

2.24 out of 10.

That’s the average approval rating truck drivers gave to the AI-enabled cameras installed in their cabs. Cameras meant to improve safety. Cameras meant to help them.

Nobody thought to ask.

The camera was sold in a boardroom to a VP of Fleet Operations who needed better safety scores. The vendor demo was clean. The contract was signed. The camera went in the cab.

And the people who sit in that cab for 10 hours a day gave it a 2.24.

This is not a technology problem. It’s a procurement problem. And it happens in every industry, every week, dressed up in the language of digital transformation.

​Let’s break it down.

Signal:

THE BUYER ISN’T THE USER. THAT GAP IS WHERE AI GOES TO DIE.

MIT Sloan’s Ganes Kesari has spent years watching AI roll out across conservative industries, and his diagnosis is worth sitting with: AI doesn’t fail because the technology is wrong. It fails because leaders underestimate the human and operational context in which AI tools are introduced.

Three blockers keep showing up:

1. AI feels inaccessible and scary.

2. AI looks like extra work.

3. AI benefits don’t seem worth the pain.

All three are symptoms of the same root problem. The person who signs the contract is not the person who has to use the tool. The VP approved the camera. The driver lives with it.

Here’s what that gap looks like in practice. The demo impresses the buyer. The UI confuses the user. The vendor builds a training program. Leadership schedules mandatory sessions. Adoption numbers come in low. The tool gets blamed. Nobody asks why.

I’ve written about the readiness gap before (From Signal to Scale - Issue 2026.06 is still worth your time if you missed it). But this is a harder problem than readiness. Readiness assumes the tool was right and the team wasn’t prepared. What I’m describing here is buying the wrong tool entirely, because nobody building it started with the person.

There’s a line from Kesari’s piece that cut through everything else: “The cost of learning feels personal, but the benefits feel abstract and impersonal.” Read that twice. The person being asked to change carries the cost. The executive who signed the deal claims the win. Nobody in that arrangement has the same stakes.

Here’s the design question that enterprise AI procurement never asks: how should this make the person feel?

Not “what can it do?” Not “what does it integrate with?” Not “what’s the ROI at 18 months?” How should it feel to the dispatcher at 6am, to the technician in the yard, to the driver 400 kilometres from home?

Apple started there. Every time. Jobs didn’t begin with a feature list. He began with a feeling and worked backwards to the technology that could create it. The iPod wasn’t “a 1GB storage device.” It was a thousand songs in your pocket. Confident. Light. Yours. The technology served the feeling, not the other way around. When Apple got it right, you didn’t need a training program. You needed about 90 seconds and a sense of wonder.

Enterprise AI does the exact opposite. It starts with capability, runs it through a procurement process designed for the buyer, and hopes that feeling follows somewhere downstream. The demo shows what the model can do. The contract specifies features. The implementation plan lists integrations. Nobody in that chain ever asked what it should feel like to the person doing the actual work. And if the answer was “it should feel like surveillance,” nobody said it out loud​

That’s how you get a 2.24.

If you need a training program to get people using your product, the product is already broken. That’s not a harsh take. That’s design. Face ID didn’t come with a manual. Uber didn’t need a certification course. Good design is invisible. The minute you’re scheduling mandatory sessions to explain a menu to a warehouse worker at the end of a 10-hour shift, you’ve already lost. You just haven’t admitted it yet.

The real blocker isn’t resistance to AI. It’s change fatigue, compounded by a decade of technology rollouts that promised everything and delivered a worse version of the old process with a new login. People aren’t wrong to be skeptical. They’ve been here before.

Scale:

THE MOAT YOU KEEP OVERRIDING

The fix isn’t complicated. It’s just inconvenient for the people who control the budget.

1. Put users in the room during vendor selection. Not as observers. As evaluators. Give the dispatcher, the technician, and the driver a scorecard. Let them run part of the demo. The gap between what a vendor shows a VP and what a front-line worker actually needs surfaces fast when the right people are in the room. You want that tension before the contract, not after. It is significantly cheaper to kill a bad deal than to manage a failed rollout for 18 months.

2. Run a 30-day shadow pilot with your most skeptical users. Not your enthusiastic early adopters. Pick five people who represent your hardest cases and embed the tool in their actual workday. Watch where they stop using it. Where they work around it. Where they swear at it. That friction map is worth more than any analyst report or vendor case study. The places where skeptics give up are exactly where the product failed to start with the person.

3. Measure the problem, not the deployment. Most go-live dashboards track seats licensed, logins, and training completion. None of that tells you if the tool is working. Tie your success metric to the pain the tool was supposed to solve. If it was supposed to cut time on work orders, measure time on work orders. New KPIs trigger debate and delay action. Familiar metrics accelerate it. Connect AI value to the numbers your people already get judged on, and the conversation changes entirely.

4. Say thank you. I know how that sounds. Do it anyway. When a front-line worker adopts something new and it sticks, their manager rarely says a word. That silence reads as indifference. A five-minute conversation from a senior leader saying “I heard the new system is working for you, what do you think?” does more for adoption than a company-wide rollout email ever will. Pride is portable. People take their best work to where it gets recognized. Issue 18 made this point harder than I will here. Go read it if you haven’t.

Deep Dive:

The organizations that get AI adoption right aren’t the ones with the best tools or the biggest budgets. They’re the ones who asked a simple question before signing anything:

At what point in our procurement process does the person who actually has to use this get a vote?

Not a survey. Not a focus group after the fact. A real vote, with the power to change the outcome.

If you can’t answer that question cleanly, you already know why your last implementation struggled.

No deep dive this week. Just that one question on your next leadership agenda. It will do more work than a consultant’s deck.

Thanks for reading!

The Kesari piece in MIT Sloan is worth your full attention. Go read it before you approve your next AI spend.

​https://sloanreview.mit.edu/article/the-human-side-of-ai-adoption-lessons-from-the-field/​

Where have you seen this go right? Where have you seen it go sideways? Drop a comment or send me a note at jt@jasontate.ca. Push back on anything I got wrong.

The newsletter isn’t the conversation. The conversation is the conversation.

Have a great long weekend!

See you next Friday.

Best,

JT

I want to talk about something that has been bothering me. Not what AI does to jobs. What it does to the work.

There is a Todd Henry piece from last week that has been rattling around in my head. The argument is that efficiency culture does not kill craft in one decision. It kills it slowly.

The people on your team who care most about the work walk in with standards. Those standards came from experience, training, the scars of past projects that fell apart. They are real. This is why you hired them. And every time you choose speed over quality, those standards get sacrificed. Quietly. Without a meeting. Review cycles get cut. Proofing gets shortened. Verification gets skipped. Each one a small concession, framed as a reasonable trade-off for the deadline.​

After a hundred of those small overrides, the standards no longer belong here. Not because anyone said so out loud. Because the message landed every time the same way. This is not where standards matter.

Apply that to AI, and you get this issue.

​Let’s break it down.

Signal:

YOUR TEAM HAS ALREADY DONE THE MATH

Resume Now surveyed more than 1,000 employed adults in March about how AI will affect work in 2026. The numbers are not subtle...

• 63% expect AI to make the workplace feel less human this year.

• 57% rank skill erosion as the biggest workforce issue, ahead of job displacement at 49%.

• 42% cite dehumanization of work as a top AI workforce concern.

• 20% rank loss of creativity and critical thinking as their personal top AI fear.

Read those again.

Your team is not waiting for the leadership offsite to debate AI strategy. They have already done the math. They are watching every time you ship something they know could have been better.

By the time you sit down to discuss culture, they are already grieving.

Scale:

THE MOAT YOU KEEP OVERRIDING

And this is where I see most leadership teams trip up.

The teams I work with are still framing AI as a productivity question. How fast can we ship. How much can we cut. How many hours can we save. That framing was fine in 2023 when access to AI was rare. It is wrong now, because access to AI is not rare anymore. Anyone can buy the tools. Anyone can prompt the model. Whatever speed advantage you got from AI last quarter, your competitor got it this quarter.

So if AI is no longer the moat, what is?

Deloitte’s 2026 Human Capital Trends report has a clean answer: technology is replicable, people are not. Their data says 59% of organizations are taking a tech-focused approach to AI, and those organizations are 1.6x more likely to miss their AI ROI targets compared to organizations taking a human-centric approach. Microsoft’s 2026 Work Trend Index says the same thing in different math. Organizational factors (culture, manager support, talent practices) drive 67% of AI’s reported impact. Individual mindset and behaviour drive 32%. Roughly 2:1 in favour of culture.

The moat is not the AI tool. The moat is the team that uses it well, judges its output well, and shapes the work into something a competitor cannot copy. That is craft. That is taste. That is judgement under uncertainty.

And that is exactly what gets eroded every time you ship “good enough.”

HOW THE MOAT FALLS

Henry’s mechanism is the part of his piece that hit me hardest. The moat does not fall in one decision. It falls in hundreds of small “good enough” overrides. The designer who wanted one more pass. The writer who wanted to rework the framing. The analyst who flagged a number that did not look right.

The first override, they absorb. The second, they adjust. By the 10th, they have stopped bringing their best judgement to the table. They have learned that the answer will always be “we do not have time.”

AI compounds this pattern, because every override is now 10 times cheaper than it used to be. The Copilot-drafted deck is acceptable. The Claude-edited memo is fine. The auto-summarized client report passes the bar. So you ship. And the person on your team who knows it could have been 30% better watches you ship.

Then the next time, they stop flagging it.

Then the time after that, they stop trying.

THREE MOVES TO DEFEND THE MOAT

Henry gives you three moves. Each one needs to go further when AI is in the mix.

1. Name the trade-off, and label the work

Henry says when you choose speed over quality, say so out loud. Acknowledge the cost. “We are shipping this knowing it is not where we would want it. I see the difference, even if we cannot close the gap this time.”

In an AI workflow, take it one step further. Label the mode of work. Microsoft’s Work Trend Index identifies 4 collaboration patterns: Author (you produce, AI helps when called), Editor (you set the intent, AI drafts, you edit), Director (you spec, AI executes), Orchestrator (you design the system, agents run in parallel).

These modes are not the same and your team should know which one applies. A board memo in Director mode is a different beast than the same memo in Author mode. Tell them which is which. Hidden AI is the new “good enough.” When craft people cannot tell where AI ended and a human started, they assume the worst.

2. Designate the human-only zones

Henry says be intentional about where quality is non-negotiable. Which projects get the extra pass. Which deadlines are real and which are negotiable.

In an AI saturated workflow, name the work that stays human. The strategic memo. The first draft of a board narrative. The framing of a client conversation. The voice of an apology email to a customer who was hurt by something your company did.

This is not a productivity sin. It is a craft preservation move. EY’s research describes a practice some of their pharma clients use, called AI Huddles. Weekly sessions where the technical team and the domain experts interpret AI output together. The point is not to debate the model. The point is to keep the human muscle in shape.

3. Ask what part of the work has their fingerprint

Henry’s third move is to ask craft-oriented people what would make them feel proud of the work. In an AI workflow, the question gets sharper.

What part of this work has your fingerprint on it.

If your team writes a proposal where AI drafted the body and a human edited the framing, the framing is the fingerprint. Make that visible. Talk about it. Credit it. If the only part a human did is hit “send,” they have nothing to point to and they will stop showing up. Pride is portable. People take their best work to where it is recognized.

WHAT YOU ARE ACTUALLY PROTECTING

The Gerlich study (n=666) and a recent Frontiers in Medicine paper on AI deskilling confirm what your craft people already feel. When humans repeatedly offload cognitive work to AI, the prefrontal cortex becomes less active. Independent reasoning capacity drops. Researchers now use the term “AI-induced cognitive atrophy.”

Translation: every override that lets “good enough” win is not just a one-time loss of quality. It is a one-time training rep that your team did not get. Skip enough reps and the muscle atrophies. The capacity to think, judge, and shape the work weakens. And the moat is gone.

The Resume Now data from the top of today’s rant is your team’s early warning system. 20% of workers are already worried about losing their creativity and critical thinking. They know what is at stake. They are watching to see if you do.

Deep Dive:

No deep dive this week. Instead, one question I would put on your next leadership agenda.

In the last 30 days, what did we ship as “good enough” that someone on the team flagged as below standard? What signal did we send when we shipped it?

The leaders who can answer this are doing the harder work. The leaders who cannot are watching their moat erode in real time.

That is your moat. Defend it.

Thanks for reading!

I am the PA announcer for the University of Calgary Dinos. McMahon Stadium has no autocorrect. Reminds me every game why some work has to stay yours.

It is kinda funny, this week’s piece spent more time being rewritten than written. Apparently ‘good enough’ is harder to ship when you wrote about it.

The irony is not lost.

See you next Friday.

Best,

JT

PS - If someone forwarded this to you and you want it in your inbox directly, subscribe HERE.

Years ago, I worked with a utility company that had just lost a worker in the field.

Their ask was simple. Move our SOPs from 3-ring binders to iPads. Substitute paper for a screen. Ship it. Done.

We didn’t stop there.

We built something that changed how the work happened, not just where it was stored.

That engagement taught me something I keep using. The technology mattered. The judgment about how to use the technology mattered more. Technology without the judgment behind it would have just been a worse binder.

I’ve been thinking about that engagement a lot this month. Three articles I read are all making the same warning, in different vocabulary, about the moment we’re in with AI. Different sources, different audiences, same line in the sand.

That’s a Signal worth slowing down for.

Let’s break it down.

Signal:

The pressure is real, and it’s pushing leaders sideways

Boards ask about AI. Customers ask. Investors ask. Most leaders I talk to feel like they’re behind, even when they’re using it more than they admit.

That pressure produces a specific failure mode: people start handing off the calls they get paid to make. Not the prep work. Not the data crunching. The judgment.

Wrong trade. AI compresses time. It does not compress responsibility. The faster you move, the more deliberately you have to think about what you’re still on the hook for.

Dr. Ruben Pedentura’s SAMR model has been kicking around education for decades. Anyone who knows me has heard me beat that drum for most of them. Substitution, Augmentation, Modification, Redefinition. Most digital transformation work, in my experience, stops at Substitution. We took a paper binder and made it a PDF on an iPad. Same job, different surface. We took an in-person meeting and made it a Zoom call. Same meeting, lower lighting.

Same thing’s happening with AI right now. People are using it as a substitute for thinking instead of as a tool that changes what thinking can produce. That’s a missed opportunity at best. At worst, well I’d rather not think about it.

Digital transformation has largely been performative. AI transformation is heading the same way unless leaders draw a clearer line.

Here’s what each of the 3 articles says about that line, and what to do about it.

Article 1: Get literate enough to ask the right question

Source: MIT Sloan Executive Education, “AI and Leadership: Navigating Strategy, Ethics, and Opportunity” (April 24, 2026)

The MIT Sloan Executive Education team makes a useful argument for leaders who feel like they need a computer science degree to keep up. They cite Gallup data showing 69% of leaders now use AI at work, and 19% use it daily. That’s a lot of people relying on outputs they don’t fully understand.

Their position: you don’t need to build models to lead well in this environment. You need enough literacy to interpret what the model is telling you, and to connect it to what the business is actually trying to do. They call it AI literacy. It’s the difference between knowing how a tool works and knowing whether the tool is solving a real problem.

The trap they flag is one I see every week. AI initiatives that aren’t tied to clear business priorities tend to produce insights nobody can act on. Pilots get launched. Decks get presented. Nothing changes. 6 months later, the tool is forgotten and the budget is gone.

Whenever a student or a client tells me they want to use AI, my first question is the same. What problem are you trying to solve? If they can’t answer, we’re not having a serious conversation. We’re doing technology in search of a problem, which is the most expensive form of theatre I know.

Takeaway: Get literate enough to ask the right questions. Tie every AI initiative to a business outcome you can measure. Then build the team’s ability to use it, not just yours.

Article 2: Speed is not the same as quality

Source: Loeb Leadership, “Why Human Judgment Is the Ultimate Competitive Advantage in the AI Era” (April 27, 2026)

The Loeb piece makes a sharper point. AI outputs sound authoritative because they’re confident, fluent, and well-packaged. They feel like answers.

They’re not. They’re patterns the system found inside the data and the design choices it was given. Every output reflects somebody’s assumptions about what counts as a good answer. The model doesn’t tell you that. You have to remember it.

Their warning lands hard: when you compress decision time, the margin for error compresses too. Speed without scrutiny gets you to the wrong conclusion faster than you used to get there. That’s not progress. That’s risk dressed up as productivity.

Loeb pulls out 4 habits worth keeping in this environment:

1. Frame the trade-offs clearly

2. Look at second-order consequences

3. Sort signal from noise

4. Resist the pull to certainty

None of those are technical. They’re the same calls a good operator has always made. The only thing that changed is the speed of the inputs hitting your desk.

The other piece of their argument I want to flag is governance. AI governance gets called a compliance issue or an IT problem. It’s neither. The piece argues, and I agree, that AI governance is a leadership job. Who validates outputs before they shape decisions. What escalates and to whom. How you watch for bias and drift. Where accountability sits when something goes sideways. Those are leader questions. They don’t get answered by a vendor or a Slack channel for the IT team.

If you haven’t had this conversation with your senior team yet, that’s your tell. Not “are we using AI enough.” But “do we know who is on the hook when AI gets it wrong.” The answer needs to be a person, not a policy.

I’ll say this part plainly. When one of the big 5 consulting firms hands you an 800-page report on how AI will transform your business but spends no actual time inside your business understanding how the work works, that report is bullshit. Governance is not delegated. It’s yours.

Takeaway: AI outputs are not neutral. Treat them like a draft from a smart but biased analyst. Speed is fine. Skipping the second look is not. And governance lives in your office, not down the hall.

Article 3: Tasks belong to software. Trust belongs to you.

Source: MIT Sloan Management Review, “When Not to Use AI” by Benjamin Laker (March 30, 2026)

This is the cleanest framing I’ve read on the subject. Laker splits your work into 2 buckets: tasks and trust.

Tasks are repeatable. They benefit from speed. Hand those off. Let the model draft the timeline. Let it crunch the numbers. Let it generate the slide skeleton. That’s what it’s good at.

Trust is the human currency of management. Feedback. Relationships. Difficult news. Hiring calls where you’re reading resilience between the lines of how someone tells a story. Strategy decisions that hit how the team feels about the future. Don’t hand those off.

His test, which I think is the best line in any of these 3 articles: would I stand by this if my name were on it alone? Would I say it out loud to someone I respect? If the answer is no, you slow down and re-engage.

Laker also flags a sneaky failure. AI tools are designed to be agreeable. They produce arguments that support whatever direction you’re already leaning. That feels like decisiveness. It’s actually a narrowing of your thinking. You ask a leading question, you get a confirmation. You feel sharp. You’re actually getting dumber, slowly.

His fix is the one I’d steal first. Occasionally instruct the tool to argue the other side. If you’re ready to reorganize a team, ask for the strongest case against. If you’re set on hiring someone, ask for the reasons it might not work. Force the counter-argument before you commit.

The line that stuck with me from his piece is about leaving the lifting to the machine and the leading to you. That’s the whole game right there.

Takeaway: Sort your week into tasks and trust. Software does the tasks. You do the trust. Once a week, ask the model to disagree with you, hard.

Scale:

The mistakes I see most often

Across the 3 articles, the same failure modes show up. Worth naming them plain.

Sending what the model wrote. If the output is going out under your name, your hands need to be on it. Skim doesn’t count. Edit until you’d defend every sentence to the person on the receiving end.

Treating velocity as virtue. Faster decisions aren’t better decisions. They’re just faster. Board pressure to “move on AI” doesn’t override your job to think. Decision quality has to outpace decision velocity, and that gap is where you actually earn your title.

Letting the model agree with you. If you only ask AI to support your view, you’re using it as a mirror. Mirrors don’t sharpen thinking. Counter-arguments do.

Outsourcing the relational calls. Performance feedback. Layoffs. Promotions. The opening of a hard conversation. None of these belong to a model. They belong to the person who has to live with the result.

Treating governance as somebody else’s problem. Compliance is not governance. IT policy is not governance. The questions about who validates, who escalates, and who’s accountable when things go wrong have to be answered at the leadership table. If they’re not, you have a hole. The hole shows up later, usually in front of customers or regulators.

Try this for a week

No new platform required. Better hygiene on what you’re already doing will get you most of the way there.

Pick 3 decisions on your calendar for the coming week.

For the first one, go AI-free. No prompts, no drafts, no summaries. Think it through yourself, with whatever inputs you’d normally have. Notice what changes when you have to carry the full weight. Most leaders find their reasoning gets sharper, not slower.

For the second, let AI do the prep. The data pull, the timeline, the meeting agenda. Then make the call yourself, in your own words, with your name on it. Keep the lifting and the leading separate. On purpose.

For the third, push back on yourself. Ask AI for the strongest argument against the direction you’re leaning. Read it carefully. If it doesn’t change your mind, you’ll know more clearly why. If it does, you just saved yourself a bad call.

That’s a workable first month. Snow melts from the edges. The leaders who handle this well will not be the ones who shouted “AI everywhere” the loudest. They’ll be the ones who quietly figured out where it belonged in their work and where it didn’t, and who kept their hands on the wheel where it counted.

The 3 articles converge on one idea, even though none of them say it quite this way. AI is not the threat to leadership. The threat is leaders who quietly stop leading because the tool is doing enough of the work to look fine from the outside.

It’s not fine. The people on your team can tell the difference between a message you wrote and a message you forwarded. The candidate you interviewed can tell whether you were actually listening. The customer reading your apology email knows whether somebody meant it.

Let the machine lift. You lead.

Deep Dive:

No deep dive this week. The 3 articles above are doing the work, and I’d rather you spend your time with them than with another piece of mine layered on top.

Go read the 3 articles. I’ll meet you back here next week.

Thanks for reading!

Where have you seen this go right? Where have you seen it go sideways?

I’m genuinely curious. The patterns are still being written, and the ones we name out loud are the ones we get better at handling.

Drop a comment, send me a note at jt@jasontate.ca, or push back on anything I got wrong. The newsletter isn’t the conversation.

The conversation is the conversation.

See you next Friday.

Best,

JT

PS - If someone forwarded this to you and you want it in your inbox directly, subscribe HERE.

This issue is a bit different…I feel like I’ve been saying that a lot lately?!

I’ve been spending time lately working on NFP strategy, through my partnership with the Osborne Group. Specifically, the technology and AI dimensions. And I keep finding the same thing: sector leaders who are smart, mission-driven, and genuinely worried about doing right by their communities, navigating a technology environment that’s moving faster than their governance can keep up.

This week I want to share three signals from that work. The numbers are Canadian. The risk is real. And the window to get ahead of it is shorter than most boards realize.​

Let’s break it down.​

Signal:

Signal 1: The NFP Sector Is Using AI. Nobody Is Running It.

A January 2026 report from Imagine Canada and the Canadian Centre for Nonprofit Digital Resilience surveyed over 900 Canadian nonprofits. 80% said their organization is using AI in some form. Only 10% have a formal AI policy. 64% of those using AI have no policy at all, and aren’t developing one.

​That gap is where the risk lives. Not in the technology itself. In the absence of any decision about how it gets used, by whom, for what, and with whose data.​

62% of respondents said they’re aware of reputational risks. 60% flagged legal and ethical concerns. Most of them are doing nothing about it.

This is the pattern I see in the field: awareness without action. The sector knows the roof has a leak. Nobody’s called the contractor yet.

​

Signal 2: The Sector Is Going It Alone.

Of those same organizations, only 9% have engaged an external consultant for AI support. Larger organizations, those with revenues over $10M, are more likely to seek outside help, but only 27% of that group has done it.

Meanwhile, the Ontario Nonprofit Network reports that 60% of small charities say they lack digital skills, lack a strategy, and struggle to fully adopt the tools they already have. Only 13% have dedicated tech staff and a roadmap.

The combination of those two data points is the problem. The organizations with the least internal capacity are also the least likely to bring in outside expertise. They’re not ignoring the problem. They’re stretched too thin to solve it.

​

Signal 3: The Barrier Isn’t Money. It’s Knowledge.

This one surprised me. In most nonprofit research, limited resources are the top challenge. Not here. The Imagine Canada report found that the biggest barriers to AI adoption and growth are uncertainty and limited hands-on experience, not finances.

Funding matters when scaling, but it’s not what’s keeping people stuck at the starting line. What’s keeping them stuck is not knowing where to start, not trusting what they know, and not having anyone in the room who’s done it before.

That’s a different problem than budget. And it calls for a different solution.

​

Scale:

Scale 1: Run a Data Audit Before You Add Anything New

Before an NFP adds another tool, the question to answer is: what data do we already hold, who has access to it, and what would happen if it walked out the door?

Start with a one-page data map. List every platform that touches donor, client, or staff information. Note who owns each one, who has admin access, and when access was last reviewed. That exercise alone will surface problems worth fixing.

Canada’s privacy law is being replaced. The expected successor to PIPEDA carries penalties up to $25M or 5% of global revenue. That’s not theoretical risk. That’s board-level exposure for an organization that’s never thought about it.

One process. One hour. More clarity than most organizations have had in years.

​

Scale 2: Get an Outside Perspective Before Someone Else Brings One In

The sector is navigating AI governance mostly by instinct. That works until it doesn’t, and when it doesn’t, the consequences tend to arrive in the form of a breach, a complaint, or a funder asking questions the organization can’t answer.

External perspective doesn’t have to mean a six-month engagement. A structured conversation with someone who’s done this work across organizations, who knows what good looks like and what failure looks like, is often enough to reset the direction.

The 9% of nonprofits that have brought in outside support are applying AI to more areas, with more confidence, and with less exposure. That’s not a coincidence.

​

Scale 3: Reframe the Question Your Board Is Asking

If your board is asking “should we have an AI strategy,” they’re a step behind the question. 80% of organizations already have staff using AI tools. The strategy ship has sailed. The governance ship hasn’t left the dock.

The better question is: which decisions do we need to make right now about how AI is used in this organization, and who owns each one?

That conversation takes 30 minutes with the right framing. It produces a short list of decisions, assigns owners, and turns an abstract risk into a workable agenda. That’s the starting point, and it’s within reach for any organization willing to have the meeting.

Deep Dive:

This week’s full piece goes further on all three signals: the regulatory exposure building under Canada’s pending privacy legislation, what the 9% of organizations with outside support are actually doing differently, and a framework for having the governance conversation with your board without it turning into a two-hour debate about tools.

Read: 80% of Canadian Nonprofits Are Using AI. 64% Have No Policy.

Thanks for reading!

I’m offering a free 30-minute call for NFP leaders, EDs, ops directors, and board members who want to work through what any of this means for their organization.

No pitch. No deck. Just a structured conversation about where you are, what’s at risk, and what a reasonable next step looks like. Interested? Book your 30-minute call.

See you next Friday.

Best,

JT

​

PS - If someone forwarded this to you and you want it in your inbox directly, subscribe HERE.

And if one of these signals hit closer to home than the others, drop a comment and tell me which one. I read everything.

This is Part 5. The last myth. The one that ties all the others together.

Over the last five weeks we’ve covered…

​Myth 1 - Fixing the Foundation​

​Myth 2 - Why Boring Tech Wins​

​Myth 3 - The Complacency Trap​

​Myth 4 - Why Your Employees Are Already Building the Future​

Every one of those, traces back to this.

What leaders say vs. what they actually do.

Let’s break it down.

Signal:

CEOs Are Saying One Thing and Doing Another. The Numbers Don’t Lie.​

BCG’s 2026 AI Radar report surveyed 640 CEOs and 2,360 senior leaders. 82% are more optimistic about AI than a year ago. AI is a top-3 strategic priority for 2 out of 3 CEOs. Half believe their job stability depends on getting AI right this year.

That sounds like commitment. Then you read the rest of the data.

60% of those same CEOs admitted they’ve intentionally slowed AI implementation over concerns about errors and malfunctions. Only 6% plan to scale back spending if AI fails to deliver. They’re slowing down the work while refusing to cut the budget. Motion without movement.

EY’s survey of 500 US senior leaders found that 96% report AI-driven productivity gains. But 65% admit they can’t tie those gains to AI adoption. They’re reporting results they can’t measure.

The spending gap is just as wide. In 2024, 65% of executives predicted they’d invest at least $1 million in AI the following year. Only 58% actually did. 34% predicted $10 million or more. Only 23% hit that number.

And the clock is ticking. An HBR piece from this month found that 71% of global CIOs said AI budgets would be frozen or cut if value can’t be demonstrated within 2 years.

BCG’s own “Widening AI Gap” report found that deeply engaged C-level leaders are 12x more likely to be in the top 5% of companies winning with AI. Which means 95% of companies aren’t winning. The difference isn’t the technology. It’s leadership.

Scale:

Myth 5: Executives Are Hungry for AI Transformation.

Reality: They’re hungry to talk about it.

Andriole was blunt about this in 2017. The number of executives who really want to transform is small. The gap between what leaders say and what they do is wide. Nothing has changed.

Today, executive AI talk creates pressure. Subordinates do something performative. Something leadership can point to on the quarterly call. The result: a lot of motion and very little movement.

This is a systems problem. Donella Meadows identified information flow as 1 of the most important characteristics of a healthy system. When information moves accurately from edges to center, the system adapts. When it doesn’t, the system stagnates.

In most organizations, the information flow around AI is broken. Signals from frontline employees never reach the people making strategy decisions. Honest feedback gets punished. Telling leadership what they want to hear gets rewarded. Every layer of management filters the signal until it’s useless.

Kim Scott called this the absence of radical candor. Without it, the information that would drive real change gets sanitized before it reaches the people who need it most.

VG’s Three Box framework explains why it persists. Box 1, managing the present, is where all the reporting happens. It’s what executives are evaluated on. Box 3, creating the future, requires admitting what’s not working. But executives who built their careers in Box 1 have every incentive to protect it. So they fund Box 3 initiatives without doing the Box 2 work. Those initiatives get strangled by Box 1 operating logic. Then leadership blames execution. “The team couldn’t deliver.” The team was never set up to succeed.

George Westerman of MIT captured the result: when transformation is done wrong, all you have is a really fast caterpillar. That’s what most AI “transformation” looks like right now. Fast caterpillars. Multiple initiatives in flight. No vision of a butterfly in sight.

Meadows argued that the most powerful place to intervene in a system is at the level of its goals. If the real, unstated goal of your AI strategy is “say the right words on the quarterly call,” the system will produce exactly that. Words. Not results.

If you want different behavior, change the goal. Make it “build the capacity to adapt” instead of “implement AI.” The tool is secondary. The muscle is primary.

Because it’s hard, it’s worth doing.

Deep Dive:

This week’s deep dive closes the series. It covers the full anatomy of performative transformation, why the information flow breaks, what the duty to dissent looks like in practice, and the 5 steps to make the goal real instead of rhetorical.

Read: Executives Love Talking About AI. The Numbers Say They’re Faking It.

Thanks for reading!

That’s the series. 5 myths. 5 weeks. If you’ve been here from the start, thank you. If you jumped in at the end, go back to Myth 1 - Fixing the Foundation.

The full Five Myths About AI Transformation paper is where this all started. If you haven’t read it yet, start there.

I’d love to hear about which myth hit the hardest for you? Which one are you living inside right now? Leave a comment. The conversation doesn’t end here. It starts.

See you next Friday.

Best,

JT

This is Part 4 of the Five Myths About AI Transformation series. First was Myth 1 - You Don’t Need an AI Strategy, fixing your foundation. Then, Myth 2 - AI Is the Technology That Changes Everything‍ ‍boring tech beats shiny AI. Followed by, Myth 3 - Profitable Companies Are Best Positioned for AI, your profits are making you complacent.

This week is my favourite myth. Because it flips the entire conversation. It is all about where change actually comes from.

Let’s break it down.

The Signal:

Half Your Workforce Is Already Using AI Without Permission. And Your Executives Are Leading the Charge.

A BlackFog survey published in January 2026 landed a number that should retire the “should we adopt AI?” conversation forever. 49% of workers at companies with more than 500 employees use AI tools without employer approval. 86% are using AI weekly.

But here’s where the story turns. This isn’t a rogue employee problem. 69% of C-suite members said they’re fine with it. Among executives and senior managers specifically, 93% admitted to using shadow AI tools themselves.

93%. The same people who haven’t approved those tools for their teams.

A Cybernews survey found 59% of US employees use unapproved AI tools, 75% have shared sensitive data with them, and 57% say their direct managers know about it and support it. 46% said they’d keep using these tools even if explicitly banned.

MIT’s “GenAI Divide” report called it a “shadow AI economy.” While only 40% of companies have official AI subscriptions, workers in over 90% of organizations use personal AI tools for work.

These aren’t rogue actors. These are your canaries. They’ve already done the discovery phase. They know where the friction is because they live in it every day.

The Scale:

Myth 4: We Need to Disrupt Our Industry Before Someone Else Does.

Reality: Snow melts from the edges.

I disagree with Andriole on this one. At least partially.

He argued that disruption almost never comes from market leaders. It comes from startups making bold bets. The evidence is strong. But I’ve seen established companies make bold bets too. Apple. Google. Salesforce. The difference is where those bets originate.

Often, it doesn’t start in the executive suite. It starts with individual employees figuring things out in their silos. Building their own workflows. Finding workarounds. Solving problems nobody gave them permission to solve.

Donella Meadows described this as self-organization. The ability of a system’s components to create new structures without top-down direction. The healthiest systems create space for experimentation at the edges. They watch for signals. They amplify what works.

VG’s Three Box Solution connects directly. Your employees are doing Box 3 work, creating the future, every time they solve a problem with an unapproved tool. They’re testing and learning what works. But the policies that ban unapproved tools, the approval processes that take months, the governance frameworks built for a different era? Those are chains, not roots. Box 2 work, the selective forgetting, means clearing them out.

Choudary’s Shein example from HBR reinforces this. Traditional fashion companies organize around the season. The executive suite decides what the market wants months in advance. Shein reduced the unit of work to continuous experiments that test demand in real time. The learning happens at the edges.

MIT’s research backs it up. The 5% of companies that succeeded with AI empowered line managers to drive adoption, not central AI labs. Vendor-purchased tools succeeded 67% of the time. Internal builds succeeded a third as often. The users knew what worked. The architects didn’t.

Academic research published in 2025 found that employees use AI roughly 3x more often than managers estimate. The researchers called it “shadow user innovation” and argued companies should treat it as decentralized R&D, not a compliance violation.

Fujitsu put it best: “Trying to ban shadow AI is like trying to ban Googling in 2002.”

The question isn’t whether to disrupt your industry. The question is whether you know what your people are already building. And whether you’ve created conditions for those experiments to surface, or whether you’re suppressing them with policies and fear.

The Deep Dive:

This week’s deep dive goes all the way in on Myth 4, including how to find your canaries, run a shadow AI audit that actually helps, and build the conditions for bottom-up innovation to become organizational capability. Read: Your Employees Are Already Building the Future. Are You Listening?

Thanks for reading!

Next week: Myth 5. “Executives Are Hungry for AI Transformation.” The final myth. And the most uncomfortable one for leadership.

If your company has an AI policy, ask yourself: was it written to protect the organization, or to make it better? There’s a big difference. Hit reply and tell me which one yours is

See you next Friday.

Best,

JT

P.S. — Read the full Five Myths Paper on Substack.

We are pushing this out a day early, in anticipation of the long weekend. No matter what you are celebrating, I hope it is full of love and joy with friends and family.

This is Part 3 of the Five Myths About AI Transformation series. First we hit Myth 1 - You Don’t Need an AI Strategy, fixing the foundation before chasing AI. Last week, Myth 2 - AI Is the Technology That Changes Everything was about why boring technology outperforms the shiny stuff. This week is different. This one is about YOU!

Let’s break it down.

The Signal:

Forrester Named Your Disengaged Employees. They Call Them “Coasters.”

Forrester’s Predictions 2026: The Future of Work report introduced a category that should make every profitable company uncomfortable. “Coasters.” Disengaged workers who don’t think their employer deserves their energy. Not quitting. Not sabotaging. Just doing the minimum.

This group hit 27% in 2024, dipped to 25% in 2025, and Forrester expects it to climb to 28% in 2026.

The why isn’t complicated. Employees watched colleagues get laid off for AI that never materialized. 55% of employers who made those cuts already regret it. Entry-level positions are disappearing, locking out the generation with the highest AI readiness (Gen Z at 22%, Baby Boomers at 6%). And companies that could afford to invest in training mostly didn’t. Only 23% of AI decision-makers said their organizations offered prompt engineering training in 2025.

Meanwhile, investors are paying attention. Mercer’s Global Talent Trends 2026 found that 97% of investors said funding decisions would be negatively impacted by companies that fail to upskill workers on AI. 77% favor companies building their workforce alongside the technology.

The profitable companies have the resources. Most aren’t using them. And both the talent market and the investment market are starting to keep score.

The Scale:

Myth 3: Profitable Companies Are Best Positioned for AI

Reality: Comfort is the enemy of adaptation.

Both types of companies come to me. The ones in pain and the ones riding high. The ones in pain are afraid it will get worse. The ones riding high have the opposite problem. They’re afraid change will work, and they’ll have to rethink how they do business.

Both end up in the same place: avoidance.

Stephen Andriole made this point in 2017. Profitable companies are the least likely to transform successfully because they have the least incentive to change. How many successful companies, without market pressure, have truly rethought their business models? Very few.

Clayton Christensen spent decades explaining why. Incumbents don’t fail because they’re blind. They fail because their business environment doesn’t allow them to pursue new approaches when those approaches aren’t profitable enough at first. Everything in the structure is optimized to protect what’s currently working. The focus on existing customers becomes locked into internal processes. Even senior managers can’t shift investment away from what’s paying the bills.

Vijay Govindarajan (VG) built a framework that names this imbalance. His Three Box Solution says companies need to manage the present (Box 1), selectively forget the past (Box 2), and create the future (Box 3). Profitable companies pour everything into Box 1. Box 2, the forgetting, gets skipped because the past is still working. And Box 3 gets lip service. VG’s sharpest insight: what you need to forget is a future weakness, but it’s embedded in your current strength. That’s why it’s so hard to let go.

Donella Meadows called these balancing feedback loops. Forces that resist change because the current state produces acceptable results. The company is profitable. The shareholders are happy. The system whispers: don’t rock the boat.

But “well enough” is a moving target. McKinsey found that organizations adopting a digitally ready setup can quadruple their 5-year revenue growth and nearly triple total return to shareholders. PwC’s analysis of close to a billion job postings found that AI-exposed industries see 3x higher revenue-per-employee growth. Wages in AI-exposed roles carry a 56% premium, up from 25% a year earlier.

Those numbers don’t describe companies replacing workers with AI. They describe companies investing in their people alongside AI. The market is pricing in scarcity of humans who know how to work with technology.

Sangeet Paul Choudary calls the resistance “architectural self-preservation.” Units of work define roles, expertise, and status. Changing them redistributes influence. Leaders sense resistance and call it culture. Choudary says it’s the system protecting itself. That protection is strongest when profits are good.

If you’re profitable today, that’s a position of strength. Not a destination. Use it to fund the foundation, train your people, and build the muscle for adaptation while you still have the resources.

Because by the time the market forces your hand, the cost of catching up will be 10x what staying current would have cost.

The best time to fix the roof is when the sun is shining.

The Deep Dive:

This week’s deep dive goes all the way in on Myth 3, including VG’s Three Box Solution applied to the AI adoption challenge, the 2 traps profitable companies fall into, and the 5 specific moves to make while you still have the resources. Read: Your Profitability is Making You Worse at This.

This is Part 2 of the Five Myths About AI Transformation series. Last week we covered Myth 1 - You Don’t Need an AI Strategy, and Microsoft's Copilot reality check. This week's story is messier, funnier, and more instructive.

Let's break it down.

The Signal:

Klarna's AI Experiment Went Exactly How You'd Expect.

In 2023, Klarna stopped hiring. By 2024, they'd partnered with OpenAI, cut roughly 700 customer service jobs, and replaced them with AI agents. CEO Sebastian Siemiatkowski told the world that "AI can already do all of the jobs that we, as humans, do." They claimed $10 million in savings. Headlines everywhere.

By early 2025, customers had a different opinion. Satisfaction dropped. Complaints piled up. Internal reviews revealed the AI couldn't handle nuance, empathy, or the kind of problem-solving that people with billing disputes actually need. Customers described the responses as generic and repetitive.

Siemiatkowski admitted it: "We focused too much on efficiency and cost. The result was lower quality, and that's not sustainable."

Then it got worse. Klarna couldn't rehire fast enough. Reports surfaced that the company started pulling software engineers and marketers out of their actual jobs and into call centers to fill the gap. The CEO who said AI could do everyone's job was now reassigning his technical staff to answer phones.

The solution that would have worked from the start: AI handles the routine stuff, humans handle the complex stuff, and you design the process before you pick the tool.

That's not a headline. But it's what works.

The Scale:

Myth 2: AI Is the Technology That Changes Everything

Reality: The biggest wins still come from boring, proven technology applied to the right problem.

Klarna didn't have a technology problem. They had a sequencing problem. They applied the most sophisticated tool available to a workflow they hadn't redesigned. That's Substitution on the SAMR scale. Swap 1 thing for another. Same process. Same structure. Just cheaper. Until the quality collapsed.

This pattern is everywhere. MIT's "GenAI Divide" report found that companies were pouring more than half their AI budgets into sales and marketing tools. But the highest ROI came from somewhere nobody was looking: back-office automation. Document processing, compliance, internal workflows. The boring stuff.

Deloitte's 2026 State of AI report put a number on the gap. 74% of organizations hope to grow revenue through AI. Only 20% are actually doing so. The companies in the 20% didn't start with the flashiest use case. They started with the most friction.

Sangeet Paul Choudary made this case in Harvard Business Review last month. He studied Shein vs. traditional fashion houses. Traditional companies use AI to speed up design sketches, but the underlying structure stays sequential. They made the season faster. They didn't rethink the season. Shein reduced the unit of work from the seasonal collection to continuous small-batch experiments. AI doesn't sit on top as a fancy addition. It connects sensing, decision-making, and action into a single learning loop.

That's not AI changing everything. That's a company rethinking how it works, then using technology to make the new structure possible.

Stephen Andriole observed this back in 2017 in MIT Sloan. Most short-term impact comes from conventional operational technology. Not from the hot new thing. Think about Uber. The technology behind Uber wasn't exotic. Phone in your pocket, GPS, mobile payments, a well-designed app. All proven. The innovation was applying it to a business model problem.

Donella Meadows wrote that the most powerful place to intervene in a system is at the level of information flows and delays, not at the level of new components. If information moves slowly through your organization, if decisions stall because data lives in 7 different spreadsheets nobody reconciles, adding AI is like putting a turbocharger on a car with flat tires.

Fix the tires first. Then talk about turbochargers.

The Deep Dive:

This week's deep dive goes all the way in on Myth 2, including a practical sequence for finding the boring wins hiding in your business and the math that shows why they matter more than you think. Read: AI Gets the Headlines. Boring Technology Gets the Results.

I WANNA GO DEEP!

Thanks for reading!

Next week: Myth 3. "Profitable Companies Are Best Positioned for AI." Why comfort is the enemy of adaptation, and why the companies riding high are the ones most likely to get left behind.

If you know someone chasing AI when they should be fixing their plumbing, forward this. They probably won't thank you. But their customers will.

Hit reply and tell me: what's the most boring fix that produced the biggest result in your business?

See you next Friday.

Best,

JT

P.S. — Read the full Five Myths Paper on Substack.

This week I published Five Myths About AI Transformation. If you've been reading along the last few months, you've seen me pull at these threads already. AI fatigue. Feature stuffing. Organizational readiness. The gap between what tools can do and what companies are actually prepared to use them for. The paper ties all of that thinking together and goes deeper, connecting the patterns I've been seeing in the field to the systems thinking and business theory that explain why they keep repeating.

Starting this week and for the next 5 Fridays, I'm pulling each myth apart in its own issue. Deeper research. More context. Specific things you can do about it.

We're starting with the myth that got the whole paper started: Every company needs an AI strategy.

Let's break it down.

The Signal:

Microsoft's Copilot Reality Check Landed. Hard.

In January 2026, Microsoft finally revealed the number they'd been sitting on for 8 quarters. 15 million paid Copilot seats. Satya Nadella called it "record AI momentum." He said Copilot is "becoming a true daily habit."

The number he didn't volunteer: Microsoft has 450 million commercial Microsoft 365 users. 15 million is 3.3%.

After 2 years on the market. After $37.5 billion in capital expenditure in a single quarter.

Microsoft Spent $37.5 Billion on AI. Only 3.3% of Users Are Paying for It.

It gets more interesting. Recon Analytics surveyed 150,000 enterprise users in January. 70% initially preferred Copilot because it was already in their Office apps. After trying alternatives, only 8% kept choosing it. Copilot's paid subscriber share dropped 39% in 6 months.

Even Microsoft is walking it back. Reports surfaced that the company is pulling AI features from Windows 11 where usage doesn't justify the investment. SemiAnalysis put it bluntly: an outside competitor shipped a better AI experience on Microsoft's own application than Microsoft's $30-per-seat product could deliver.

The Scale:

Myth 1: Every Company Needs an AI Strategy

Reality: Most companies need to fix their foundation before AI can help.

Microsoft's Copilot adoption numbers aren't an AI failure. They're a foundation failure. Companies bolted Copilot onto dirty data, broken processes, and workflows nobody had examined. The tool worked. The systems underneath it didn't.

This pattern is everywhere. MIT's Project NANDA reviewed over 300 AI deployments in 2025 and found that 95% of enterprise AI pilots delivered no measurable impact on profit and loss. Between $30 and $40 billion in enterprise investment. Almost all of it producing zero return.

The core issue wasn't the models. MIT found that enterprise AI systems didn't retain feedback, didn't adapt to context, and didn't fit into real workflows. A CIO told the researchers: "We've seen dozens of demos this year. Maybe 1 or 2 are genuinely useful. The rest are wrappers or science projects."

Gartner put a number on the data problem. 63% of organizations either don't have or aren't sure they have the right data practices for AI. Their prediction: through 2026, organizations will abandon 60% of AI projects because the data isn't ready.

RAND Corporation measured an 80% overall failure rate. But here's the number that should change how you think about this: projects with sustained executive involvement succeeded 68% of the time. Projects that lost executive sponsorship within 6 months? 11%. That gap isn't technology. It's commitment.

This is the Substitution trap. Dr. Ruben Puentedura's SAMR model maps it precisely. Substitution swaps 1 tool for another with no functional change. That's what most AI adoption looks like. Swapping ChatGPT into a broken workflow and calling it transformation.

Sangeet Paul Choudary made this case in Harvard Business Review last month. He studied Figma vs. Adobe. Adobe did everything right. Moved to the cloud. Subscription model. Collaborative features. Still lost. Because Adobe changed the delivery mechanism. Figma changed how design work was organized. Substitution vs. Modification.

Donella Meadows wrote that you can't improve a system you can't describe. A Microsoft survey of 500 enterprise decision-makers across 13 countries found that only 22% strongly agreed their organization has clearly documented key processes and data dependencies.

If you're in the other 78%, the first honest question isn't "what's our AI strategy?"

It's "can we describe, in detail, how our business actually works right now?"

If the answer is no, start there. The AI can wait.

The Deep Dive:

This week's deep dive goes all the way in on Myth 1, including the practical steps for what to do instead and the business theory that explains why this pattern keeps repeating. Read: You Don't Need an AI Strategy. You Need to Know How Your Business Actually Works.

I WANNA GO DEEP!

Thanks for reading!

Next week: Myth 2. "AI Is the Technology That Changes Everything." Why the biggest wins still come from boring, proven technology applied to the right problem.*

If this landed, forward it to someone who's about to spend 6 figures on an AI strategy without documenting their processes first. They'll thank you later.

If you disagree, hit me up. I’d love to hear your thoughts. That's what fills my cup.

See you next Friday.

Best,

JT

P.S. — Read the full Five Myths Paper on Substack.

Companies moved from Microsoft Exchange to O365. From on-premise SharePoint to SharePoint in the cloud. From Word to Google Docs. They called it digital transformation.

It wasn’t. It was substitution. Swapping one tool for another and expecting different results.

Now it’s AI. Companies are swapping ChatGPT into broken workflows, bolting Copilot onto processes nobody has examined in a decade, and telling their boards they’re “transforming.” The budgets are bigger. The stakes are higher. The script is identical.

I’ve consulted with over 100 businesses on automation and AI implementation. The same mistakes show up in almost every engagement, regardless of industry, company size, or how much money they’ve spent. These aren’t random errors. They’re patterns.

In 2017, Stephen J. Andriole, a professor at Villanova University who had served as CTO at multiple organizations and directed a technology office at DARPA, published a piece called “Five Myths About Digital Transformation.” He had spent decades watching companies stumble through technology transitions and distilled the most common mistakes into five myths, each with a corresponding reality.

Nearly a decade later, every one of those myths has resurfaced in the AI conversation. Same patterns. Same blind spots. Same gap between what leaders say and what they do.

The question that kept nagging at me was: why? Why do capable people, running successful companies, keep making the same mistakes every time a new technology wave arrives?

The answer is systemic. Donella Meadows, who wrote Thinking in Systems, argued that a system’s purpose is revealed by its behavior, not by the goals it announces. When you look at how companies actually behave during technology transitions, not what they put in the press release, but what gets funded, staffed, and measured, the patterns become obvious. And predictable.

This paper walks through Andriole’s five myths, updated for AI, and uses systems thinking to explain why they persist. More importantly, it provides a practical framework for breaking the cycle. Because understanding the pattern is only useful if you can do something about it.

The Thesis

AI is not a project you implement. It is not a line item on a roadmap with a start date and a completion date.

AI changes the conditions under which business is done. It changes what your customers expect. It changes what your employees demand. It changes what your competitors are capable of, often before you realize they’ve moved.

The question is not “should we adopt AI?” The environment has already shifted. The question is whether your organization can adapt continuously, or whether it will keep treating each technology wave as a one-time event with a beginning, middle, and end.

Meadows would recognize this immediately. She described systems that survive as ones capable of self-organization, learning, and adaptation. The ones that fail are rigid. They’re built to protect a stable state that no longer exists.

Most companies are still protecting a state that no longer exists.

Myth 1: Every Company Needs an AI Strategy

Reality: Most companies need to fix their foundation before AI can help.

I worked with a major Canadian airline during the early mobile era. The smartphone had arrived, and every business felt the pressure to build an app. They needed an app because everyone needed an app.

The problem was that nobody asked the right first question: what does the customer actually need?

The airline’s underlying systems weren’t ready to provide the information needed to power an app. There was no clear signal about where to start. They ignored the client journey, jumped straight to “we need an app,” and built web apps that performed terribly. No offline capability. Slow. Disconnected from the data systems the app depended on. The functionality was basically useless.

Sound familiar? Today, companies are doing the same thing with AI. They’re swapping ChatGPT into existing broken workflows and calling it transformation.

It isn’t.

Dr. Ruben Puentedura developed a framework called the SAMR model that maps exactly to what I see in the field. The S stands for Substitution, exchanging one tool for another with no functional change. A is Augmentation, adding some functional improvement. M is Modification, redesigning the task because the technology makes it possible. R is Redefinition, doing something that was not even conceivable before.

Most companies are stuck at Substitution. At best, they reach Augmentation. That’s where “digital transformation” stopped for the majority of organizations I’ve worked with. AI is following the same pattern.

Andriole made this point in 2017: not every company, process, or business model requires digital transformation. The same applies to AI. Some companies genuinely do not need AI right now. What they need is to fix the systems they already have. Document their processes. Clean up their data. Connect the tools they’re already paying for.

Meadows wrote that you can’t improve a system you can’t describe. If you can’t model your existing processes, if your employees can’t articulate how work actually flows through the organization, adding AI will only magnify the mess.

Even Microsoft is backpedaling on some of their AI-everywhere approach because the discovery is proving out: most of the implementation is performative and delivering little value.

The first honest question is not “what’s our AI strategy?” It’s “can we describe, in detail, how our business actually works right now?”

If the answer is no, start there. The AI can wait.

Myth 2: AI Is the Technology That Changes Everything

Reality: The biggest wins still come from boring, proven technology applied to the right problem.

When companies come to me saying “we need AI,” the vast majority of them could benefit from it. But what they actually need first is algorithmic, systematic ways to solve business problems. More process automation than artificial intelligence.

The irony is that the opportunity for AI after building those systems is massive. But you have to build the foundation first.

Andriole observed the same thing in 2017. He argued that most short-term impact comes from conventional operational technology: networking, databases, enterprise software. Not from the hot new thing. The companies that won were the ones who applied mainstream tools already in consumers’ hands.

Think about Uber. The technology that made Uber possible wasn’t exotic. It was the phone in your pocket, GPS, mobile payments, a well-designed app. All built on infrastructure that already existed. The innovation was in applying proven technology to a business model problem.

AI gets the headlines. But in my experience, the companies that see the fastest return are the ones fixing something boring. Connecting two systems that don’t talk to each other. Removing manual steps from a workflow that hasn’t been updated in a decade. Building a simple process that handles routine decisions so humans can focus on the exceptions.

Meadows would frame it this way: the most powerful place to intervene in a system is often at the level of information flows and delays, not at the level of new components. If information moves slowly through your organization, if decisions take weeks because data lives in seven different spreadsheets that nobody reconciles, adding AI to that mess is like putting a turbocharger on a car with flat tires.

Fix the tires first. Then talk about turbochargers.

Myth 3: Profitable Companies Are Best Positioned for AI

Reality: Comfort is the enemy of adaptation.

Both types of companies come to me equally. The ones in pain and the ones riding high.

The companies in pain are afraid the pain will get worse if they don’t use AI. They worry they’ll become obsolete. They worry the investment will bankrupt them. They worry the implementation won’t work.

The companies riding high have the exact opposite problem. They’re worried it will work. That they’ll have to fundamentally change how they do business.

Both scenarios lead to the same behavior: avoidance. Both types of companies ignore the foundational work. Their data structures, their people structures, their technology structures were built and weathered for a different era. Both are looking for quick, easy solutions. Those solutions don’t exist.

Andriole nailed this in 2017. He argued that profitable companies are actually the least likely to transform successfully because they have the least incentive to change. How many successful companies, without market pressure, have truly rethought their business models? Very few.

Meadows explains this through what she called balancing feedback loops. These are forces within a system that actively resist change because the current state is producing acceptable results. The company is profitable. The shareholders are happy. The processes work well enough. The system whispers: don’t rock the boat.

But “well enough” is a moving target. What was sufficient last year might be inadequate next year. The gap between where your company is and where it needs to be widens every quarter that you wait. Putting your head in the sand because you’re profitable is foolish.

Employees of all ages want to work for digitally mature companies. Companies with the cultural characteristics, the research mechanisms, and the opportunity to experiment. Customers expect experiences that match the best consumer apps they use every day. Partners expect integration that actually works.

If you’re profitable today, that’s great. Use that position of strength to build the muscle for adaptation while you still have the resources.

Because by the time the market forces your hand, the cost of catching up will be ten times what the cost of staying current would have been.

The Transformation Illusion

Before we get to the last two myths, we need to name the trap that makes them so persistent.

George Westerman of MIT captured it perfectly:

“When digital transformation is done right, it’s like a caterpillar turning into a butterfly, but when done wrong, all you have is a really fast caterpillar.”

George Westerman, MIT

A fast caterpillar. That’s what most companies are building. And they’re calling it transformation.

Right now, countless organizations are informing their workforce and stakeholders that they are proud to be “transforming” their business with AI. That they are “leaders in their industry.” The reality for many of them is that they are simply making faster caterpillars.

Westerman and others identified three dangerous downsides to this illusion.

First, companies become so busy creating fast caterpillars that they stand still in the real transformation stakes. They’re heads-down implementing AI widgets and chatbots and copilots, and they can’t see the disruption they’re not preparing for. Their efforts are neither defensive nor offensive in their market.

Second, they devote all their limited time, energy, and resources to faster caterpillars because those “change” initiatives have become the priority. There’s no bandwidth left for actual transformation. The busy work of incremental improvement crowds out the hard work of rethinking how the business operates.

Third, and this is the most dangerous: companies are lulled into a false sense of security. They look at all the AI projects on their roadmap and feel good about themselves. Multiple initiatives in flight. But there’s no vision of a butterfly in sight.

John Hagel III at Deloitte observed that most executives he spoke with were still focused on digital as a way to do the same things, just faster and cheaper. He saw little evidence of leaders stepping back to rethink, at a basic level, what business they were actually in.

That’s Substitution on the SAMR scale. That’s performative transformation. And that’s the script AI is following today, almost perfectly.

A snake sheds its skin and it changes. A caterpillar becomes a butterfly and it transforms. The difference matters. Most leaders are confusing the two.

Myth 4: We Need to Disrupt Our Industry Before Someone Else Does

Reality: Snow melts from the edges.

I disagree with Andriole on this one, at least partially.

He argued that disruption almost never comes from market leaders. It comes from startups making bold bets on old industries. Airbnb. Uber. Netflix. Amazon. The evidence is hard to argue with.

But I’ve seen established companies make bold bets too. Apple. Google. Meta. Salesforce. The difference is where those bets originate.

Often, disruption doesn’t start in the executive suite. It starts with individual employees figuring things out in their silos. Those are the canaries in the coal mine. Those are the signals you should be paying attention to.

In every organization I’ve worked with, there are people on the front lines already experimenting with AI. They’re building their own workflows. They’re finding workarounds. They’re solving problems nobody gave them permission to solve. These people have their finger on the pulse of what’s actually happening. They know where the friction is because they live in it every day.

This is what I mean by “snow melts from the edges.” Real change doesn’t start with executive mandates or company-wide rollouts. It starts with one person solving one problem. That insight spreads. The tools spread. Before long, the organization starts to shift from the bottom up.

Meadows described this as self-organization, the ability of a system’s components to create new structures and patterns without top-down direction. The healthiest systems allow this. They create space for experimentation at the edges. They watch for signals. They amplify what works.

If you spend all your time looking at what competitors are doing, you miss the signals coming from inside your own building. If you want to look externally for opportunities, your industry or vertical is a decent place to start. But the real insights come from exploring the edges in completely different disciplines. How are hospitals thinking about this? How are logistics companies? What’s working in education? The patterns that show up in vastly different organizations are the ones worth paying attention to.

The question for leaders is: do you know what your people are already building? And have you created the conditions for those experiments to surface, or are you suppressing them with policies and fear?

Myth 5: Executives Are Hungry for AI Transformation

Reality: They’re hungry to talk about it.

Andriole was blunt about this in 2017: the number of executives who really want to transform their companies is relatively small. He described a wide gap between what executives say about transformation and what they do.

Nothing has changed.

Today, executives talk about AI in their strategic plans. That talk creates pressure. Subordinates are forced to do something performative, something the executives can point to during quarterly business calls. The result is a lot of motion and very little movement.

This is a systems problem, not a people problem. Meadows identified information flow as one of the most important characteristics of a healthy system. When information moves accurately from the edges to the center, the system adapts. When it doesn’t, the system stagnates or makes bad decisions based on bad data.

In most organizations, the information flow around AI is broken. The signals from frontline employees, the canaries who know what’s actually working, never reach the people making strategy decisions. The system punishes honest feedback and rewards telling leadership what they want to hear.

Executives manage up to shareholders the same way individual contributors manage up to their bosses. Saying the words they think people want to hear. Forgetting their duty to dissent. Forgetting to call out when something doesn’t make sense, because that feels risky. This behavior has been taught and reinforced for years.

Kim Scott, a former Google and Apple executive, gave this problem a name in her 2017 book Radical Candor: Be a Kick-Ass Boss Without Losing Your Humanity. People are afraid of radical candour. And without it, the information that would drive real transformation gets filtered, softened, and sanitized until it’s useless.

External consultants can help tease this out. But ultimately, the work has to be done in partnership with the people inside the organization. If you go to one of the big consulting firms, you’re going to get an 800-page report that tells you all the things you could be doing. That’s nice. But the real power is giving the gift of exploration and experimentation to the people within your organization who can actually implement, not just talk about or theorize on the opportunities.

Meadows argued that the most powerful place to intervene in a system is at the level of its goals. If the real, unstated goal of your AI strategy is “say the right words on the quarterly call,” the system will produce exactly that. Words. Not results.

If you want different behavior, change the goal. Make the goal “build the capacity to adapt” instead of “implement AI.” Make it about the muscle, not the tool.

The Gut Check

Before you read another word about AI strategy, answer these three questions honestly.

One. If you asked one of your junior engineers why this product deserves their nights and weekends, what would they say? How close is that to the actual reason you claim drives your technology strategy? In a technology organization, the gap between the engineers’ answer and your stated purpose reveals whether you’re leading a mission-driven team or just shipping code to improve market value.

Two. Open your Jira board, your roadmap, and your operating expenses. Where, line by line, do excellence in user experience, systems that don’t break, and developer growth actually get funded, staffed, and celebrated? In every digital business, the real priorities show up in what gets sprint capacity, headcount, and budget, not in the “transformation” narrative.

Three. Which specific feature or experience in your product is so remarkable that power users already rave about it in public? If you turned it off tonight, who outside your company would immediately complain, loudly? If you can’t name a loudly-missed capability, you are likely running a solid but invisible product that blends into the noise.

If those questions made you uncomfortable, good. That discomfort is information.

The Path Forward

This section could fill 800 pages. It won’t. Because the companies that succeed don’t follow 800-page playbooks. They follow a sequencing principle.

Remember the airline? We didn’t start with electronic boarding passes. We started with showing the availability of flights because that’s the first step in the customer journey. Then we proved we could handle ticket purchases. Then the boarding pass. Then in-flight entertainment on passengers’ personal devices, which moved us from Augmentation into Modification on the SAMR scale. And now the platform extends well beyond travel into services that customers told us they wanted.

At each phase, we solicited feedback from customers and employees. We talked to the teams responsible for building and maintaining the enterprise systems. We explored why previous initiatives had succeeded. Only then did we proceed. No wasted money. No performative checkboxes.

That’s the principle. It’s journey-based. And it applies to AI just as much as it applied to mobile.

Jakob Nielsen, the usability researcher, documented how organizations mature through predictable stages when adopting any new capability. They move from hostility (“we don’t need this”), through isolated experiments by individual advocates, to dedicated budgets, to managed programs, and eventually to a point where the capability is woven into corporate strategy. His research showed that skipping stages creates fragility. Companies that leap from “we don’t need AI” to “AI everywhere” without building the intermediate capabilities will fail. The same way companies that skipped from hostility toward usability to mandating user-centered design, without building the internal muscle to execute, always failed.

Nielsen estimated this progression could take twenty years for usability. The timeline for AI maturity is compressed because the environment demands it. But the principle holds: you cannot skip the hard parts. You can compress them. You cannot eliminate them.

The Sequence That Works

Step One: Describe your system. Map your actual processes. Not the ones in the documentation nobody has updated since 2019. The real ones. How does work actually flow through your organization? Where does information get stuck? Where do people do the same task twice because two systems don’t talk to each other? You can’t improve what you can’t describe. Meadows was adamant about this. If you skip this step, everything that follows is guessing.

Step Two: Find your canaries. Identify the people in your organization who are already experimenting. The ones using AI tools on their own. The ones who have built workarounds because the official systems don’t work. These people have already done the discovery phase for you. Listen to them.

Step Three: Pick one problem. Not ten. One. Choose the problem with the most friction, the best data, and the highest visibility. Solve it. Measure the before and after; hours saved, errors reduced, and capacity created. Make that win visible to the entire organization.

Step Four: Build the feedback loop. Before you move to the next problem, ask the people who lived through the first one: what worked? What didn’t? What would you do differently? Ask the customers affected. Ask the partners affected. This is where most companies fall apart. They skip the feedback and jump to scaling. Don’t.

Step Five: Sequence the journey. Use the SAMR model as your compass. If you just did a Substitution, ask: what would Augmentation look like for this process? What would Modification look like? At each level, the feedback loop repeats. At each level, the organization builds capability. Not just a new tool.

The goal is not “implement AI.” The goal is continuous adaptation to a constantly changing environment. AI is one part of that. The culture, the information flow, the willingness to listen to the edges, the discipline to sequence instead of leapfrog. These matter more than any specific technology.

Westerman’s caterpillar-to-butterfly distinction is the test at every stage. Did you change something, or did you transform something? If you swapped a tool and the work looks the same, you made a faster caterpillar. If you rethought the work because the technology made something previously impossible now possible, you might be building a butterfly.

The Conversation Starts Here

I’ve watched this pattern repeat; mobile, cloud, digital transformation, and now AI. Same myths. Same mistakes. Same gap between what leaders say and what they do.

The reason the pattern persists is systemic. Organizations are systems. They have feedback loops that resist change, information flows that filter honest signals, and goals that are revealed by behavior, not press releases. Until you address the system, no amount of AI spending will produce transformation. Just faster caterpillars.

The environment has already changed. Your customers know it. Your employees know it. Your competitors know it, even if they’re no better at responding than you are.

The work is not easy. It’s not fast. It won’t fit in a slide deck.

Because it’s hard, it’s worth doing.

I’d love to hear where you think your organization sits in this picture. What myth are you living inside? Where are your canaries, and are you listening to them?

The conversation doesn’t end with this paper. It starts.

References

Andriole, S.J. (2017). “Five Myths About Digital Transformation.” MIT Sloan Management Review.

Hagel III, J. Cited in Llewellyn, R. “Transformation Illusions.” Referenced in MIT Sloan Management Review.

Meadows, D.H. (2008). Thinking in Systems: A Primer. Chelsea Green Publishing.

Nielsen, J. “Corporate Usability Maturity: Stages 1-8.” Nielsen Norman Group.

Puentedura, R. (2006). SAMR Model: Substitution, Augmentation, Modification, Redefinition.

Scott, K. (2017). Radical Candor: Be a Kick-Ass Boss Without Losing Your Humanity. St. Martin’s Press.

Westerman, G. “The Transformation Illusion.” MIT Sloan / Digital Business Transformation.

About the Author

Jason Tate has led international teams at Apple, scaled a technology consultancy to $48 million in revenue, and provided AI adoption consulting for over 100 businesses. He has built multiple post-secondary programs at the Southern Alberta Institute of Technology (SAIT) and Mount Royal University from the ground up, designed courses for Alberta Women Entrepreneurs, Women Entrepreneurs of BC and Alberta Catalyzer, and is currently an Adjunct Professor at SAIT and serving as Entrepreneur in Residence at Platform Calgary.

His philosophy is simple: We solve business problems with technology. We’re not technology in search of a problem.

Contact: jt@jasontate.ca | jasontate.ca

Newsletter: From Signal to Scale (weekly at jasontate.ca/blog)

You open your project management tool on a Monday morning. You've got a deadline, three fires, and a client who's been waiting since Friday.

​But before you can do anything, there's a banner. A big one. "Introducing AI Summaries, AI Task Sorting, and AI Workflow Suggestions!" Fourteen new features. A guided tour you can't skip. And your actual work is now buried under three layers of menus that moved since last week.

​You didn't ask for any of this. You just wanted to assign a task.

​If that scenario made your eye twitch, you're not alone. And if you're a product leader who just shipped something like this, we should talk.

​This week, I look at the five-stage feature stuffing playbook, three filters that separate signal from noise, and a framework I first used at Apple that tells you whether a tool is worth the money before you spend it.

Let's break it down.

The Numbers Are Real. The Strategy Isn’t.

Here’s what’s happening. SMBs are adopting AI at a pace that nobody predicted five years ago. Fifty-eight percent already use generative AI. Ninety-six percent plan to adopt it. In Canada, that number is even higher. Microsoft found that 71% of Canadian SMBs are actively using AI and GenAI tools right now.

Those numbers are real. The demand is real.

But here’s where it goes sideways. Tech companies see those adoption stats and hear one thing: “Put AI in everything.” So they do. Every button gets a copilot. Every text field gets a summary. Every dashboard gets a prediction nobody asked for.

IDC calls the winners “companies that focus on pragmatic use cases that are easy to deploy and deliver measurable ROI.” Read that again. Pragmatic. Easy to deploy. Measurable ROI. Not “fourteen features in a press release.”

​The gap between what buyers need and what builders ship is getting wider, not narrower. And both sides are paying for it.

Here’s the thing that makes this moment different from previous technology cycles. When mobile apps were the hot thing, feature bloat was annoying but survivable. You ignored the features you didn’t need. AI features are different. They change your workflow whether you asked for them or not. They rewrite your interface. They insert themselves into processes that were working fine. A bloated mobile app wasted screen space. A bloated AI product wastes your time and, if you’re not careful, your judgment.

The Difference Between Feature Stuffing & Solving a Problem

I teach this concept using a simple test. Take any feature your product ships (or any feature you're evaluating as a buyer) and ask one question:

If I removed the word "AI" from this feature description, would the value still make sense?

If the answer is no, it's probably feature stuffing.

Here's what I mean. "AI task prioritization" sounds great in a press release. But what does it actually do? If the answer is "it sorts your tasks by due date and tags," that's not AI. That's a filter. You dressed up a filter in a lab coat.

Now compare that to a tool that reads your last 90 days of project data, identifies which tasks consistently get pushed, and flags the patterns causing bottlenecks. That's solving a problem. The AI part is secondary. The value is in the pattern recognition you couldn't do manually without a spreadsheet and four hours you don't have.

The difference is simple. One starts with the technology and works backward to find a use. The other starts with a problem and uses whatever tool fits best.

Or as I like to say: "We solve business problems with technology. We are not a technology company in search of a problem.

There's a model in education called SAMR, developed by Dr. Ruben Puentedura. It describes four levels of technology integration: Substitution, Augmentation, Modification, and Redefinition. Most feature stuffing lives at Substitution. You had a filter. Now you have an "AI filter." Same function, fancier label.

The features worth paying attention to live at Modification and Redefinition. They let you do something that genuinely wasn't possible before. Reading 90 days of project patterns to surface bottlenecks you didn't know existed? That's modification. Connecting customer support data to product development priorities in real time so you can fix the issue before the next ten tickets come in? That's redefinition.

The SAMR test is useful because it forces honesty. If a feature is just doing the old thing with new packaging, it doesn't matter how much AI is under the hood. It's substitution. And substitution is where digital transformation goes to die.

The Feature Stuffing Playbook (and Why It Fails)

Feature stuffing follows a predictable pattern. If you’re building products, check yourself against this list. If you’re buying them, use it as a filter.

Stage 1: The Announcement. A competitor ships an AI feature. Doesn’t matter if it works. Doesn’t matter if customers wanted it. It’s in TechCrunch. The board asks why you don’t have one.

Stage 2: The Sprint. Product team gets six weeks to “add AI.” Nobody stops to ask what problem they’re solving. The brief is the competitor’s press release, not a customer pain point.

Stage 3: The Launch. Fourteen features ship at once. There’s a banner. There’s a webinar. There’s a blog post drowning in adjectives that make a sort filter sound like a research lab. The marketing team is thrilled.

Stage 4: The Silence. Adoption is 3%. Support tickets go up 40% because the UI changed. Power users are angry. New users are confused. The AI features sit untouched while the filter sort that worked fine before is now three clicks deeper.

Stage 5: The Pivot. Six months later, they quietly remove half the features and call it “simplification.” Nobody talks about the sprint. The cycle starts again when the next competitor ships something.

Techaisle flagged this pattern directly. SMBs are now drowning in what they call “point solution sprawl.” Too many tools doing too many things, none of them connected, most of them half-finished. The response from the companies getting this right? Fewer tools. Better data. Systems that actually complete a workflow end to end instead of eight half-baked features that each do 20% of the job.

Three Questions That Separate Signal from Noise

Whether you’re building features or buying software, these three filters will save you time and money.

​

1. Does this move a line on my P&L within 90 days?

If a feature can't be connected to revenue, cost, or capacity within a quarter, it's a nice-to-have. Nice-to-haves are fine. But they're not your priority, and they're definitely not worth disrupting your workflow for.

Alex Bratton figured this out during the mobile app craze. I worked with Alex during my time at Apple, and his framework stuck with me. In Billion Dollar Apps, he built something called Return on App (ROA) that forces you to calculate two things before spending a dollar: the New Revenue Potential (how much more revenue your team can generate with the freed-up time) and the Cost Savings Potential (how much you'll save by cutting process time and labor). Those two numbers, plus the organizational side effects, give you a real return calculation before you write the check.

The math works just as well for AI tools as it did for mobile apps. I broke down the full ROA framework in this week's Deep Dive, including the formulas, a worked example, and the five mistakes that blow up the calculation. Deep Dive: How to Tell If a Feature Is Worth Your Money (Before You Spend It)​

​For builders: if you can't articulate the P&L impact in one sentence, the feature isn't ready to ship. For buyers: if the vendor can't tell you specifically what changes in your business after 90 days, keep walking.

2. Does this complete a workflow, or does it add a step?

The OnDeck and Ocrolus research on SMBs entering 2026 made this point clearly. The businesses getting results from AI are the ones using it to improve real outcomes, things like cash-flow visibility, decision speed, and lending accuracy. Complete workflows. Not half-built processes that still require you to copy-paste between three tabs.

A feature that saves you from opening a second application is valuable. A feature that requires you to open a second application to verify its output is a cost wearing a benefit’s clothing.

​3. Would this exist if AI didn’t?

This is the gut check. Some features exist because they genuinely solve a problem better than the previous approach. Others exist because someone needed to check the “AI” box on a product roadmap.

If the underlying need would still exist without AI, and AI just makes the solution faster or more accurate, you’re looking at real value. If the feature only makes sense because AI makes it technically possible, but nobody was asking for it before, that’s a solution looking for a problem.

Think about it this way. Businesses have always needed to categorize expenses, route invoices, and flag anomalies in spending. AI does all of that faster and with fewer errors than a human scanning a spreadsheet at 4pm on a Friday. The need existed before the tool. AI just made the execution better.

Now contrast that with “AI-generated mood analysis of your Slack channels.” Was anyone struggling to figure out team morale before this existed? No. They talked to their people. This is a feature built because the technology made it possible, not because the problem demanded it.

The Mistakes Both Sides Make

Builders: Shipping for the press release, not the user.

Your AI feature doesn’t need to be impressive. It needs to be useful. A tool that auto-categorizes expense receipts with 95% accuracy and saves an office manager four hours a week will never make TechCrunch. But that office manager will never leave your platform. That’s the feature that builds a business.

IDC’s research is clear on this: successful implementations start with honest assessments of infrastructure, skills, and current gaps. Then they focus on a few high-value workflows. Not a buffet. A few things done well.

Buyers: Confusing “more features” with “better product.”

When a vendor shows you a demo with forty features, your instinct might be to feel like you’re getting more value. You’re not. You’re getting more complexity. More training. More things that can break. More surface area for something to go wrong.

The better question when evaluating any tool: how many of these features will my team actually use in the first 60 days? If the answer is three, buy the product that does those three things well. Skip the one that does forty things you’ll never touch.

Both: Ignoring the switching cost.

Every new feature changes the product. For builders, that means every feature ships with hidden costs: documentation, support load, UI complexity, performance impact. For buyers, every update means retraining, workflow disruption, and the very real chance that the thing you loved about the product just got buried under something you didn’t ask for.

The best software companies I’ve worked with understand this. They ship less, not more. They run honest assessments before adding anything. And they ask the uncomfortable question before every release: are we building this because our customers need it, or because our competitors have it?

Where to Start

If you’re buying software right now, here’s a simple filter you can use this week.

Take every tool you’re currently paying for. For each one, write down the one thing it does that you can’t live without. Just one. If you can’t name it, that’s a problem. If you can, you’ve just identified what you’re actually paying for. Everything else is noise.

Then look at any new tools you’re evaluating. Apply the same test. What’s the one thing? If the sales pitch focuses on a list of features instead of one clear problem solved, you’re looking at feature stuffing.

For builders, the filter is even simpler. Before any feature gets on the roadmap, require a one-sentence answer to this: “What does this let a customer stop doing?” If the answer is vague, the feature isn’t ready.

Fewer tools. Fewer features. More outcomes. That’s not a slogan. It’s math.

Deep Dive

Want the actual math? I broke down Bratton's full ROA framework, including the formulas, a worked example, and the five mistakes that blow up the calculation, in this week's Deep Dive.

I WANNA GO DEEP!

Thanks for reading!

If this saved you from one bad software purchase or one unnecessary feature sprint, it did its job.

I'd love to hear where you're at. Got a feature-stuffing horror story? I want to hear it. Hit reply.

See you next Friday.

Best,

JT

P.S. — If your project management tool just added a feature that predicts your mood based on how aggressively you click, we need to talk.

Over the past several years, I’ve had the privilege of working alongside hundreds of women through AWE, WeBC, in my courses, and the classrooms I’ve had the honour of teaching in.

And I’ll be honest. They’ve taught me more than I’ve taught them.

I’ve watched women walk into a room doubting themselves and leave having built something they didn’t think was possible. I’ve watched founders wrestle with hard decisions and make the right ones. I’ve watched students ask the questions that made me rethink things I thought I already knew.

The women who shaped my career didn’t do it with grand gestures. They did it by showing up, doing the work, and refusing to accept “that’s just how it’s done” as a good enough answer.

My mom. My daughter. My partner Colleen. Colleagues who called me out when I needed it and backed me up when it mattered.

I carry all of them into every room I walk into.

To every woman I’ve had the chance to work with, teach alongside, or learn from. Thank you. What you’re building matters. The standard you hold is raising the bar for everyone around you.

Go do something remarkable today.

No three signals. Instead, I want to talk about one thing that's been showing up in every classroom, every consulting engagement, and every honest conversation I've had in the last month.

AI fatigue.

Not "I'm tired of hearing about AI." The other kind. The kind where you shipped more work last quarter than any quarter in your career and somehow feel worse. The kind where the tool that was supposed to save you time has consumed your entire day. The kind where you sit at your desk at 11pm reviewing AI-generated content and wonder what happened to the afternoon.

I've been circling this topic for weeks. A brilliant piece by engineer Siddhant Khare finally put language to what I've been watching play out in real time. I've been ranting about it to my colleagues. And I think it's time we had this conversation out in the open.

Let's break it down.

The Buffer Zone Is Gone

Richard Banfield at Second Harvest gave me a way to frame this that I haven't been able to shake. He said, "Technology shrinks the time between what you want to happen, to it actually happening."

He's not talking about hours getting shorter. He's talking about our subjective experience of time. The distance between desire and fulfillment is collapsing.

Think about it. Historically, someone had to invent the technology. Then people had to adopt it. That adoption period, the messy middle where we figured out how a new tool fit into our lives, gave us time to adapt. By the time the next wave hit, we felt ready. Or at least ready enough.

That buffer is gone.

The invention-to-adoption cycle used to take years. Now it takes weeks. Sometimes days. Claude ships sub-agents, then skills, then an Agent SDK. OpenAI launches Codex CLI, then a model that helped code itself. New agent frameworks show up weekly. And somewhere in the middle of all this, someone on LinkedIn posts that if you're not orchestrating AI agent swarms, you're already obsolete.

Our nervous systems were not built for this.

The Paradox Nobody Warned You About

Here's what's actually happening on the ground, and I see this with my students and my consulting clients, not just in articles.

AI genuinely makes individual tasks faster. That's not a lie. What used to take three hours now takes 45 minutes. Drafting a document, building a workflow, researching an unfamiliar topic. All faster.

But your days got harder. Not easier.

When each task takes less time, you don't do fewer tasks. You do more. Your capacity appears to expand, so the work expands to fill it. Your boss sees you shipping faster, so expectations adjust. You see yourself shipping faster, so your own expectations adjust. The baseline moves.

Before AI, you might spend a full day on one problem. You'd sketch on paper, go for a walk, come back with clarity. One problem. One day. Deep focus.

Now you might touch six different problems in a day. Each one "only takes an hour with AI." But context-switching between six problems is brutally expensive for the human brain. The AI doesn't get tired between problems. You do.

Here's the part that really matters for anyone running a business: AI reduces the cost of production but increases the cost of coordination, review, and decision-making. And those costs fall entirely on the human.

You became a reviewer. A judge. A quality inspector on an assembly line that never stops. Creating is energizing. Reviewing is draining. And most of us didn't sign up to be full-time editors of machine output.

Faster Is Not Better

This is where I need to get something off my chest, because I think we've lost the thread on this one.

Somewhere along the way, we started to conflate "faster" with "better." But pace and efficiency are different levers, and good operators know when to push each one.

Efficiency is about getting to the outcome you care about while wasting as little time, energy, or resources as possible. When you consistently produce more value from the same (or fewer) inputs, that's efficiency. Not speed.

AI does not equal better.

The key is to use technology to produce more value and provide the time to pursue your gifts. Or, as Dan Martel would say, "Buy back your time."

But that only works if you actually buy it back. If every minute AI saves you gets immediately filled with more AI-generated work that needs reviewing, you haven't bought anything. You've traded one hamster wheel for a faster one.

I say this to every cohort I teach: find the efficiencies that allow you to pursue your gifts, and give those gifts to the world. If AI is just making you a faster hamster, something is broken.

What Actually Helps

I'm not going to pretend I have this figured out. I spend 15 to 20 hours a week trying to stay current so I can produce quality material for my students and consulting clients. The FOMO treadmill is real, and I've ridden it hard.

But between Khare's experience, the practitioners I talk to, and my own trial and error, some patterns are emerging that actually work.

First, before you open a single AI tool, write one sentence about what you're actually solving. One sentence. That's it. If you can't explain the problem in a single line, AI is going to generate noise. And you're going to spend your evening editing that noise instead of living your life.

Second, give yourself a three-shot rule. If AI doesn't get you to 70% usable in three prompts, close it and write the thing yourself. I've watched people burn 45 minutes refining a prompt for something they could've written in 20. That's not productivity. That's a slot machine.

Third, and this one is going to sound old-fashioned, separate your thinking time from your AI time. I do my best thinking in the morning with a voice recorder and a coffee. No tools. No prompts. Just me talking through the problem out loud. It feels slow. It is slow. But when I sit down with AI in the afternoon, I'm sharper. I can actually judge whether the output is good because my own reasoning showed up to work first.

Fourth, set a hard stop. A "last prompt" time, the same way you'd set a "last email" time. I've talked to people who are prompting at midnight, 2am, "just one more." I have a rule I stole from my years at Apple: the best time to fix the roof is when the sun is shining. Set the boundary before you need it. Because the most productive thing you can do at 11pm is sleep. The work will be there tomorrow. Your brain won't be if you don't let it rest.

And fifth, the one that matters most: ask yourself whether you're doing more things or the right things. Are you actually closer to your goal? Or did you just ship a bunch of stuff that felt good in the moment? If you can't answer that honestly, you're busy. Not productive. There's a difference, and AI has made it dangerously easy to confuse the two.

The Part Nobody Wants to Talk About

One more thing, and I'll keep this brief because it deserves its own conversation.

Many of us have our identities wrapped up in our careers. I know I do, despite my best efforts. The impact this has on our psyche when the ground keeps shifting underneath us is real. It's not weakness. It's human.

Whole industries are about to get reshuffled. Most of the jobs we recognize today won't exist in ten years. New roles will pop up, burn hot for a while, and vanish just as fast. If we see our identity as "marketer" or "developer" or "consultant" or "accountant," we're going to experience a lot of suffering in this new reality.

The skill worth building isn't mastering the latest AI tool. It's building an identity that strengthens under the pressure of constant change and chaos. That's a different kind of work entirely.

The Real Skill

Khare put it better than I could:

“The real skill of the AI era isn't prompt engineering. It isn't knowing which model to use. It isn't having the perfect workflow. It's knowing when to stop.

We design our systems for sustainability. Circuit breakers. Backpressure. Graceful degradation. We should do the same for ourselves.

AI is the most powerful set of tools I've ever worked with. It's also the most draining. Both things are true. The people who thrive won't be the ones who use AI the most. They'll be the ones who use it wisely.

If you're tired, it's not because you're doing it wrong. It's because this is genuinely hard. The tools are new, the patterns are still forming, and too many people are pretending that more output equals more value.

It doesn't. Sustainable output does.

Take care of your brain. It's the only one you've got.”

Deep Dive

No deep dive this week, that was deep enough.

The signals will keep moving. I'll be back next week with a full Deep Dive.

Thanks for reading!

This one was personal, and I suspect it hit close to home for some of you too.

I'd love to hear how you're managing this. What boundaries have you set? What's working? What's not? Reply and tell me. The best ideas I've found on this topic have come from real conversations, not articles.

See you next Friday.

Best,

JT

P.S. — AI writes v1. But you decide if v1 should exist at all. That distinction is worth more than any prompting course you'll ever take.

Anthropic turned Claude Cowork into a full enterprise platform with department-specific agents for finance, HR, legal, and engineering—connected to Gmail, DocuSign, and your internal systems.

Perplexity launched Computer, an autonomous digital worker that coordinates 19 specialized models to run end-to-end workflows across your browser and SaaS tools.

Google shipped Nano Banana 2, an image model fast enough and accurate enough to produce marketing-ready visuals, infographics, and localized creative directly from a prompt.

The pattern? AI stopped being a tool you talk to and started being a worker that delivers finished output.

Let's break it down.

Signal:

Signal One: Cut Department-Level Busywork — Anthropic’s Enterprise Agents

Anthropic launched an enterprise agents program that turns Claude Cowork into a managed, department-specific productivity platform. Companies can now deploy AI agents with stock plugins for finance (market research, competitive analysis, financial modeling), HR (job descriptions, onboarding materials, offer letters), legal, engineering, and design—all customizable to internal workflows. New enterprise connectors for Gmail, DocuSign, FactSet, and Clay let agents pull live data from existing systems, while IT gets centralized controls: private plugin marketplaces, managed data flows, and per-org customization. Anthropic says 80% of its business is already enterprise, and this positions Claude as a direct competitor to the SaaS tools currently handling these tasks.

Signal Two: Stop Toggling Between Tabs — Perplexity Computer

Perplexity launched Computer, a general-purpose digital worker that takes natural language instructions and autonomously executes complex, long-running workflows across the same interfaces humans use—browsers, Gmail, Slack, Notion, Calendar, and common SaaS tools. Under the hood, it deploys subagents coordinating 19 specialized models to handle parallel research, browser automation, content creation, and tool integration in the background. It can build production-ready apps, websites, and reports, monitor tasks continuously, and iteratively refine its own output. Currently available on desktop for Perplexity Max subscribers, with Enterprise, Pro, and mobile coming soon.

Signal Three: Produce Marketing-Ready Visuals in Minutes — Google’s Nano Banana 2

​Google released Nano Banana 2, a Gemini 3.1 Flash image model that matches Nano Banana Pro’s quality at significantly faster generation speeds. It renders accurate text inside images (marketing mockups, greeting cards, localized creative), generates infographics and data visualizations from notes, and maintains consistency across up to five characters and 14 objects in a single workflow—at resolutions up to 4K. It’s rolling out as the default image model across the Gemini app, Google Ads, AI Studio, Vertex AI, and Flow. Google is also pairing SynthID watermarking with C2PA Content Credentials for provenance tracking, so teams can verify how AI was used in any generated asset.

Scale:

Scale One: Cut Department-Level Busywork — Anthropic’s Enterprise Agents

Start here: Pick one department where a recurring, document-heavy workflow is already well-defined—offer letter generation in HR, competitive research briefs in finance, or onboarding packet assembly. Deploy the stock plugin for that function with read-only connector access to the relevant systems (Gmail, DocuSign, FactSet). Have the person who currently owns that workflow review AI-generated outputs side-by-side against their manual process for two weeks. Don’t roll out across departments simultaneously—prove it works in one team first, then use that team’s results to build the business case for the next. Restrict agents to read-only access on connected systems until the pilot team confirms output quality. Keep human review on anything client-facing or legally binding. Track time-to-completion per task, number of manual edits required, and pilot team confidence for 30 days before expanding to additional departments or granting write access.

Scale Two: Stop Toggling Between Tabs — Perplexity Computer

Start here: Identify one multi-step workflow your team currently handles manually across three or more tools—weekly competitive monitoring, prospect research that feeds into a CRM, or content briefs that pull from multiple sources. Give Computer the task in natural language and compare its end-to-end output against your current process. Start with workflows where the output is internal (reports, briefs, summaries) rather than anything published or sent externally. Limit initial use to internal-facing deliverables. Review every output before it touches a client, prospect, or public channel. Keep a human in the loop on anything that triggers a notification or message to someone outside your team. Track total workflow time (end to end, not just AI execution), output accuracy, and number of manual corrections for 30 days. If quality holds and time drops, expand to the next multi-tool workflow.

Scale Three: Produce Marketing-Ready Visuals in Minutes — Google’s Nano Banana 2

Start here: Pick one repeatable creative task where your team currently waits on design—social media graphics, internal presentation visuals, or product mockup variations. Generate a batch using Nano Banana 2 and compare against your current process for speed, quality, and accuracy of any embedded text. Start with internal creative (decks, internal comms, concept mockups) before moving to customer-facing assets. Require human review on all generated visuals before external use. Verify text rendering accuracy on every output—especially for localized or translated creative. Use C2PA Content Credentials to maintain provenance tracking on anything published. Track production time per asset, revision cycles, text accuracy rate, and designer time freed up for higher-value work over 30 days before replacing any step in your external creative pipeline.

Deep Dive:

No deep dive this week. I'm stepping away from the screen for a few days to do something radical, and go outside. Real sunlight. Real vitamin D. The kind you can't get from a monitor, no matter how good your display settings are.

The signals will keep moving. I'll be back next week with a full Deep Dive. In the meantime, the Scale sections above have enough to keep you busy, pick one and start small.

Thanks for reading!

If any of these three signals hit close to home, reply and tell me which one—and whether you're testing it, planning it, or still deciding if it's real.

See you next Friday. I'll be the one with the tan.

Best,

JT

P.S. — Three different companies, three different bets on the same idea: AI that does the work, not just talks about it. Pick one workflow where the output is already well-defined and the data is already digital. That's your starting point. I'll bring the deep dive next week.

Anthropic shipped a new flagship AI model that can hold an entire codebase in its head. OpenAI released a coding agent that owns tasks end-to-end instead of just writing snippets. And Microsoft updated Power Platform to put AI copilots and agents directly inside the business apps your team already uses.

The pattern? AI just moved from "helpful assistant" to "semi-autonomous coworker." The companies that figure out where to let it own whole workflows, not just answer questions, will pull ahead fast.

Let's break it down.

Signal:

Signal One: Anthropic's Opus 4.6 Lets You Feed It an Entire Codebase and Get Useful Answers Back.

Anthropic released Claude Opus 4.6 with a 1 million token context window (in beta). That means it can ingest an entire codebase, a multi-year regulatory filing, or a full patent family, and reason across all of it without you chopping it into pieces first. It also ships with better agent capabilities: it can split work into subtasks, run tools in parallel, and keep multi-step workflows moving with less hand-holding. On a 1M-token retrieval benchmark, it scores 76% vs. ~18.5% for the previous version, meaning it actually finds what you need in large documents instead of drifting.

Signal Two: OpenAI's GPT-5.3-Codex Turns "AI for Coding" Into an Actual Software Worker.

OpenAI merged its coding model with its reasoning model into GPT-5.3-Codex. It runs 25% faster than the previous version and uses fewer tokens for the same tasks. The difference from earlier models: this one is built for long-running, tool-using tasks. It can plan, execute, self-check, and keep going across terminals, repos, and environments. OpenAI used it internally to debug its own training runs. It's rolling out inside GitHub Copilot, and early reports say it produces fewer half-baked fixes and handles repo-scale reasoning better, especially for bug-hunting and refactors.

Signal Three: Microsoft Power Platform's February Drop Puts AI Agents Inside Your Business Apps.

Microsoft's February 2026 update pushes Copilot and agents deeper into Power Platform. M365 Copilot chat is now embedded inside model-driven apps (preview), so it can reason over your app data plus docs, email, and collaboration content. A new MCP Server lets agents use app capabilities as tools, starting with data entry from unstructured content into forms. There's a shared feed so humans can supervise, compare, and approve agent actions before they go live. And "Code Apps" hit general availability, meaning dev teams can host React or Vue apps as governed Power Apps assets.

Scale:

Scale One: Anthropic's Opus 4.6 Lets You Feed It an Entire Codebase and Get Useful Answers Back.

Start Here: Pick one recurring research or analysis task where your team currently preps documents for AI. Feed the full document set into Opus 4.6 without chunking. Compare the output quality to your current pipeline. The beta 1M context is available on the Claude Developer Platform. Start with read-only analysis, not anything that writes back to your systems. Have a subject-matter expert compare AI answers to known-good answers on 3-5 test cases before you trust it on new questions. Track prep time eliminated, answer accuracy vs. your current process, and time-to-answer for 30 days. If it's faster and accurate, start moving more document-heavy tasks over.

Scale Two: OpenAI's GPT-5.3-Codex Turns "AI for Coding" Into an Actual Software Worker.

Start Here: Pick 3-5 bugs from your backlog that have clear reproduction steps and existing test coverage. Point GPT-5.3-Codex at them one at a time. Review every PR before merging. Treat this like onboarding a junior developer, you check everything. Only use it on code with strong test suites. No production-critical systems on the first pass. Keep a developer in review mode for every change. Set up CI gates so nothing merges without passing tests. Track time-to-fix per bug (agent vs. human), PR quality (revision rate), and developer time spent reviewing vs. writing. Run this for 2 weeks before expanding to more complex tasks.

Scale Three: Scale Three: Microsoft Power Platform's February Drop Puts AI Agents Inside Your Business Apps.

Start Here: If you're on Microsoft's stack, enable M365 Copilot chat in one model-driven app where your team already works daily. Pick a read-heavy use case first, like answering questions about existing records, not writing new ones. ssign someone to monitor the agent feed daily for the first 30 days. Define which actions the agent can suggest vs. which require human approval. Assume Copilot coverage is patchy right now. Canvas apps and Power Pages don't have this yet. Track how often the team uses Copilot in-app vs. going back to their old method, quality of answers (spot-check 10% weekly), and operational cost of monitoring the agent feed. If the supervision cost is higher than the time saved, narrow the use case.

Deep Dive:

Your AI Agent Has a UX Problem. What Apple's Research on Computer Use Agents Means for Your Business.

Last week I watched a demo where an AI agent booked a flight, reserved a hotel, and rented a car. All by itself. Then it bought the wrong insurance, upgraded to a suite nobody asked for, and sent a confirmation to the wrong contact. The technology worked. The experience around it didn't.

A new Apple/Carnegie Mellon research paper puts structure around what's been obvious to anyone actually deploying these tools: the hard part isn't getting AI to click buttons. It's making sure humans can work with these things without losing control of their own business.

This deep dive breaks down the four areas that matter most, from how you talk to agents, to what they should show you, to where they need to pause and ask permission, and gives you a practical framework for evaluating any agent tool before you hand it to your team.

[GET THE DEEP DIVE]

Thanks for reading!

I'd love to hear which of these three signals hit closest to home for you. Reply and let me know what you're testing, what's working, or what still feels like vendor noise.

See you next Friday.

P.S. Three different companies, three different tools, same lesson: AI that lives inside the work beats AI that lives in a separate tab. Pick one workflow where the data is already digital and the process is already documented. That's your starting point.

It's not about what AI can or can't do anymore. The technology works. The problem is the gap between what leaders want to achieve and what their people can realistically support. Harvard Business Review found that 45% of CEOs believe employees are resistant or hostile to gen AI, and most companies lack both a change management strategy and formal training to close that gap.

So this week, when Anthropic launched Cowork to handle cross-app workflows, OpenAI upgraded Deep Research to clear analysis backlogs, and quietly released Frontier to manage AI agents like digital employees, my question wasn't "Are these tools powerful enough?" It was: "Can your team actually adopt them without breaking?"

The pattern? The bottleneck shifted from technology capability to organizational readiness.

Let's break it down.

Signal:

Signal One: Cowork Handles the Handoffs. Remove the app-switching bottleneck.

Anthropic released Claude Opus 4.6 and a major upgrade to its Claude Cowork agentic AI tool. It reads files, organizes directories, composes documents, and completes complex workflows across multiple business applications. Think daily briefings that pull data from Slack, Notion, and GitHub, or research that turns into PowerPoint presentations and Excel workbooks without you toggling between six tabs.

Signal Two: Deep Research Clears the Analysis Backlog. Faster decision-making.

OpenAI upgraded ChatGPT's Deep Research feature with GPT-5.2, adding selective source control, real-time tracking, editable research plans, and a fullscreen report view. It's positioned as a controllable, business-ready research tool that competes with specialized analysis software.

Signal Three: Frontier Fixes the Deployment Bottleneck. Remove the "one-off " experiments.

OpenAI quietly launched Frontier on February 4, a platform for building, deploying, and managing AI agents as digital employees with identity, permissions, onboarding processes, and performance reviews. Unlike standalone chatbots, Frontier provides shared business context (data warehouses, CRMs, internal apps), parallel agent execution across real workflows, built-in evaluation and improvement tools, and governance controls (identity management, audit logs, access controls).

Scale:

Scale One: Cowork Handles the Handoffs. Remove the app-switching bottleneck

Start Here: Pick one recurring cross-app workflow where the person currently doing it manually is willing to pilot the change. Don't force it on resistant team members first. Run a side-by-side pilot where the team member reviews AI-generated outputs against their manual process, documents what "good" looks like, and defines which elements require human judgment. Grant read-only access to connected applications first. Track time saved, output quality (edits required), and, critically, employee confidence and buy-in for 2-4 weeks before expanding to additional workflows or team members.

Scale Two: Deep Research Clears the Analysis Backlog. Faster decision-making

Start Here: Start with one repeatable research type where an analyst champions the pilot and helps define what "good AI research" looks like versus what requires human verification. Build from buy-in, not mandates. Train analysts on evaluating AI research quality, validating sources, and adding strategic synthesis AI can't infer. Define source filters and report structure upfront, and keep human review for all strategic recommendations. Track research turnaround time, source quality, decision impact, and, critically, analyst confidence in using AI outputs for 30 days before expanding to additional research categories or team members.

Scale Three: Frontier Fixes the Deployment Bottleneck. Remove the "one-off " experiments.

Start Here: Choose one high-volume workflow where the team currently doing the work helps define which tasks agents should handle versus which require human judgment. Involve them in designing the change, not just experiencing it. Run a visible pilot where employees see every agent action. Provide training on reviewing agent outputs and handling exceptions, and commit to redeploying employees to higher-value work as the system scales. Connect agents to one system with read-only access first. Track output quality, cycle time reduction, exception rate, and, critically, employee adoption, confidence, and role evolution for 30-60 days before expanding agent permissions or deploying across additional teams.

Deep Dive:

The tools are ready (see last weeks Deep Dive). Your people aren't. A Kyndryl survey found 45% of CEOs believe their employees are resistant or hostile to gen AI. Meanwhile, 31% of workers admit to actively sabotaging their company's AI initiatives.

This week's Deep Dive digs into why new AI tools keep failing at the organizational level, what three psychological needs drive the resistance, and how companies like Siemens, Dell, and Moderna are building adoption that actually sticks.

Plus: the one conversation, one experiment, and one measurement you can run this week to find out if your team is ready.

[GET THE DEEP DIVE]

Thanks for reading!

My hot take, pick one thing from the signals above and do something about it. Security audit. Workflow documentation. Infrastructure mapping. The work that matters isn’t sexy, but it’s the work that compounds.

See you next Friday!

The pattern? The real work isn't in the shiny capabilities. It's in the boring foundations that let you deploy them safely and systematically.

Let's break it down.

Signal:

Signal One: Ignore the Agent Hype. Fix the Security Holes.

Critical security gaps in AI agent frameworks (like ​Claudebot/Moltbot/OpenClaw​) expose API keys and credentials to attackers, with malicious plugins seeing thousands of downloads. This creates supply-chain attack vectors that show up as breach costs and compliance failures. If you own infrastructure or AI operations, rotate all API keys immediately, move off default ports, and establish plugin governance—start with agents touching customer data or production systems.

Signal Two: Forget Better Prompts. Orchestrate the Workflow.

Anthropic released ​Cowork​ plugins that handle complete workflows like “feature request → spec → stakeholder doc” or “support ticket → triage → response → KB update” by encoding team preferences and integrating tools. This reduces context-switching friction that shows up as inconsistent output quality and slow time-to-draft. If you own product, operations, or support, start with 2–3 high-volume workflows, connect read-only tool access first, and automate the drafting step while keeping human review for edge cases.

Signal Three: Stop Chasing Models. Build the Infrastructure.

Businesses that build AI infrastructure—documented context, digital assets, processes, and standards—let any model handle 80–90% of lead generation, nurturing, and sales work autonomously, while those chasing new models stay stuck in research mode. This removes the “learning every new model” treadmill and unlocks model-agnostic leverage where your role shifts to designing reusable assets. If you own growth, content, or operations, map one revenue-critical workflow this week, document the assets AI needs to execute it, and automate the content-generation step while keeping human editing for brand.

Scale:

Scale One: Ignore the Agent Hype. Fix the Security Holes.

Start Here: Audit your most privileged agents first—those touching customer data, financial systems, or production infrastructure. Implement weekly credential rotation and restrict agents to read-only access until you verify plugin safety. Track credential exposure incidents and time-to-rotate for 30 days before expanding agent permissions.

Scale Two: Forget Better Prompts. Orchestrate the Workflow.

Start Here: Pick 2–3 high-volume workflows where output format is consistent (specs, briefs, responses) and quality standards are documented. Connect tools with read-only access first and keep human review for final outputs, edge cases, and customer-facing content. Track turnaround time, revision count, and stakeholder satisfaction for 4 weeks before removing human review steps.

Scale Three: Stop Chasing Models. Build the Infrastructure.

Start Here: Map one revenue-critical workflow (lead nurture, sales follow-up, content creation) and document the assets AI needs to execute it in your voice. Start with stable, high-volume content generation (emails, briefs, responses) and keep human editing for brand consistency and edge cases. Track output quality, time saved, and pipeline impact (open rate, reply rate, conversion) for 60 days before expanding to additional workflows.

Deep Dive:

This week’s deep dive focuses on Signal Three: AI Infrastructure Beats Model-Chasing.

If you’re tired of testing AI tools that never make it to production, here’s a practical framework for building AI infrastructure that works regardless of which model you use. It includes the four-layer infrastructure model (business context, digital assets, structured processes, and specialized skills), real examples from companies using this approach, and a step-by-step guide to documenting your first revenue-critical workflow.

Best for founders, ops leaders, and growth teams who need to shift from AI research to AI deployment.

[GET THE DEEP DIVE]

Thanks for reading!

My hot take, pick one thing from the signals above and do something about it. Security audit. Workflow documentation. Infrastructure mapping. The work that matters isn’t sexy, but it’s the work that compounds.

See you next Friday!

The rules:

• Collaborative tone. Talk like a teammate, not a robot commander

• Be explicit. Use action verbs, specify quantities, name your audience

• Set boundaries. Constrain length, style, format. Tight boxes = better creativity

• Draft → Plan → Act. Outline first, refine, then produce. Kill the one-shot habit

• Demand structure. Tables, schemas, frameworks. Make outputs usable

• Explain the why. Context, goals, brand values. Help it optimize for what matters

• Control verbosity. "Brief," "expert-level," or "ELI5." Say what depth you want

• Provide scaffolds. Give templates so it fills structure, not invents it

• Use power phrases. "Think step by step," "critique your response," expert personas

• Divide and conquer. Break big asks into subtasks. Blueprint → sections → synthesis

The takeaway: a well-architected prompt beats a clever one-liner every time.

Source: https://www.youtube.com/watch?v=Xob-2a1OnvA

https://substack.com/profile/385653226-jt/note/c-186766177?utm_source=substack&utm_content=first-note-modal

What 570 Experts Predict the Future-of-Work will Look Like

Ah, the crystal ball of the Harvard Business Review strikes again!

We know technology is evolving at a rapid pace—think AI, ML, and the latest app that can order your groceries for you while reciting Shakespeare—workers must continually learn and adapt. It’s not enough to ace your job skills once and call it a day. You’ll need to regularly upgrade your knowledge, like updating your phone’s software, or risk being left behind faster than you can say “deprecated feature.”

Four considerations...

1. Stay Relevant: In a tech-dominated world, skills can become outdated quicker than a new meme. Constant learning ensures you’re not just relevant but at the forefront of industry trends.

2. Competitive Edge: Companies will want employees who are excited to embrace the latest tools and technologies. Being a tech-savvy wizard can set you apart from the crowd, making you the go-to person at the office for anything “techy.”

3. Job Security: A continuously evolving skill set can be your safety net. In a landscape riddled with automation, the more skills you have, the harder it is for a robot to take your job.

4. Innovation Fuel: When you’re constantly learning, you’re primed to innovate. New skills can lead to new ideas, which is where the magic happens. Who knows? You might even stumble upon the next big thing while learning a new skill.

In short, if you want to be more than just a cog in the machine, embrace the lifelong learning journey. Think of it as your personal evolution—a way to keep stepping up to new challenges, not just keeping up but standing out. Because let's face it, being average is just not remarkable!