Lightly edited transcript: names and terms corrected, filler, false starts and crosstalk trimmed.

Francis Wade (FW): [Episode introduction] It’s the JumpLeap Long-Term Strategy Podcast, Episode 38. Your company knows that once a corporate strategy has been developed, it needs to move quickly to get it into the hands of staff. But what does it mean to get it into their hands? Send them a PDF? Have them watch a video, attend a town hall, or sit in on a briefing from their manager right after she’s received hers? This is a massive problem, and here today to apply an AI-driven solution is Amie Devero. Tune in to hear us tackle the issue of strategy activation together, using AI.

I’m Francis Wade, and welcome to the JumpLeap Long-Term Strategy Podcast. We’re joined here by my partner and colleague. We were actually business partners for a long time. Her name is Amie Devero. If you haven’t met her, she’s the founder of Contextuum, and she’s also the owner of Beyond Better Strategy and Coaching. She wrote a fantastic book called Powered by Principle, which covered some of the work we did together a few years ago. Let’s just leave it at that. She’s been a friend of mine since somewhere in the ’90s, and we’ve worked together as much as we possibly can. Amie Devero, welcome to the show.

Amie Devero (AD): Thank you for having me. It’s always awesome to be on your show with you, Francis.

FW: Great. As usual, we’re tackling a difficult problem, which is what we like to do. Folks, if this is your first time visiting the JumpLeap Podcast, you should know this isn’t a regular interview where we ask someone what they put in a book, or for their pearls of wisdom. We’ve picked a problem we’re trying to solve together, and we’re going to spend a good amount of time diagnosing it before we get to solutions. We’re following what Einstein said: if he had an hour to solve a problem, he’d spend 55 minutes thinking about the problem and five minutes thinking about the solution. We’re also hoping that by taking this approach we’ll uncover some serendipity that moves the needle in a way we hadn’t planned. I’ll be using an insight bell, so you’ll hear it when we come to a moment where we’ve discovered something one of us didn’t know up to that point.

So let’s start, as we do, with a story. This story is about Larry. He runs a boutique consulting firm of twelve people with one specialty: post-merger integration for hospitals, and nothing else. The strategy in one line is, “We’re the best in the world at one thing, and we say no to everything else, even great money, because the day we become generalists is the day we’re worth nothing.” Everyone in the company can recite it.

The firm has another rule it’s proud of, and it’s a good one for existing clients: they go the extra mile. You don’t nickel-and-dime your clients. You help them, even with the odd thing a little outside the usual scope. It’s half the reason clients love them.

Fast-forward a bit. Last month Theo, a junior partner, got a call from a hospital whose integration with an acquisition they had just finished. Thrilled with the work, the hospital asked if his team could take on a big project fixing their supply chain costs. Different work. Not integration, not the specialty. And Theo said yes, warmly, on the spot. It’s an existing client, it’s real money, and it’s a relationship he owns. He books it as a win and tells the team. But it’s a direct violation of the company’s strategy. They do post-merger integration, not supply chain management. And this supply chain project will take three consultants for two months, which keeps them from doing other strategic work.

Larry knows Theo isn’t greedy or careless; maybe a bit ambitious. If you ask Theo the strategy, he’ll recite it perfectly, and he’ll tell you honestly that he wasn’t ignoring it. He was doing the other thing the firm believes: take care of your existing clients, even somewhat beyond scope. From where he sat, this wasn’t a strategy moment. It was a client service moment, and the client service rule answered it clearly and cleanly before “we only do one thing” ever popped into his mind. So it’s a juicy problem. Folks, as you listen, imagine yourself in a similar dilemma, where there’s a stated strategy and there are other activities that seem adjacent to it, not necessarily in opposition, and you or your employee or your client has to make a difficult choice. Amie, this is a problem you’ve been dwelling in for some time. How have you seen it manifesting in your work?

AD: There are so many instances of this exact issue. For so many years, you and I and all our fellow strategy consultants have worked under the belief that if you have enough metrics and KPIs and training and sharing and communication, your employees will do what the strategy says. But I started to see this thing when I stopped being just a consultant and implicit coach, and started being a strategy consultant and the coach for the founder, and often the executive team, for years after the strategy was complete. That’s really the only way you could possibly observe it. You can’t observe it unless you’re in the wild. And what I saw was that we would create an amazing strategy, and then my conversations with the founder would very quickly fill up with all the tactical emergencies they were dealing with.

I’m thinking of one client with a digital enterprise product that was a very long, very costly sale. Part of how they had originally positioned the product was as high-touch, and that had been effective for a while. One of the problems with that, of course, is that if your clients expect high touch, you can’t have volume. As AI became more ubiquitous, their product became better able to do things without help, and they started thinking, “We need more volume.” Not only that, they had run out of multinationals to sell to. They needed to sell to smaller, scrappier organizations. So they started modifying the product to make it more self-serve. That made a lot of strategic sense. They had a great retreat that included everybody, since they were still a small, growing startup with a staff of about thirty, and everybody understood: we’re modifying the product, and we’re going to start selling a self-serve product that clients can onboard onto themselves and solve problems with as they go.

Six months after the retreat, I’m talking to the founder, and she says, “My customer success team is driving me crazy. I cannot get them to work on upsells.” I said, “I don’t understand. Why not? We’ve talked about that extensively.” She said, “Because when I look at what they actually spend their time on (I have Granola running and everything’s recorded, so I hear every call or read the transcripts), they’re spending all their time on the same handholding they’ve done with our clients for the last eight years.” Why? Because the clients were trained to expect high touch, and the customer success people don’t want to lose clients, because client retention is a key KPI. As a result, they have no facility for saying to a customer, “That’s a really interesting thing you’re trying to do. Let me send you these new videos and how-tos on doing it yourself. Give it a go, and if you hit a snag, let me know and I’ll help.” Instead, they’re doing the work for the client, exactly as they have for eight years. It’s burning all their time and making it impossible for them to do what would really drive the business, which is to upsell or help the sales team implement new contracts.

So I’m thinking about this one client: how can we help this customer success team understand? And of course I end up talking with the client about tactical fixes. More how-tos, more videos, more this, more that. But as I percolate on the problem and start to see it more and more, I realize that the whole strategy vanishes at a kind of terminal node of its execution: the place in the organizational chain where employees have to make choices and judgment calls about what to prioritize, what to do and how to do it, in a moment when they don’t have time to think and don’t have their manager there telling them what to do. The strategy isn’t in their back pocket the way politicians carry the Constitution in theirs. It’s a memory. It’s a concept. It’s not real in that moment.

FW: And it’s a fleeting concept, or at least a diminishing one.

AD: It’s both, right? Fleeting and diminishing. They’ve had ten, twenty, thirty years of habitual experience doing something. Along comes a presentation in a town hall, or a video, or a PDF they get sent, and it says, “Do this differently.” Just because they’ve been told to do it differently doesn’t mean they have the wherewithal, or the concept, or even the understanding, let alone the will or the willpower. We haven’t even gotten to that; we’re just talking about understanding. And remember, this was a really small organization where everybody was in the room. These people didn’t get it as a PowerPoint. This was the perfect strategic rollout. They were in the room, they understood the insights, the breakthroughs, the decisions, the map, all of it. And they still couldn’t execute against it in a way that was consistent with transitioning the company to a new state of being. Forget about thirty years of doing the same thing; some of them had been doing it for five years, or one year, because they’re relatively young. It’s very hard for people to take a conceptual or even operational idea and turn it into a judgment call.

FW: Yes. It’s easy to see, and to be excited in the moment. That’s the part I specialize in. I’m on the strategy creation side, where it’s all rainbows and fairies and pretty pictures of the future, and everybody leaves excited and hugs me at the end, filled with possibility. Then months, or usually years, later: “Hey, how’d it go?” “Oh, that.” It’s not that they didn’t believe it or didn’t understand the case. But other things get their attention, their time and their energy. It’s a change management problem. Those things have so much inertia and so much continuity that this new thing, the strategy, just doesn’t penetrate their day-to-day activity, processes or patterns in any meaningful way. Is that an accurate way to put it?

AD: Even if you change processes, you still run into this judgment problem. Take Theo in your story. This is the key thing: it’s not just habit. It’s also all the other things that are incentivizing behavior. Theo is juggling these two strategic themes. And it’s an eat-what-you-kill firm, so getting new business is a real feather in his cap as a young consultant. We’ve all been in firms like that. He’s ambitious; he wants to climb the ladder. So bringing in new business builds kudos. And supply chain, post-merger integration, whatever: it’s all a lot of the same stuff. We’re consultants; we know how to do everything, right? So he’s not thinking all that through. And there are a million other things that could be in the mix. He could be competing with another associate for attention. This could be a client they thought they’d never see again, and now he has a way to keep them around. You don’t know how many competing things are in the room. The customer success person’s manager could be dating the high-maintenance guy at the client company, so she can’t pivot him to self-serve without causing a whole mess. But all that’s real.

FW: Right.

AD: That’s what is actually happening, and no amount of KPIs or metrics or even training can anticipate it: the number of different forces working on any one individual as they have to make a decision quickly and act.

FW: Yes. And I think it’s fair to say that in all the sixty-something retreats I’ve been in, the challenge we’re talking about has never explicitly been labeled or discussed. There’s just an ongoing assumption that the communication will take care of itself. Maybe a long time ago it did, when there were fewer distractions, strategy was simpler, and there was no AI. But the prevailing assumption is that it’s just a matter of telling people.

AD: And measuring and incentivizing.

FW: And then measuring and incentivizing. But it’s always assumed those would be easy. No one puts up a hand and says, “Actually, last time we tried this, we didn’t really know how to create this kind of behavior change and induce this kind of in-the-moment judgment,” which is what we really need, isn’t it? Nobody initiates that conversation, partly because I probably don’t allow it, so mea culpa, but also because the sensitivity and awareness aren’t there. And I read a fascinating piece of psychological research: when you think about yourself in the future, it’s like thinking about a third party. The same part of the brain lights up when you think about a stranger as when you think about yourself in the future. So even the team in the room, when they thought about themselves a year out, were in effect thinking about strangers. That’s profound.

AD: And what’s even more interesting about that phenomenon is that it really works for you when you’re trying to change your own behavior in the present, if you can harness it. Like saying to yourself, “Maybe I shouldn’t get drunk tonight, since I have that podcast in the morning. What would future me want to say if she were hungover during that podcast?” I actually had that conversation with myself last night. I wasn’t planning to get wasted; I get migraines from even a sip of alcohol. But I had the thought, “It’s Saturday night; it would be awesome to have a cognac.” Then I thought, “How would future me feel about that tomorrow morning?” And I had to admit she’d be extremely annoyed. So when we harness it, the fact that our future self is a third party feels really compelling. But when we have to predict, the fact that our future self is a third party makes it really hard.

FW: Right, and hard to parse and grok, because the landscape we’d have to encounter would probably stop us in our tracks at the moment of creation. But I know that when we get to solutions, you’re going to suggest ways to think about the future in the present that are more useful than just shrugging our shoulders, which is what usually happens.

AD: I think there are two things. I actually think this was invisible, at least to me, until I saw it. You and I weren’t oblivious, thinking about all the retreats we led together. We definitely understood that a strategy needs legs, that there needs to be a plan to make it real. We weren’t naive or in denial. So it’s not that nobody thinks about what we usually call execution. We do. We think we need a scorecard, KPIs, measurement, training, incentives. None of that is unimportant; it’s really important. It just doesn’t get at this other thing. There’s the expression that culture eats strategy for breakfast. I think the reality is that culture emerges out of strategy, if strategy is alive. So what we’re talking about is really the culture piece. But culture doesn’t come out of nowhere, and it doesn’t come out of the KPIs. It comes out of something else. And what’s interesting (I’ll tell you a little later how I came to see this as clearly as I do now) is that it wasn’t seeing the problem that showed me. It was seeing it solved: not by me or by anything I did, but by encountering organizations that have solved it.

We’ve all studied the same iconic companies. Think about the Four Seasons, as a case study. What’s so amazing about the Four Seasons is that the bellhop knows how to do what the Four Seasons culture, and strategy, says to do. And you know what the turnover is among bellhops. This is not a career, right?

FW: Right.

AD: So how does the Four Seasons integrate its strategy so thoroughly that bellhops who have worked there for a month know they should know the guest’s name, know what room they’re going to, and already know to tell them where the tennis court is because they’re really into tennis? That’s not just operational.

FW: Yes, but there’s another side to this, isn’t there? Every embedded behavior that a new strategy is intended to change was once a strategy at some point. Some companies don’t care about things, and things just drift and solidify into behaviors nobody can trace, because they didn’t come from a strategy. But often, companies that endure had a strategy that worked forty years ago, and the person who set it up created some culture around it. They didn’t call it that, but it worked, which is why the company still exists forty years later.

AD: But that’s my point: that’s why it’s mysterious. There’s this history. Everybody knows what Costco is, and anybody who gets a job at Costco knows what Costco is. Anybody who applies knows what they’re in for. But that didn’t emerge out of nothing. So we know this thing happens. The question was how. How can I get my client’s customer success people to operate like a new cashier at Costco?

FW: Right. Let’s call that the first insight. Because understanding where these things come from gives us some clues about how to repeat the trick, and put in place a strategy that will allow behavior to change the next time. That’s the game we’re in. If we accept that at some point people learned a particular behavior, and now we’re asking them to unlearn it, then we know it was successfully created or changed once. So un-creating it is in the realm of possibility. It’s just that last time, perhaps it wasn’t intentional or conscious. It just became, and nobody remembers or can explain where the becoming came from.

AD: There was some amazing founder who was super charismatic.

FW: Exactly.

AD: What I found so interesting is that I have all these founders with the same category of problem: what I’ve called activation, taking the strategy and transmitting it to the moment of judgment, in the moment. And by the way, this isn’t just an entry-level employee problem. It’s a senior executive problem. I see the head of sales go to the product team and ask them to create a whole new function because it’s going to close a deal, even though we did a retreat a month ago in which we nailed down the product roadmap and said we’re not going in that direction.

FW: Wow.

AD: But he needs to make his number for the quarter.

FW: The temptation to make the number for the quarter is just so strong. There are other parts of the culture, accepted forever, that pull them in that direction. Did they feel the irony of abandoning the strategy to chase the sales figure in that situation?

AD: They struggle, because there are all these different forces. There’s a bunch of investors saying, “Sales, sales, sales. We need revenue, revenue, revenue. Forget the strategy; get us some sales.” So how does the strategy ever survive all the forces pushing at the point of judgment? When you look at something like Costco or the Four Seasons, you discover that it does happen in the wild. You can find instances. They’re just uncommon and mysterious. Something happened fifty years ago, and they’ve propelled it through time, iterated on it and reinforced it. But how do you do that with the strategy you created last weekend?

FW: Right. How do you influence that point of judgment in a non-accidental way?

AD: Or a non-historic way, where you have the fullness of time. The world doesn’t operate like that anymore. Your strategy won’t necessarily last fifty years. You may have to iterate on it next year. So how do you get judgment made strategically right away?

FW: Right, almost instantly. Let’s pivot to solutions, and go to the moment where you saw something: the moment with Trader Joe’s.

AD: Okay. So I’d found this product at Trader Joe’s (you probably just saw my dog poke his nose in): freeze-dried bits of chicken that are fantastic to sprinkle on top of a dog’s food. My dog didn’t like his food at first. It turned out the dog loves this stuff, and the cat loves it too, and now neither will eat unless it has the freeze-dried chicken sprinkled on top. Trader Joe’s was the only place I’d ever seen it, so I’d buy ten bags at a time. One day I went in knowing I had one bag left at home, and I’m wandering around the store looking for it because it’s not where it should be.

As Trader Joe’s staff always do, somebody noticed my befuddlement. A staff member walked up and asked if she could help, and I told her what I was looking for, and that the dog won’t eat and the cat won’t eat and I don’t know what I’m going to do. She went in the back, looked for it, checked the order list, and said, “It’s not scheduled to come in. I don’t know when we’ll get it again. It could be gone forever; it could be six months. I just don’t know.” Then she took out her phone and said, “I searched on Amazon and found a similar product. It’s not ours, but it’s freeze-dried bits of chicken you can sprinkle on a dog’s food.” I’m positive that nobody at Trader Joe’s trained her to tell me how to buy something from their competitor, Amazon. I’m sure of it. But her doing that made me a devoted Trader Joe’s customer for life. That was the moment I said, “Holy cow.” There was no training, no communication, nothing that would have incentivized, measured, or caused an employee to make that particular judgment at that moment with that customer, except a deep understanding of Trader Joe’s strategy. How in the world can we do that? That’s what I took as my challenge. That’s what we need to create.

FW: In the moment she was helping you, did she transmit a sense that she was doing something extraordinary, or did it feel mundane and routine, like “this is what we do”?

AD: It felt like, at Trader Joe’s, they’re going to make sure that if they can possibly get you what you came in for, they will. After that I went and learned a lot about Trader Joe’s strategy. They don’t have a lot of SKUs. They have a limited inventory of white-labeled things, and things are there and then they’re gone, and they’re really gone. I’m sure I wasn’t the first person to walk in completely attached to some product. Often they can tell you it’ll be back in a week, or tomorrow morning, or that they’re only getting a little at a time at opening, which they’ve also told me. But what she understood was that her job was to make sure I got the thing I needed, and also that I knew if it was ever back at Trader Joe’s, that’s where I’d come. In one fell swoop she delivered the entire strategy. I don’t care what it costs on Amazon; I’ll always buy it at Trader Joe’s if I can get it there.

FW: It’s a vivid description of what you call the point of judgment. It would be different if you could tell she was grappling with it, if she’d said, “I really shouldn’t tell you this,” or if there were something about it that suggested she knew it was against policy, or that this was an extraordinary moment, or that she was trying to earn one of those little rewards for good service. Any of that would take away from the fact that they’ve been able to deliver this imperative, whatever you want to call it.

AD: It is. It’s like a strategic imperative. Their job is to wander around the store and find ways to help customers and delight them. I had another incident like this. They also have the very best sardines in water. I don’t know what they feed these sardines; they’re the fattest sardines I’ve ever seen. I love them. I think I eat them four nights a week instead of cooking. All of a sudden I go in to buy my weekly quota and there are no sardines, and this goes on for weeks. Finally somebody tells me, “We’re getting a shipment tomorrow, but it’s small, so you need to come in when we open, because it may be gone by nine in the morning.” So I go in at opening, which is already huge, that they’d tell you that. It’s essentially rationed: one box a day, it comes in at seven, they open at eight, and that’s when you can be sure to get it. Already very helpful. Then I’m standing in line with my two tins of sardines, and the cashier says, “We don’t get very many of these in.” I said, “I know.” He said, “If I were you, I’d take a bunch more.”

FW: Whoa.

AD: He said, “Take as many as you can, because they’re in such short supply and people buy them out so fast that if you don’t grab them, you may not see them for a while.”

FW: She’s not upselling, is she?

AD: This was a guy.

FW: He’s not upselling.

AD: He’s like my buddy. He’s saying, “Seize the moment.”

FW: Yes. But it’s coming from a place that reinforces the strategy, the same as your other encounter. The source seems to be one place, and it shows up in two different points of judgment where they deploy the strategy with you. And they do it without a lot of fuss or forethought, as if “this is what we do here, and this is the obvious thing to do.” In most other stores, the second one would be a cross-sell and you’d feel the pressure. And the first one would be forbidden, or if it were whispered, it would be clear she was breaking a rule and would get fired if she were found out.

AD: Yes, it would be some kind of subversion. Here, three different people each invented something to do in a moment when expressing the strategy called for inventiveness. Each one did something different, so they made it their own. It was internalized. And there was no way to predict that any of those moments would happen: that sardines would be rationed, that a product wouldn’t be there or coming but would exist elsewhere.

FW: Right.

AD: There’s no way you could create a training manual that would anticipate those instances. And that’s how strategic decisions actually show up. They just show up.

FW: All right. Let me ask you to eliminate the obvious explanations. Someone listening, and as strategists who care about implementation (which I imagine is everyone listening to this podcast), would go first to things like, “They have an incentive structure,” or “Their boss debriefs them every day and coaches them on meeting their numbers for the week.” We go to the mechanical answers first: strategy reporting, a balanced scorecard, a KPI, an OKR. But I think you’ve eliminated those as the primary cause. Is that accurate?

AD: There probably are KPIs. There’s probably something about those interactions that gets captured in a KPI. I’m now a regular customer, so if they’re tracking debit card usage, they’re seeing my card with increasing frequency, and I’m sure they’re tracking same-store sales. They’re tracking all of those metrics, but none of those metrics implies these particular actions, and none of these actions was captured by anything. The cashier said, “Don’t worry, I’ll wait; go grab more.” If there were a metric for cashiers, it would probably be something like throughput. As it happened, there was no one waiting, largely because they keep opening new registers when people are waiting. I could give you a million other instances like this. These are just particularly visceral, because we’ve all been in situations where the person helping us is nice and helpful and says, “I’m really sorry, we don’t have it. Try back in a week.” And that’s okay; that’s not bad. In many stores, what these people did would be anti-strategic. But it was entirely Trader Joe’s strategy, every bit of it: noticing me looking befuddled, offering me an alternative, telling me when the box comes in and that it’s all they get, and getting me to grab more so I wouldn’t run out before the next batch.

FW: So having had this epiphany, noticing this behavior and its consistency and connecting it to strategy, is it the case that as you look at other companies, or at the general case, you now can’t unsee it?

AD: Exactly like that.

FW: It’s exactly—

AD: —like that. It’s like the red car phenomenon. You get a red car and suddenly everybody has a red car. Now all I see are instances where it’s present or not, and mostly it isn’t. But I still see moments where it is. I was listening to the radio the other day, and they did a story on Snap-on, the tool company. I’ve seen their tool boxes, but what I didn’t know is that Snap-on’s strategy is focused entirely on professionals, not DIYers, and that they’ve driven that successfully through not just their direct employees but their franchises. Every aspect of their strategy and their people’s performance ties back to “We are for professionals. We are not for home DIYers, period, end of story.” They’re not looking to be in that market. They’re not in Home Depot; they’re not accessible to the public that way. And the employees and franchisees I’ve met make it crystal clear; it permeates every decision, every marketing decision, every selling decision, every philanthropic decision. It’s everywhere. It makes them that much more valuable to professionals, and it has made them a company that has endured a very long time and outperforms on every financial measure we could look at. When this is present, it’s an extraordinarily powerful producer of value.

FW: Mm-hmm. I think that’s our second insight: having it happen consistently is a bit of a miracle. And your argument is that it’s not an accident, and this sent you looking for how to activate strategy in a way that produces these moments consistently. I’m imagining you’re not going to say more town halls.

AD: Yeah.

FW: More PowerPoint. That’s what we need. We really need more videos.

AD: Those are okay. I’m not saying not to do that stuff. That’s all fine. But I don’t think we could have done what I see as the solution until now. I’m not saying this is the only way. I’ve spoken with consultants who do the kind of work we do and who have developed very granular, ongoing work to get the strategy into the hands of every employee and constantly reinforce it. They have constant coaching sessions with managers and directors and direct reports, and they constantly collect feedback from employees about whether they understand how to do this and that. You can do it that way. But, man, that’s hard. That takes a lot of work. It’s so far beyond managing a balanced scorecard and cascading it down to performance reviews, and I’m assuming companies already do all that. Doing what you’d have to do mechanically to achieve this on a replicable basis is more than most companies have the bandwidth for, big and small. Could IBM do it? Yes. IBM does do it. But they’re IBM.

FW: And especially in an age when things change so fast, by the time you’ve presented the PowerPoint for the hundredth time to the same employee, things have moved. Keeping it dynamic is the problem. Do you vary the PowerPoint slightly and present a new version next week, and have them sit through it at scale? Have all 10,000 employees sit in an auditorium while you say, “We made a few changes to the PowerPoint”? Everybody’s dying. There’s no way to use the traditional methods and have them be dynamic. Even with a cascade, the time it takes to cascade from the top through levels of management and supervisors, and have them not just parrot it but actually internalize it, is so long that you’d spend the whole time doing that and nothing else. So I think you got to the point where you said there are better ways than the traditional ones.

AD: Yes. And actually, the origin of this whole idea of contextualizing the company was you. It wasn’t even my insight. It began with you, because you asked me if there was a way I could vibe code something that would let your client at the time communicate the strategy interactively to their team after a retreat. I thought, “That’s an interesting problem,” and I went to work on it. That’s how Contextuum started. You and I had that conversation when NotebookLM had just come out, and suddenly there was a way to take a corpus of information and turn it into something. Claude could suddenly make an artifact that was a little quiz, or NotebookLM could turn it into a podcast or a review sheet. I was already vibe coding because I was trying to make a gardening app. That started me down the road of thinking through how we could use AI to communicate strategy. And once I was in that world, it combined with these observations I’d made, whether at Trader Joe’s or with all my clients where this doesn’t work, and culminated in this idea: what’s really amazing about AI isn’t that it produces interactive things, because we’ve been able to produce interactive things for as long as there’s been digital technology. It’s that it does it on the fly. So once you can figure out how to schematize a strategy, you can ask the AI to do something with it. You’d asked me how you could give your clients an interactive AI tool to present the strategy to their board of directors, I think it was. And I went, “Okay.” Then I started to understand how hard a project this is, not because of the end result but because of the initial problem: how do you schematize a strategy? What does the AI need to understand to do this?

FW: In favor of activation.

AD: Right.

FW: Not tracking, reporting, balanced scorecards or performance management; none of the downstream stuff. The very first contact, which is usually a town hall or something. What I hear you saying is: how do you influence that first contact so that activation occurs right at the beginning?

AD: No, I’m saying something even earlier than that. Assume you’re going to make some kind of interactive artifact, as Claude would call it, to teach the strategy. Assume you already know you’re going to do that. We already know Claude can do that; NotebookLM can do a podcast; there are all these other ways. I’m sure ChatGPT can do it. The first problem is: what is a strategy, such that an AI could learn it well enough to do that? Is it a book? A strategy map? A PowerPoint? If you gave it all of that, would that be enough? What is the schema? If you’re going to do this over and over, you need some kind of format you hand the AI and say, “Here, this is the strategy. Now turn it into an interactive teaching tool.”

FW: I’m going to label that bell number four. Number three was when I discovered that my conversation with you led to all this, because I had no idea. So there’s the job of translating the strategy that comes out of a retreat into an interim format, the schema you speak of. And there’s a specific challenge in doing it in a way that an LLM, or an artifact, or an interactive tool can then use. This is new territory. It’s not going through the PowerPoint or doing a video or the chat with your boss or the other channels of communication. It’s a whole new channel, and in this channel there’s a new demand for strategy as schema, so that an LLM can use it, presumably in some kind of interactive conversation. You have to define some new thing in between.

AD: At that point it could have been any kind of interactive thing. I didn’t know if it was going to be a quiz, a chat conversation or an instruction manual. I really didn’t know, because as soon as I started doing what you’d requested, I hit that wall. And the wall was: what do I give an LLM so it can teach the strategy? The transcript of the retreat? The PowerPoint? And of course I’m already thinking about scaling, because that’s how my mind is wired. Whatever I give it, if I build it for what Francis produces from a strategy, will it work for any other output from a strategy retreat? Because every organization I work with produces something different. It’s a PowerPoint, or a Confluence page, or a shared Miro board, or a—

FW: —a 50-page PDF. I remember us having this conversation, because I was one of the first guinea pigs of Contextuum. I had a document from a strategy retreat, and I figured that was all I needed to input. Then Amie’s process for developing the schema turned out to be far more rigorous than what I had written down. The questions were far deeper, so it took me a good while, because I now had to create content around the original output, content that didn’t exist. The answers to those questions usually do touch on some or most of what’s discussed, but in the conversation you blow past them because there are too many things. You can’t write everything down, and you need to encapsulate it in a way that makes sense at the end and isn’t too long to read or listen to. Nobody’s going to listen to eight or sixteen hours of transcript all over again. And I remember being forced to rethink what a designer, or a protagonist like Larry, would need in order to produce a Trader Joe’s kind of outcome. There’s something more that isn’t conveyed, and can never be conveyed, in PowerPoints, something more profound, and it’s a requirement if you want even the basic conversation. You know what we call Chinese telephone: a CEO announces a strategy at a high level and tries to cascade it through levels of management, and by the time it reaches a supervisor who must teach a frontline employee, it bears no resemblance to what was decided at the top, because it’s been morphed five times into a completely different document. So this is a profound problem, and it’s a problem of great strategy, because great strategy calls for game-changing stuff, but the game-changing stuff has to fit into the old stuff. There has to be a way to have these organized conversations at scale, and it doesn’t sound easy to do manually.

AD: When you first did that strategy setup inside the Contextuum platform, it didn’t yet have the capacity to ingest whatever you gave it. It could ingest a little, but it didn’t really know what to do with it, and the schema wasn’t fully fleshed out. So I’d been forced to create user interfaces where you literally entered the answer to every strategy question. What problem do you solve? Who do you solve it for? How many different things do you solve? What makes you different? Where do you stand in the competitive matrix? I’d used every framework I knew, and some I didn’t know and had to research, everything from Porter to jobs to be done to competitive matrices. They’re all in there. You don’t have to do them all.

What doing all that work on my side created was the bones of the schema. The schema is what, in computing, you’d call a JSON, which is basically a data set. It has about 200 fields, and not every strategy has an answer for every field, which is what you ran into, because you were determined to fill every field. You don’t have to. Now, if you give it a transcript from a strategy retreat, or have a really good AI summarize that transcript well and give it the summary, plus your website and other places where you’ve articulated the strategy in broad, conceptual, more subjective terms, it can extract many of those fields. So you don’t have to go through every screen anymore. But at the end of the day, it doesn’t matter how the information gets in. You still need the schema. The schema is the kernel of everything else that gets delivered.

The most germane piece for this conversation is the employee introduction to the strategy. The platform takes the schema, and the employee is assigned, or however it works at an organizational level, and goes onto the platform and enters their name, job title, job description, level in the organization and some basic information. It’s a one-screen thing that asks five questions: Do you manage anyone? How many people do you manage? That kind of thing. With that information and the schema, it generates an introduction of 30 to 40 interactions that both teaches the strategy and invents real, lifelike scenarios for the employee, each one increasingly challenging to decide. A typical scenario would be exactly the one Theo had. If Theo had done the introduction to the strategy for his firm, he’d get a scenario like: “You just finished work for the municipal hospital in Phoenix. After you finished their integration, you got a call from John, the VP, who wants your firm to take on a big supply chain management project. He trusts you now and thinks you’re just the folks for it. It’s a half-million- or million-dollar project. Do you take it?” It presents four possible answers, each less consistent with the strategy than the last, in random order. “Yes, we take it.” “We take it, but restrict it to one consultant.” And so on. One is the most strategic answer and one is the least. You get it right or wrong, and it tells you why. It does about 30 of those. So by the time employees get through the introduction to the strategy, they’ve shown they can make decisions like the one the Trader Joe’s employee made with me, because they actually get to try it out. How do you take this abstract thing, the strategy, the PowerPoint, and know what to do on a Tuesday at 11, when Mr. K calls from the hospital?

FW: So you’re simulating these strategic dilemmas, which were created because they cut to the heart of what’s in the strategy. And I’m also hearing that they’re customized for the job description, so the AI reads the job description, goes into the strategy, and creates a custom experience where people go through these 30 or 40 little simulations and make strategic choices. And you almost force them to get a bunch of things wrong, because they can’t know what’s in the new strategy.

AD: No, no. It teaches them the strategy first. It starts with five, six or seven instructive interactions, and it shows them sample problems. It’s an introduction to the strategy, so you learn it, then you try it, then you’re tested on it, all in one package. It takes about an hour for an employee to do it well, depending on the employee and the complexity of the strategy. It’s gamified and scored, so they can pass or fail. And it’s the entry point into the entire platform, which also has ways for them to work through real decisions on the fly, if they face really tough choices and can’t decide. But that’s a different part of the platform.

FW: Because in the future, I imagine, you could always go back to them with five new questions.

AD: You could go back to them with a whole new strategy, and they’d get a new introduction. You can do that at any time. Say you change the strategy in a month because something happened technologically and your whole idea has turned out to be a failure, so you create a whole new strategy. Every employee does an introduction, and each one is generated fresh, because remember, LLMs don’t really have memory. Everything is created in the moment.

FW: Right. So this is very different from putting together a quiz in SurveyMonkey, sending it to everyone, and then a year later, with a new strategy, sending a new SurveyMonkey quiz. That’s very static. Collating the answers is impossible, and real-time feedback doesn’t exist. This is at a whole other level, because it’s a simulation led by the AI. Folks listening might be thinking, “Hold on. Can’t I do that on paper?”

AD: Well, yes, if you have a full-time curriculum creator who can do this for every job description, and then create a new one whenever someone fails. Because it’s not a test; it’s an experience. They go through the whole thing, and say they get 45%, which suggests they didn’t try very hard. It says, “Great, you’re learning. Try it again.” They can come back next week and try again, get different scenarios this time, score 60%, and now they’ve passed.

FW: Right. Let me ask you a question, because of my own experience using Contextuum. So far in this conversation we’ve only talked about employees. But after I used Contextuum for the first time, I went back into my regular work with customers and prospects, and I started to see use cases everywhere.

AD: I know, because you keep asking me to build new things.

FW: Because the principle you’ve carved out, the process, the core thinking, applies to a board member who wasn’t at the strategy retreat. It applies to a new executive who just joined, or a new manager. It actually applies to the executives who were at the retreat.

AD: And it does have that. First of all, any employee can use it, and it will make it appropriate for them. It doesn’t have to be a junior employee; it can absolutely handle an executive. But there’s another part of Contextuum. It’s broken out into the employee portal, the strategy owner portal and the executive portal. In the executive portal, there’s a refresh of the strategy, where you get walked through the thinking behind the creation of the strategy. So it introduces a new executive to a sort of facsimile of how the strategy was created: what it was like before, what prompted this, what’s new about it, and then the actual strategy. The new executive learns the strategy and its evolution. And you can do that as a group.

FW: But that’s not a movie, right? It’s schema-informed. I hate to call it a recap, though I’m sure it is a recap in a sense.

AD: It’s slightly more interactive than a recap. One day it’ll be much more than that; today it’s still using the AI we have, but Contextuum will progress with the technology. Right now it is a recap, but it’s interactive, and it can be facilitated, which is great for consultants. You can take it into a room with everybody who was at the retreat and revisit the strategy, refreshing everyone’s understanding and recall of what you did and why. And you can modify things in it as you go, because you realize, “You know what? We were wrong on that bet. That’s not really how it is.”

FW: But the secret sauce, the element you’ve isolated (and I’m going to make this bell number five) is the schema. Because people like me could do the same thing with PowerPoint. We could show the PowerPoints from the retreat and say, “We’re doing a recap,” and everyone is dying: “Don’t we already know this?” Without the schema, I’m just repeating what came out in the report.

AD: I think you’re being a little unfair to yourself. A good facilitator like you could do a really dynamic recall. The problem is that you don’t work in these companies. You’re not there until five years later, when they do it again. I’m not there either. So this is a tool consultants can use, and also one a consultant can recommend, set up in part for the client, and continue to support. And by the way, I’m not committed to this being a standalone tool. Someday I absolutely intend it to be connected to other strategy platforms. It’s never going to measure KPIs, but it has what we call the CEO dashboard, which pulls in, not so much insights employees have, as the complaints and problems employees have, and combines that with observations of the world at large to create insights the CEO might use to iterate the strategy. You can imagine it also pulling in the current balanced scorecard and KPIs, putting it all together and saying, “You’re probably going to see this in your lagging indicators in two quarters, but your pricing is off. You might want to check that out.” All the CEO needs to do is check it out. Maybe it is, maybe it isn’t. But it’s an early warning system.

FW: Right. I think what makes all of that possible is the schema. The schema isn’t replaceable by the PowerPoints, or by my memory, for that matter. If I’m in front of a group, you’re right: I can recreate it. I’d probably end up recreating elements of the schema, because I’d have to, in the moment, seeing where they’re befuddled and trying to refresh them in the right way. But I’m not following a methodology or a standard as I do it. What I found with Contextuum is that it made me think about the output of the retreat in a standardized way that went well beyond what I normally produce, and I generally produce the same kind of report each time. This was far more rigorous, because I’m normally not thinking deeply about their problem once they need to go and speak to all their different constituencies and get all those stakeholders on board. I’m just thinking about the outputs they could understand, because there’s a whole argument that strategy should be simple. I believe it, and I don’t believe it. Someone said something about the simplicity on the other side of complexity.

AD: I do think that’s true.

FW: Oliver Wendell Holmes said it. He didn’t give a fig for the simplicity on this side of complexity; he cared about it when the complexity had been dealt with first. That’s when I care about simplicity too. And I think that’s the catch, because when I did Contextuum, I wondered whether it was over-engineered, whether it was asking too many questions. But by the end I realized that the answers to those questions don’t exist anywhere. And if they don’t exist anywhere, a year later it’s my memory and their memory, and we’re trying to cobble together bits and pieces of PowerPoint. Suddenly we don’t have what we need to refresh the strategy, because the complexity wasn’t handled in the first place. We got to the simplicity first, but we skipped over so many answers that the answers we’re giving each other are 5% or 10% of the picture.

AD: Let me just add a caution. There’s a lot in the Contextuum strategy setup, probably much more than most people would ever do in a strategy retreat. But since the product is meant to live in the organization, what I was thinking as I built it is that they might not do this at a retreat, but the marketing team will do the competitive matrix at some point. Maybe it’s not done at the retreat; it’s done later. Jobs to be done isn’t done at a strategy retreat. Why would it be? The product team will probably do that later. And then all of it gets integrated into the schema. So I don’t think it’s that the retreat doesn’t produce something sufficiently rigorous or robust. It’s that once you start down the road of creating a schema, you realize it’s modular. You can keep adding nuance and complexity, and more and more sides of the prism through which to see how the organization executes against its strategy. Not making that available to the schema seemed insane. It’s there. You don’t have to use it to do anything else in the product. You can give it something as simple as a crappy website like mine, and it will populate the schema and produce employee experiences. They won’t be as nuanced or subtle as the ones you got by filling in every single screen, but they’ll be much, much better than a PowerPoint, a town hall or a manager.

FW: Or a one-page summary. There’s a whole line of thinking that says if your strategy can’t fit on one page, it’s not a strategy. In the context of what we’re talking about, there’s something appealing in that notion, if you just want a summary. But if you want Trader Joe’s behaviors, and competitive analysis, and jobs to be done, and process analysis, and so on, then in an average-sized company there’s no one-page cheat sheet that will get you where you want to go.

AD: It depends what you mean by enough. I don’t believe anyone needs to do every screen of the setup to have a very robust, workable strategy. In fact, Trader Joe’s strategy, as far as I could glean it from public sources, is a sample case in Contextuum, and it uses only about 45 or 50 of the fields in the schema. The truth is that Trader Joe’s strategy isn’t complicated or highly complex. It’s just very, very distinct. And that’s what Contextuum will push you toward: making your strategy distinct enough that you could tell a Trader Joe’s employee from a Whole Foods employee in five minutes.

FW: In a way that, I’d suggest, the one-pager couldn’t. The one-pager isn’t enough to get that job done. The Whole Foods one-pager could look like the Trader Joe’s one-pager.

AD: I don’t think so. They wouldn’t look the same at all. The problem is that the one-pager doesn’t deliver the judgments. I think you could fit a strategy on one page. The problem is that the page, that medium, doesn’t deliver the judgments. It’s the ability of the LLM to generate, in real time, experiences that build the ability to make judgments. It’s not about the complexity of the strategy. If the strategy is differentiating and distinct and can be applied to every job, if it has consequences for every role, then one page can become a schema. The problem is that a page doesn’t do that. An LLM does.

FW: Right. I’m going to capture that as the next bell: the schema isn’t a function of complexity; it’s a function of distinctiveness. And I imagine you’ve engineered Contextuum to have the minimum number of screens needed to capture the distinctiveness, and that if something looks vanilla, it will dig in and say, “Hmm, do you mean this or that?”

AD: It doesn’t do that yet, but it will one day. It does push back now. It’ll say, “That mission sounds like every other dry cleaner’s,” or “That vision isn’t a vision.” “We’re the best car dealer in the world” is not a vision, and it will say so. Eventually it will have the intelligence to prod you in real time. And there are two easy, no-commitment ways for anybody to try it. One is to try the employee experience as an employee of my take on Trader Joe’s, or Zoom, or WeWork, or the Savannah Bananas, which I love. There are about four levels of roles in each company, and you can try the employee experience, see how you do, learn their strategy and answer their questions. The other is to take your own company and answer six questions, and it will schematize your strategy. You won’t see the schema, but those six questions are enough for it to build one, and then it will produce an employee experience for any job title and description you give it. That’s worth trying, because it shows whether you know the answers to those six questions. And it will push back if your answers are generic: “That’s kind of a generic answer. Are you sure? It’s going to affect the employee experience.”

FW: Right. And that’s the distinctiveness. Having used the earlier version of the product, I felt it very distinctly, partly because of all the holes I was trying to fill. It became apparent that the distinctiveness was always there; it just hadn’t been captured or explicitly placed in a framework from which other things could happen. It was just assumed: “Oh, sure, that’s something we talked about.” But that’s not enough.

AD: It might be enough. I don’t know. I think it’s in your nature to do things very thoroughly. You have this engineering brain; you wouldn’t feel you’d done it correctly if you didn’t do every screen. I don’t want people listening to think they have to do every single screen, because you don’t, and the platform tells you that. Fill in what you can, skip what you don’t want to do, give it material, and it will extract everything it can. You’ll get as robust an employee experience as the information you give it. And there are screens it doesn’t use in the employee experience at all. The competitive framework: which of your employees is that relevant to? No one, unless they’re selling against one of those competitors, and then it’ll pull it in. Is it going to pull it in for your cashier? No. The employee experience is tied to the role.

FW: What makes using an LLM or AI so powerful, back to what you said before, is that you can customize the strategy activation per job description. You could even customize it if someone says, “I work in this job now, but I worked in these other jobs before.” It could say, “Okay, let me customize it for you, because you’re at a different level.”

AD: It doesn’t do that yet, but it could.

FW: Okay. But it makes sense, because you can do things that are simply impossible face-to-face at scale, or with a handout, or with a decision tree where you answer one question and get one answer, the kinds of basic things that have been tried in the past. This really does open up a whole new frontier of interactivity. Amie, how can folks get their hands on Contextuum?

AD: Go to contextuum.ai. It’s spelled the way it sounds: contextuum.ai. You’ll learn all about it, and you’ll see a video of me. Then you can do either the six-question Quick Start or one of the demos with those four companies. Understand that these aren’t the companies’ actual strategies; they’re created from public data. I don’t want anyone to think Trader Joe’s is a sponsor or gave me this; it didn’t. But try out the employee experiences. The product is free right now. It’s in beta, and I’m looking for users. I want people to try it. I want to hear what sucked and what was great. I’m building this, and I’m not charging money until I think you should pay for it, which isn’t yet.

FW: Great. So folks, contextuum.ai. Don’t skip away from this podcast without going to check it out in the next few minutes. Thank you for joining us, Amie.

AD: Thank you for having me, Francis. It’s been awesome.

FW: [The episode closes with a trailer for the next episode and an announcement of the Long-Term Strategy Conference, LTSP26.]