Agentic AI Strategy: Retrofit or Reimagine the Work

Enterprise AIAI AdoptionAI Agents
Written By Christina Ellwood

Episode Summary

Executives have goals with concrete numbers against them, and agentic AI keeps putting new possibilities in front of those goals. Blaine Mathieu, founder of The River Group and author of The River Doesn’t Wait, says the fundamental strategic decision is between retrofitting the work you already do and reimagining it. Retrofit takes the goals and processes you have and makes them faster or more efficient, which he calls a valid strategic choice, though he notes it is probably just as easy for your competitors to do the same. Reimagine usually rethinks three things together, the workforce, the workflow and the context underneath, and tends to be more differentiating because it is harder to copy. He is explicit that there is no wrong place on that spectrum. What he insists on is that a position is set at one point in time, so it has to be revisited as the technology moves rather than settled once a year.

Key takeaways

  • Start from a business outcome you own, then ask what agentic AI can do against it. His instruction is to stop asking what we can do with these agents and start with what you need to achieve, then work out which people in the organization help you get the technology there

  • Read that outcome back before you take it any further. He flags this very deliberately as he says it: not an outcome related to agentic AI or to technology, but a business outcome that you own

  • Decide where you sit between retrofitting the work you already do and reimagining it, and decide it for one named outcome rather than for the organization in general. He calls this the fundamental strategic decision executives are faced with making

  • Treat a retrofit as efficiency rather than as differentiation. He is direct that it is probably just as easy for your competitors to do the same kind of retrofit as it is for you, so it is probably not creating anything lasting

  • Budget a reimagine as three changes rather than one. It usually means rethinking the workforce, whether human or agentic or hybrid, redesigning the workflow itself, and then working through the context and some of the governance issues underneath

  • Give an agent the context you would give a person on your team: what you own, what information it can reach, and the rules, responsibilities and softer organizational elements it works inside. He says you can and probably need to do this if you want decisions of a similar quality

  • Put three questions on a standing agenda rather than an annual one. Has the capability crossed from demo-ware to deployable against your work, has anyone reset what good looks like for that outcome, and is anything you assumed no longer true

  • Build the habit of revisiting a decision before the numbers force you to. He names this as the new leadership skill, and says most senior executives are already good deciders but are not practiced at it

About Blaine Mathieu

Blaine Mathieu is the founder of The River Group, which works through agentic AI decisions with the executives who run a business and with the technology leaders who need those decisions made. The firm is deliberately vendor-neutral: no AI platform to recommend, no stake in what a client decides, and no implementation work. He published The River Doesn’t Wait: A Senior Executive Guide to Navigating the Surge of Agentic AI in June 2026, and a second book, on what agentic AI changes about how organizations reach their markets, is due this fall. AI has been the thread through his work for thirty-six years: his first published writing on it appeared in Ray Kurzweil’s The Age of Intelligent Machines in 1990. He was a Gartner analyst, where he published the first global e-commerce forecasts, ran market intelligence at Adobe, and has been a CMO, chief product officer and chief executive across software companies including Corel, Mindjet, GoodData and Vantiq. As chief executive of Pratexo his clients included Ford, SoftBank and ABB. He is running the Retrofit or Reimagine executive workshop with AI Realized in San Francisco on 15 September 2026.

 

In this episode

00:56 The persistent challenge in enterprise AI has been organizational, not technical
02:40 No product to sell: the obstacles, and their root cause
04:03 A speed that dwarfs the cloud and SaaS revolutions
05:49 A business outcome you own, and not an outcome related to agentic AI
07:08 The river, the boat, and your organizational scope of responsibility
08:42 Context engineering, and why that is not what he means
09:20 What you own, what you can access, and the softer organizational elements
09:51 The analogy: context equals your scope of responsibility
11:36 Not a book on AI strategy. A book on corporate strategy
11:52 The more open you are, the less you have to micromanage
12:13 From deterministic instructions to wider transparent context
13:47 The fundamental decision: retrofit the work, or reimagine it
14:14 Retrofit defined: the goals and processes you already have, faster
14:41 Why a retrofit is probably just as easy for your competitors
15:39 What a reimagine changes: workforce, workflow, context and governance
16:07 Harder to copy, and it tends to be more differentiating in the long run
16:43 Why the failed pilots failed: they were thought of as technology initiatives
17:02 Rejecting the premise: corporate strategy in the age of agentic AI
17:21 Start with the business outcome, then ask how to retrofit or reimagine it
18:11 Lost from the get go, and why the ROI does not arrive
19:11 There is no wrong place on that spectrum
19:53 Flooded with information, and the difficulty of filtering the noise
20:52 Why AI Realized wanted a workshop on this, and the obstacle it addresses
24:01 Not one decision: the ability to keep watching
25:05 Setting a position at one point in time, because the water keeps moving
25:25 The three questions: demo-ware to deployable, has good been reset, is an assumption dead
27:20 The uncertainty around us moves 360 degrees, the way weather does
27:59 Take every AI reference out and it is still a method for rapid change
31:02 Operationalizing the watching, and building a new muscle
32:25 Annual reassessment, and the next decade
32:58 The book, the site, the free discussion guide and the monthly briefing
34:43 The leadership skill most valuable in the AI revolution
35:14 Not practiced at revisiting a decision before the numbers force it
35:46 The new skill: watching the ever-evolving context, and why it is uncomfortable
36:17 There is no such thing as an AI strategy
36:42 Sign-off

In Blaine’s words

“The fundamental strategic decision that these organizational leaders and executives need to make or are faced with making is the choice between retrofitting the work they’re doing and reimagining the work they’re doing in this age of agentic AI”

Blaine Mathieu   (13:47)

“The challenge with that choice is it’s probably just as easy for your competitors to do that kind of retrofit as it is for you.”

Blaine Mathieu   (14:41)

“The closer you can get to reimagine in terms of how you want to achieve your goals and outcomes, the more differentiating it tends to be in the long run because it’s harder for competitors to just copy that change.”

Blaine Mathieu   (16:07)

“I wanna make clear first of all that there is no wrong place on that spectrum.”

Blaine Mathieu   (19:11)

“This is not really a book on AI strategy. It’s a book on corporate strategy.”

Blaine Mathieu   (11:36)

“As soon as you think about it as our AI strategy or our agentic AI strategy, I think you’ve lost it from the get go, which is why so many of these pilots and initial programs are not achieving an ROI.”

Blaine Mathieu   (18:11)

“The context you can provide to your AI agents is really equivalent to your scope of responsibility in your organization”

Blaine Mathieu   (09:51)

“They’re not practiced at revisiting a decision before the numbers make them”

Blaine Mathieu   (35:14)

“Let’s say there’s no such thing as an AI strategy. There’s a business outcome that you own, and the medium under it, we call it the river of agentic AI, has just moved.”

Blaine Mathieu   (36:17)

 

Resources

Blaine Mathieu

  • Blaine Mathieu on LinkedIn: Where he posts, and the best place to follow his work with executive teams on agentic AI strategy

  • The River Group: His advisory firm and the book’s site, where the frameworks and processes described in the book are published. He points listeners here at 05:40 and again at 32:58

  • The River Doesn’t Wait: His book, subtitled A Senior Executive Guide to Navigating the Surge of Agentic AI and Pulling Away From Your Competitors. The retrofit to reimagine spectrum, the watching and the context argument all come from it

  • The free discussion guide: The study and workshop guide he mentions at 30:09 as being on the book’s site now, so a team can take it and begin working through the method on their own

  • The monthly briefing: His free monthly read of the AI and agentic AI landscape, described at 32:58 as sorting through the noise down to the content a busy executive might think could be strategic for their organization

The workshop

  • Retrofit or Reimagine: Agentic AI Strategy for Executives: The half day workshop he and AI Realized are running on 15 September 2026 at K&L Gates on the Embarcadero in San Francisco. Executives work one business outcome they own through the method, and a registration covers two seats so a colleague can come

Ideas and terms discussed

  • The retrofit to reimagine spectrum: The strategic decision the episode is named for, and he describes it as a spectrum with retrofit at one end and reimagine at the other. Retrofit takes the goals and processes you already have and makes them faster or more efficient; reimagine rethinks how the work is done at all, across the workforce, the workflow and the context underneath. There is no wrong place on it, and where you sit is set at a point in time rather than chosen once

  • The river and the boat: The image the conversation runs on, and it is built by both of them. His half, at 06:22, is the water underneath the boat you are steering toward a particular outcome; he names the river of agentic AI itself at 24:39 and again in his closing answer at 36:17. Christina works it out further at 07:08 and 07:27: the boat is your organizational scope of responsibility, the speed, currents and obstacles change as you move down the river, and unlike many other change management she and her audience have led in the past, this one has no solid foundation under it, only a river

  • A business outcome that you own: The starting point he substitutes for an AI strategy, and he flags the wording as he says it: not an outcome related to agentic AI or technology, but one the executive is actually accountable for. Everything else in the method is applied against one or a few named outcomes

  • Context as your scope of responsibility: His answer to what context means for an agent, and he separates it from context engineering. Not the dataset you connect the model to, but what you own, what information you can reach, and the rules, responsibilities and softer organizational elements your team works inside. The claim is that you can and probably need to give an agent the same context you already give a human on your team

  • Watching the water: The practice the workshop teaches, and what he wants operationalized. Continuous observation of what the technology can now do, resolved into a decision about whether it means acting now, rather than the long-term planning cycle most leadership teams run, which he says may only reassess strategy annually

  • The three questions: What he works through with executives against one outcome. Has the capability crossed from demo-ware to deployable against your work. Has someone, possibly you, reset what good looks like. Is anything you assumed, about the technology or your organization, no longer true

Named on air

  • Gartner: Named twice for different reasons. Christina names it at 02:15 as where he was an analyst, and he confirms it at 03:14. He then names it at 16:43 as one of the sources reporting the failed AI pilots and initiatives whose framing he disputes

  • Adobe: Named by him at 03:14 as the company whose market intelligence unit he ran for a few years

  • Pratexo: Named by Christina at 02:15 as the company he most recently led, with Ford, SoftBank and ABB among its clients

  • Ray Kurzweil: Named by Christina at 02:15. She credits Blaine with first publishing work on AI in 1990, in Kurzweil’s book

Related AI Realized episodes and events

 

Frequently Asked Questions

  • There is no separate agentic AI strategy, on Blaine Mathieu’s argument: there is a business outcome you own and a technology underneath it that keeps moving, so the strategy is the corporate strategy and agentic AI is the medium it now runs on. Mathieu, founder of The River Group, puts it directly at the end of this conversation: there is no such thing as an AI strategy, there is a business outcome that you own, and the medium under it has just moved. He makes the same point about his own book twice, that it is not a book on AI strategy but a book on corporate strategy in the age of agentic AI. The practical consequence is the starting question. You begin with what you need to achieve, then ask what the technology now makes possible against it.

  • Nobody should own an AI strategy as a separate thing, on Blaine Mathieu’s argument: you start from the business outcome you need to achieve, and you partner with the right people in the organization, the CIO or CTO among them, on how the technology delivers it. Mathieu, founder of The River Group, thinks too many AI and agentic AI projects have been driven from the board down to the CIO or CTO with the words you are in charge of our AI strategy, and he is careful to add that those leaders are critically important to implementing any strategy that technology fundamentally touches. His objection is to the category rather than to the people: as soon as you think about it as our AI strategy, he thinks you have lost it from the get go, which is why so many of these pilots and initial programs are not achieving an ROI.

  • Retrofitting keeps the goals and processes you already have and uses AI to make them faster or more efficient, while reimagining rethinks how the work is done in the first place. Blaine Mathieu of The River Group treats the two as ends of one spectrum rather than as alternatives. A reimagine, on his description, usually changes three things together: the workforce, whether human or agentic or a hybrid, the workflow and the way the work is done, and then the context and governance questions underneath it. He is direct about the trade. A retrofit is probably just as easy for your competitors as it is for you, so it is probably not creating lasting differentiation; the closer you get to reimagine, the more differentiating it tends to be in the long run, because the change is harder for competitors to copy.

  • Either retrofitting or reimagining can be the right choice, because there is no wrong place on the retrofit to reimagine spectrum. Blaine Mathieu of The River Group says he is not at all arguing that reimagine is the right end and retrofit is not, and that it depends on which outcomes you own and what targets you are trying to hit. On his account a near-term, quicker and easier retrofit is sometimes absolutely the right strategic move to make, while you watch the ecosystem around you and wait for the technology to reach the point where a reimagine has a higher probability of paying off. What he does treat as fixed is that the position is set at one point in time rather than decided once, because the water underneath keeps moving.

  • Agentic AI pilots fail to deliver ROI mostly because they are framed as technology initiatives rather than as business outcomes, on Blaine Mathieu’s reading. Mathieu, founder of The River Group, notes that Gartner and others have been reporting the failed pilots, and says the main reason is that they are being primarily thought of as an agentic AI project or an AI project. His fix is a change of starting point: ask what your business outcomes are and what you need to achieve, then ask how to retrofit or reimagine them with these systems, and partner with the right people in the organization to work out how the technology delivers it.

  • For an enterprise AI agent, context means what you own in the organization, what information you have access to, and the rules, responsibilities and softer organizational elements your team works inside, not just the dataset the model is connected to. Blaine Mathieu of The River Group separates this deliberately from context engineering, which he describes as giving the model or the agent the proper data to operate on. One analogy he says he draws in the book is that the context you can provide to your AI agents is really equivalent to your scope of responsibility in your organization, and that you can and probably need to give an agent the same context you already give a human member of your team if you want decisions of a similar quality.

  • A leader can and probably needs to share as much organizational context with an AI system as they already give a human member of their team, if they want decisions of a similar quality, because to some degree the more willing and able you are to share it the more likely the outcome is to stay controlled and governed. Blaine Mathieu, founder of The River Group, acknowledges that leaders are used to being very selective about sharing that context, or what they think of as their scope of control, and says it can be uncomfortable. He grounds it in something older than AI: transparency is what effective leaders have done for eons, and the more open you are the less you have to micromanage, because your team already understands the goals.

  • A company should reassess continuously rather than annually, because the technology underneath the strategy now moves faster than a yearly planning cycle can track. Blaine Mathieu of The River Group calls this operationalizing the watching, understanding what is happening around you and responding closer to real time, and describes it as a new muscle a lot of these organizations need to build. He rules out the overcorrection in the same breath: he is not saying enterprises need to become startups that constantly pivot, which he calls a failure mode for sure. What he does say is that companies only trained to even think about reassessment on an annual basis are, he thinks, going to have some big challenges over the next decade.

  • The ability to revisit a decision before the numbers force you to is the skill Blaine Mathieu names as most valuable. Mathieu, founder of The River Group, says most senior leaders and executives are already really good deciders but are not practiced at revisiting a decision before the numbers make them, and that waiting until the numbers tell you something is probably too late when the underlying change is this fast. The new skill, on his description, is being really good at revisiting decisions by watching and understanding the ever-evolving context. He adds that it is a bit uncomfortable, especially in organizations with very well-structured long-term planning processes.

 

  • [00:56] Christina Ellwood: The persistent challenge in enterprise AI adoption has been organizational and management rather than technology. We’ve heard that from our community since the very first summit, and our guest today puts a sharp point on it. When executives ask why AI investments are not paying off, he doesn’t point to the technology. It’s-- He points to the fact that this powerful new engine has been bolted onto an operating logic that was never built to run at this speed. That really makes it a leadership question rather than a technical one. Who decides what agent’s allowed to do? What is it allowed to see? Where authority sits when there is no human in the loop on every decision. That’s his argument in "The River Doesn’t Wait: A Senior Executive Guide to Navigating the Surge of Agentic AI." And h- he-- and that author, Blaine Mathieu, is here today with us, and he’s the founder of The River Group. Blaine, welcome, and thank you for joining me this morning.

    [01:51] Blaine Mathieu: Thank you, Christina. I’ve been looking forward to this conversation.

    [01:54] Christina Ellwood: I am too, because it’s really been such a joy to work with you on this workshop that we have coming up on September the 15th. And we’ll get to that in just a minute, ’cause I, I think this is really the key element that has for many people been missing. But before we get into it, I do want listeners to have a sense of your vantage point, because it go- it does go back further than most. You were a Gartner analyst and published the first global e-commerce forecasts. You’ve been a CMO, a CPO, and a CEO several times, and most recently when you were at Pratexo, and your clients included companies like Ford and SoftBank and ABB. And you first published work on AI way back in 1990 in Ray Kurzweil’s book. So a couple of decades before the rest of us were paying attention, you were actually wor- working in this particular area. But today, your work is with enterprise leadership teams on decisions related to agentic AI and the decisions that, that agentic AI is really forcing. There’s no product for you to sell today, and you’re not really here to discuss anything about technology solutions or things like that. You’re really here to talk about those obstacles, how to overcome them, and why they’re there in the first place so we can get to the root cause and really address them.

    [03:14] Blaine Mathieu: 100%. To touch on a few points you raised, I’ve been working in the area of corporate strategy and mostly from the startup perspective, engaging with large enterprise clients, although I also was running market intelligence for Adobe for a few years, and as you said, a Gartner analyst working with a lot of large enterprise clients across the world. And what I’ve learned over the last, call it 30 years in, in these kind of interactions and engagements and leading workshops with these kind of clients is fundamentally the challenge is rarely a technology challenge. Usually it boils down to what I’d call more of a, a strategy challenge or le- and/or a leadership challenge, and those are often, very tightly connected, and the underlying technology is just the driver. But what made me write the book and really engage i- with the River group is The technologies behind AI and now in particular agentic AI are moving at a speed that, organizations and enterprises have never experienced before, dwarfing the speed of the cloud revolution or the SaaS revolution. And so leaders and executives need to think about how they’re moving and how they’re thinking about this in a very different way.

    [04:33] Christina Ellwood: Yeah. I do-- I think it’s really amazing to see both how fast the technology is moving, but also how fast the or- our organizations are moving to adopt it. And there’s a lot of things written about how we’re failing or it’s moving slowly or what have you, but in fact, we have made an amazing amount of progress in leveraging this technology. There’s just more possibilities that are unfolding as the technology unfolds, and our organizations aren’t designed to change that quickly. In fact, part of the job of executive leadership is to create consistency and stability and predictability. And so when we start to throw in lots of changes in very narrow windows of time, it really is antithetical to the kind of leadership responsibilities and approaches that we’ve used in the past. So I’m really glad that you have taken this this issue on, and I’m so excited about our September 15th workshop. So let’s talk a little bit about that workshop, why we’re putting it together, and what you will be accomplishing with the executives in the room.

    [05:40] Blaine Mathieu: Absolutely. Fundamentally, the frameworks and and processes described in the book are all on the book’s website, and we’ll talk about that later. But what I’ve seen is, leaders at enterprises and organizations can use some, some help in being led through these and applying them to a real outcome that they own. And so what we’re gonna be doing here in a few weeks is bringing these leaders together, either individually or perhaps with another member of their team, and Related to one particular business outcome that they own, and I say that very specifically, not an outcome related to agentic AI or technology, but a business outcome that they own. We’re gonna work with them to identify how do they watch on a continuous basis the evolution of these technologies and systems that are underlying that outcome, and how do they make sure that they are able to choose the best outcome and the best path there given that constantly changing foundation. I call it the water underneath the boat that they’re steering toward this particular outcome. So doing an appropriate watching of those changes and learning how to do that on a continuous basis and applying that to an actual outcome they really own in the business, I think is the fundamental premise of what we’re gonna be doing in that workshop.

    [07:02] Christina Ellwood: Yeah, I think that’s that’s really helpful when we take things that we are learning conceptually and apply them right away. I, I ... that all by itself I think is really valuable. But this collaborative element I think is as well. In, in your model you talk about how the agentic technology is like the river, and it’s moving very quickly, and our organizational scope of responsibility in our organization if you’re the CEO of the whole organization, but whatever you’re running is like the boat. And what’s happening is the as we’re moving down the river, the speed of the river is changing. The currents are changing. The, the obstacles are changing and so forth. And so unlike many other change management maybe that we have experienced and led in the past, this one has not a solid foundation under it. It has a river under it. It has this moving this moving substrate underneath us, and that is part of the nature of what makes it different. I think we’ve all experienced that. We talk about the whiplash of the y- the change in the the technology and so forth, but I don’t think we’ve really translated that as so crisply as you have in this analogy. Now s- having said that the, there are several other things that you pull out that I think are also very critical to what makes this transformation really different. And one of those is this the w- fact that the context in which we are giving these agents a-agency is part of what makes it important for them to op- operate properly. But you f- highlight how that context is actually derived from the larger context of the responsibility of the executive. So would you talk about that?

    [08:42] Blaine Mathieu: Yeah, it’s a-- context is a really interesting topic these days because from the technology perspective, we’ve been talking about context engineering for some time, right? Which is how do you give the AI model or now the agent the proper data in which to operate on? But in the wider organizational context, if you’re trying context, no pun intended, if you’re trying to navigate this, this boat on this constantly moving water underneath you to the appropriate outcome or to achieve your goal, you have to think of context as more than just how you connect your AI to a particular dataset or a database. It’s really about what do you own in the organization? What information do you have access to? What are the rules and responsibilities and even the a lot of the softer organizational elements that your team exists within, and how can you provide your agentic AI system with that kind of context and so that it can make the appropriate decisions for you and just as a member of your team would, right? A, a human member of your team would. And so one way, one analogy I describe in the book is the context you can provide to your AI agents is really equivalent to your scope of responsibility in your organization. And you sh- and just as you use that overall context to lead human employees, you can and probably need to provide that kind of context to your agentic employees if you want them to be able to make decisions of a similar quality.

    [10:19] Christina Ellwood: Yeah, I think that’s a really important observation that you’ve made there that is a key element for the leader to be revealing, and we don’t normally reveal that, do we?

    [10:29] Blaine Mathieu: No, and it can be uncomfortable because a- as a leader, we often think about, okay, what am I ultimately responsible for? What do I control? What is my overall organizational context? And then we want to be very selective in terms of sharing that, that context, or in some cases you think it a- as your scope of control with others. And in this rapidly changing space of agentic AI, to some degree, the more willing and able you are to share it, the more likely it is that you’re going to achieve your outcomes in a controlled and governed way.

    [11:05] Christina Ellwood: So that’s both sharing information that you might have that you otherwise wouldn’t need, if you will, to share, but also having some transparency about the logic behind what you think that information means. Now, that’s a very different style of leadership than many people use, but it’s not completely unfamiliar. So does it change primarily the approach that the executive is taking or the structure of the way they are actually leading?

    [11:36] Blaine Mathieu: Fundamentally, I think it goes back to the fact that this is not really a book on AI strategy. It’s a book on corporate strategy. And what you said about transparency is exactly what effective leaders do and have done for eons in organizations. The more transparent and open you are with your team, with your organization, the less you have to micromanage what they do every day because they understand the overall context. That’s transparent. They understand the fundamental goals of what you’re trying to achieve. With the use of technology in these organizations, it used to be very deterministic. You just laid out precisely you need to do A, B, and C using this data. But these new agentic system is, and agentic systems and technologies need that kind of wider, transparent context that you as an effective leader were probably giving to your team already. Now you have to think about how you provide that same context to these underlying technical systems.

    [12:37] Christina Ellwood: So when we’re doing strategy work, how does it change the strategy work to be thinking in this way? I understand what you’re saying about the the need to be transparent about the context because if we don’t give ... Th- there’s the, the exception handling and the rationale and the intention and these kinds of things that are related to what we y- we don’t really build into our processes, we expect people to bring to the party. So if we wanna have a, something like an agent bring that to the party instead, then we have to make that clear. I get that whole element of it. I think there’s been a fair amount of communication about that part. But you’re really speaking to something that’s above that So tell me about how you bring that piece out in your work. ’Cause you do consulting work as well as y- you’re an author and you speak do public speaking, and you’re doing this workshop. So you are in the mode of helping executives to connect the dots from what they’ve been doing to what they need to do in this new world, and I’m asking you really to help me understand that in this conversation so that our listeners can start to take action on this even before they get exposed beyond this conversation.

    [13:47] Blaine Mathieu: Let me make it a little bit more concrete and maybe step back a bit. So the fundamental strategic decision that these organizational leaders and executives need to make or are faced with making is the choice between retrofitting the work they’re doing and reimagining the work they’re doing in this age of agentic AI, again, given this constantly moving water and the surges of agentic AI that are happening underneath their organization. So they’re trying to achieve important strategic goals. They’ve got very concrete, numbers and and targets against these goals. And now they’ve got this new set of possibilities in front of them that are constantly changing. And on one end of the spectrum, they can retrofit. They can decide to basically take the goals they’ve already got, take the processes they’re already using, and perhaps speed them up or make them more efficient using these technologies. And that’s a very valid strategic choice to make. The challenge with that choice is it’s probably just as easy for your competitors to do that kind of retrofit as it is for you. So while increasing efficiency and speed is sometimes s- very important to do, and there are lots of good examples of it these days it’s probably not creating lasting differentiation.

    [15:05] Christina Ellwood: Yeah. When we, when, when speakers come and talk about their use cases at our events and so forth, and these amazing things that they’re doing with AI to transform their organizations, they most commonly fall in that category that you’re talking about, where they have made something that they’re already doing better. Sometimes they bring in things though that are really what we would refer to as redesigning.

    [15:27] Blaine Mathieu: Yes. And and I go even maybe a little bit conceptually further than that and refer to it as reimagination So retrofit on the other hand, on the one hand, and then reimagine on the sort of other end of the scale. And that’s when you think, "Okay, let’s not just take what I’m doing and make it more efficient or faster. Let’s absolutely rethink how I’m doing that in the first place." And that engenders usually a rethink in terms of how you’re using the workforce, either human or agentic or hybrid redesigning the workflow, the way that work is done, and then thinking through the context and and s- and some of the governance issues. But if the closer you can get to reimagine in terms of how you want to achieve your goals and outcomes, the more differentiating it tends to be in the long run because it’s harder for competitors to just copy that change.

    [16:21] Christina Ellwood: Do you think maybe one of the reasons that agentic projects are struggling to survive deployment in production is because they’re, the dec- the decision was to maybe be a little too aggressive in the reimagination end of it? Or do you think there’s a different reason that the, these projects are struggling?

    [16:43] Blaine Mathieu: I think the main reason, Gartner and all the rest have been talking about all these failed AI pilots and initiatives is because They’re being primarily thought of as technology initiatives. They’re being thought of as an agentic AI project or an AI project. And, like I said before I fundamentally reject that premise in the book. Again, this, this book is not a book about AI strategy, it’s a book about corporate strategy in the age of agentic AI, right? And so don’t start with a thought process of, "Okay, we’ve got this AI, we’ve got these agents. What can we do?" Start with, "What are my business outcomes?" "What do I need to achieve?" And then, "How can I either retrofit to make them more efficient and effective, or potentially reimagine them using these systems, these technologies?" It’s not a technology challenge, it’s a business challenge. And then make sure that you’re partnering with the right people in the organization to help you figure out how to use the technology to achieve the outcome. I think just as a quick aside, I think too many of these AI and agentic AI projects have been fundamentally driven from the board to the CIO or CTO. You are in charge of our AI strategy, okay? And of course, those leaders are critically important to implementing any strategy that’s fundamentally impacted by technology. But as soon as you think about it as our AI strategy or our agentic AI strategy, I think you’ve lost it from the get go, which is why so many of these pilots and initial programs are not achieving an ROI.

    [18:25] Christina Ellwood: Okay. So that I think is a good way to think about how to build a successful one. So if you say, "Oh, let’s look a little bit at why we- things are failing, and let’s look at how we can actually use this in an effective way." So going back to your spectrum of retrofit to reimagine, and picking up on your thread that the reimagine gives you a more competitive advantage, whereas the retrofit is something your competition could catch up to Do you think the the desire to ch- to approach your outcome in a different way in order to create competitive differentiation is the main reason to choose reimagine, or are there other o- reasons, drivers that would take you down the reimagine path?

    [19:11] Blaine Mathieu: Yeah that’s a great question. So I think you’ve, you definitely-- I wanna make clear first of all that there is no wrong place on that spectrum. I’m not at all arguing that reimagine is the right end of the spectrum and retrofit is not, right? It really depends on, again what outcomes do you own? What targets are you trying to achieve? And sometimes a near-term, quicker and easier retrofit is absolutely the right strategic move to make while you watch the, the, as I describe it, watching the water, w- watching the ecosystem that’s going on around you, waiting for the technology to get to the point where you can drive a higher probability outcome on the reimagine. So there is no wrong answer to where it is on that spectrum. But I do think the challenge a lot of leaders have right now is they are just being flooded with information about AI. The volume of reports and newsletters and everything that comes out is full of so much noise, and it’s a lot of, quasi-technical information that’s actually, that actually doesn’t have a strategic outcome for the organization It’s very hard for a leader to filter that out and really say, "Okay, but what really affects my strategic outcomes and what do I have to do based on that?" And having a clearer view of what’s going on around the organization in terms of the, the water as I describe it, will, I think, in- increase the possibility in the long run that a reimagined outcome is, is embraced and ultimately is successful.

    [20:52] Christina Ellwood: One of the reasons that I w- was so excited about the possibility of offering a workshop on this issue is, as is from the beginning, this issue of organizational management challenges has been on the table. And it’s been the most persistent issue actually from the very earliest days. And we like to focus on the strategies to overcome those obstacles rather than describing them. And this felt like a strategy for overcoming the obstacle. Follow a model that is about making decisions that are going to lead to success and that are going to be business-based, b- rooting in the outcome of the business rather than rooting in the adoption of the technology. So we’re, we’re really excited to see how this helps the executives who participate in this workshop. It’s also practical. That’s another element that we think is really important, is giving people a way to operationalize what they’re doing. And one of the things I like that you’ve done is you have coupled the work of the leader with whoever they need to collaborate with or whoever they need to convince to be a part of it. And have... Are offering that they bring a colleague and you’re doing that under the same ticket, which is a really wonderful way of signaling to the executive, we really want you to bring these people into the room. And then by having a mixed room of executives, you’re also learning from other industries, and that’s a, a element that- people like about our community is that we give them that cross-industry fertilization because they think, look, I know a lot about what’s going on in my industry. I need to know what’s going on in other, in, in other industries." So I think that’s a really well-constructed room, but put some color for us on what will you will do with that room. So we’ve got the ability to collaborate with people from other industries, and we have the ability to work with a peer if we bring one.

    [22:38] Blaine Mathieu: Let me just touch first on, on the idea you bring up of the value of having, more than one person from an organization in the room, because that loops exactly back to the context discussion we started with, right? The closer you get to the reimagine end of the spectrum, it’s likely that you will need a wider context in order to achieve the outcome that you’re shooting toward, obviously. And so that’s why, getting a larger number of stakeholders together across your team or across multiple teams, sometimes even including folks outside the organization, can be really important to having the wide enough context in order to effectively drive that reimagine. So that’s why it definitely makes sense to at least begin by bringing a partner with you into the room, into a workshop like we’re hosting, so you can begin that process of enabling solutions with that wider outcome.

    [23:32] Christina Ellwood: Yeah, clearly if I ... If we wanna do, say re- redesign the outcome f- of our funnel, we wanna have a higher performing funnel, we can’t do that with not having both sales and marketing and even success and support in the r- in the conversation because they’re gonna be impacted by it as well. Yeah. So that makes a lot of sense that you would wanna have more than one player in there. We might need the support of of other groups too that have a little wider reach, like maybe product.

    [23:54] Blaine Mathieu: Absolutely.

    [23:55] Christina Ellwood: Okay. That makes really good sense. So n- now let’s talk about how we would experience that in the room.

    [24:01] Blaine Mathieu: Yeah. Yeah. Let, so let me go back to that. And very closely related to that, I want to make sure people understand something clear. This is not about making one decision and then moving on to execute against that decision. The fundamental thing we’re working through in the workshop and in the framework is about the ability for you and your organization to be continuously watching, as, as I describe it in the book, what’s going on around you, the context in particular, the context of the capabilities of agentic AI in, within an understanding of what your goals are. And so what we’re gonna do in the workshop is really, take a part- one or a few particular outcomes that you’re trying to achieve with your business, and then teach you the methodology for watching for the important moves in the underlying technology, this river of agentic AI, that will allow you to Set your position on the retrofit to reimagine spectrum. And again, that’s setting at this one point in time because the water’s continuously moving, and learning how to navigate that continuously moving water is exactly what we’re doing. And fundamentally, there’s three questions that we’ll be talking through and working through with workshop attendees that are relative to a particular specific outcome that you have. One is, the first is, has capability of AI and agentic AI crossed from demo, demo-ware, to actually deployable against your work and your outcome? How do you know that is happening? The second question is, has someone reset, and maybe that someone is you, reset what good looks like for an outcome? Maybe a goal has got harder to achieve or has been not very often, but sometimes easier to achieve, and how does that reset how you think about applying these technologies and solutions? And then finally, has anything you assumed, either about the technology, your organization is, are any of those assumptions no longer true? Okay? And again, these are not one-time decisions, but these are things that you should be continuously reevaluating against your outcome because, again, the, the technology and underlying river is changing so fast. And we’re gonna show you how to do that, but not just in the abstract. We’re gonna actually do that against one of your particular outcomes in the workshop.

    [26:32] Christina Ellwood: That’s great. So we’re gonna leave with a new skill set really.

    [26:36] Blaine Mathieu: Absolutely.

    [26:37] Christina Ellwood: Yeah. That’s fantastic. O- one of the things that strikes me about this time that is also different besides the rapidly changing technology underneath us is the level of uncertainty around us There-- I’ve never seen a more uncertain time. We have changes at the economic level, at the political level, at the social level, et cetera. At every level, we have very high uncertainty. So the uncertainty of the technology, in many ways, is unidirectional. This makes a river really a great analogy here, right? It doesn’t f- usually flow backwards. It only flows in one direction. The uncertainty around us is is changing i- in more of a 360-degree way, the way weather changes, right? So it strikes me that this approach that you are teaching or imparting on people, this method of thinking about our outcomes, is well-served if you w- just were to isolate it in these u- uncertain times. So yes, it’s driven by the adoption of a, a particular type of or the a possibility of adoption of a particular type of technology, but it seems really well-suited to the way in which our businesses are-- or the con- context, if you will, in which our businesses are operating.

    [27:59] Blaine Mathieu: I c- I couldn’t agree more, and I, I’ll repeat what I said earlier, which is fundamentally, you could remove every reference to AI and agentic AI out of the book, and what you’d have is a methodology for dealing with extremely rapid change in your organization at a speed that we’ve, as you said, have never really had to deal with before. But the reason it fundamentally does relate to AI and agentic AI is that is a technology driver which is driving the kind of organizational and even societal change at a speed that we have never seen before, and we are only at the beginning of it. This is... we’re not even at first base on, on what’s happening right now. So over the next, I’d say, five years we’re gonna have to buckle up and be ready for massive change, and ensuring your team and your organization has the skills to navigate that appropriately is what fundamentally what the book is about, what our, and what our workshop will be about.

    [29:03] Christina Ellwood: I think it’s really exciting to have a approach that is so well thought out and so accessible to people. As you’re working with enterprises, executive teams in your consulting work, tell us a little bit about what you’re seeing in the before and after. If we were to talk with them about the outcomes that they’re experiencing as a result of having worked with you, what would they tell us?

    [29:27] Blaine Mathieu: It’s so interesting, and I’ve been doing this, as we said at the beginning for decades. I’m always working with senior executives, very smart people. They didn’t get to the level they are unless they are good leaders, great operators. That, that’s just a given. And yet I’m continuously surprised Where at the value of having an outside force either a moderator or somebody who can help them put the pieces together and see things that they just can’t see, again, because they’ve got a shared con- back to context, they’ve got a shared context for what they’re trying to do, what they’re trying to achieve. And having a framework and sometimes even a person that helps break through that context is always valuable. I’ve n- I’ve never seen it not be valuable to the point where the team goes, "Wow, I, that, that was incredible." And so I’m certainly trying to do that with the framework in the book itself, break the context and the study guide and the workshop guide is actually on the book’s website now. So a team can take this and begin to start working through it. But this is why I do think the workshop will be will be really effective because we can help help guide them through that process as well, and I’ve only ever seen it be very powerful.

    [30:51] Christina Ellwood: Now, what about the specifics of the model itself? What are people telling you about how the model has helped change their trajectory or the outcomes their organizations are experiencing?

    [31:02] Blaine Mathieu: It’s really about I think one, one of the critical changes is around operationalizing what I call the, the watching, right? And most leadership teams are working through some kind of long-term strategic planning process, right? Which may have, maybe per- perhaps annualized outcomes in terms of changing the direction and strategy and goals of the organization. But like we just discussed, the pace of change is now moving to the to the place where some kind of annualized reassessment of strategy is just not enough if you want to stay ahead and if you want to stay competitive. And so rethinking how that works and operationalizing the ability to understand and respond to the change that’s happening in closer to real time is, I think, a new muscle that a lot of these organizations need to build. And don’t get me wrong, I’m not saying that these enterprises need to become startups and be constantly pivoting. That would be a failure mode for sure. But they need to operationalize at least the, what I call the, the watching The understanding of what’s going on around you, especially again with the push and the surges in, in AI and agentic AI, and then be able to make decisions, "Okay, does this mean we have to do something now or not? And if yes, what is it? Is it a small retrofit of our operations, or is it opening up the possibility of a reimagine?" Companies that are only trained to even think about doing that on an annual basis are, I think going to have some big challenges over the next decade.

    [32:44] Christina Ellwood: That’s really, I think, a valuable insight and a good outcome for people to have a new way that they are being they’re able to respond to these rapidly changing situations. So what resources would you recommend for listeners who wanna learn more?

    [32:58] Blaine Mathieu: First obviously the book itself is a resource, and I’ve mentioned a couple times the website riverdoesntwait.com where the discussion guide is posted and free. There’s also available on the website a free monthly briefing where I read what’s going on in the AI and agentic AI landscape and, I have the unenviable task of sorting through all the noise so I can get just down to the content that a busy executive can actually, potentially think it could be strategic to their organization. Now, of course, in the workshop itself, we’re gonna show you how to begin to do that yourself so you can actually, with a deep understanding of your context, be able to do that, that noise filter for yourself and actually make strategic outcomes. So the book, the website, and then of course the workshop we’re doing together would be a great way of learning more and being able to use this in the real world.

    [33:57] Christina Ellwood: And you also could come into the organization and work with people too, ’cause you do consulting work. And we should have said early on that the book is called The River Doesn’t Wait: A Senior Executive Guide to Navigating the Surge of Agentic AI. And we’ll make sure that there’s a link to the book and to the website and to, all of Blaine’s work in the resources section and any other suggestions he has. There’s a wonderful area of his website, riverdoesntwait.com, that has all the references that he cites in the book, and he keeps that up to date so you can, go back to either access the resources again or find updates as he makes them. And I think that’s a really lovely addition that you’ve made to what most authors obviously don’t do that, but I think it’s a good addition that you’ve added. So let me ask you a different question similar to what you’ve been talking about, but in a little different way. We’re gonna personalize it. So in the AI revolution, what leadership skill do you find most valuable in your work today?

    [34:58] Blaine Mathieu: In, in my work and I think this applies to leaders generally and this is a, I think a fundamental insight that I’ve gained over the last decade or two, is most senior leaders and executives are really good deciders. All right? But they’re not practiced at revisiting a decision before the numbers make them, right? Once, once they’re not, no longer on track to achieve the goal, then they can start making decisions, right? And taking actions. But with the speed of the underlying change that we’re hap- that’s happening right now driven by AI and agentic AI, I think waiting until the numbers tell you something is probably too late because things are moving too quick. The new leadership skill, I think, is about being really good at revisiting decisions by watching and understanding the ever-evolving context before the numbers force you to do that. That’s a bit uncomfortable, especially in organizations that have very well-structured long-term planning processes, but that’s a skill that has to be built in the near term here.

    [36:11] Christina Ellwood: Great. So if our listeners remember just one thing from our conversation today, what should it be and why?

    [36:17] Blaine Mathieu: Let’s say there’s no such thing as an AI strategy. There’s a business outcome that you own, and the medium under it, we call it the river of agentic AI, has just moved. Even during the time we’ve had this conversation, it’s been moving. So start with the outcome, not the technology, and then understand how to assess this continuously moving context so you can make the best decisions.

    [36:42] Christina Ellwood: Blaine Mathieu, founder of The River Group and author of "The River Doesn’t Wait," thank you so much for sharing your experience with us today.

    [36:50] Blaine Mathieu: Thank you, Christina. It was a lot of fun.

 
Listen to other episodes
 
 
 
Join AI Realized
 
 
 
Subscribe to the newsletter
 
 
 
Christina Ellwood
Next

Enterprise AI at Ground Level: Beyond the Model