297. Boards in the age of AI

Boards need to understand the current trends in the marketplace, the current trends in technology. Not to fully be able to code AI themselves, but rather to be able to push the management team on how they're using AI in the business and to provide the relevant oversight for that.

For McKinsey Company, I'm Sean Brown and welcome to Inside the Strategy Room. You just heard Celia Huber talking about the imperative for boards to keep up with the rapid evolution of AI. As AI transforms how companies compete, many boards are grappling with how to provide effective oversight in this new era. To discuss how board directors can more quickly get up to speed with AI and effectively serve their companies, we're joined today by Celia and two other board and AI experts.

Celia is a senior partner in our San Francisco Bay Area office and she leads our North America board services practice. She advises organizations on strategies to help them thrive within the context of global industry change and on the evolving role of the board. Celia's recent board work and research have focused on board effectiveness, role clarity between the CEO and the chair, and broader governance, as well as the increasingly important role of boards in navigating evolving risks and uncertainties.

Celia is also an active board director for AMN Healthcare. Celia, welcome to the podcast.

Thanks, Sean.

Also joining us today is Aamer Baig. Aamer is a senior partner based in our Chicago office, a global leader in our technology practice. He's also chair of our knowledge council, and he advises CEOs, boards and executive teams on how to harness AI and enterprise technology to drive measurable performance and long-term value creation. Aamer's work spans advanced industries, banking, energy and materials, consumer and other global sectors, where he supports clients on tech and AI transformations. Aamer, welcome to the podcast.

Thank you, Sean. Glad to be here.

And Rebecca Schechter is a senior advisor with our healthcare practice, where she works with our private equity, corporate and government clients. Rebecca previously led one of the most successful deployments of AI and generative AI in healthcare at Microsoft Nuance. And prior to that, Rebecca was the CEO of Optum Behavioral Health, serving over 50 million members. She's also had President and C Suite roles at Liberty Mutual and State Street. Rebecca is a public board director at IA Financial Group, one of Canada's largest insurance and wealth management companies.

And she also serves on the board of the Center for Social Dynamics, a Goldman Private Equity Portfolio company. Rebecca, welcome to the podcast.

Thank you for having me.

Celia, you and Aamer recently published a new McKinsey.com article titled The AI Reckoning How Boards Can Evolve. And we'll provide a link to it in the show notes for today's podcast. But to start us off, maybe you could share some background on why you embarked on this research that led to the article.

Well, one primary reason we did this is we were getting a lot of inbound phone calls from board directors. And both Rebecca and I in our in our own boards were seeing a relative level of confusion, Sean, about, okay, AI is out there. Do we actually know how it's being used in our companies? What kind of oversight should the board be providing to management? And how, as board directors, do we stay abreast of these changes and stay current with what's happening in technology?

And Sean, uh, on the other side, we were also getting inbounds from management teams and CEOs and the rest of the C-suite. And in terms of how do we engage the board appropriately, we seem to be responding to the news of the day. So is there a structured way of of approaching it? So, so when Celia and I compared notes, it was, uh, easy to see that this is something that we can help bridge the gap.

So now, just to establish the background for today's discussion, Celia, your previous work on boards identified four board archetypes that have evolved over the past 25 years. Was hoping you could take us through them now.

So let me see if I can do it justice. What you see is what we call the complacent board. Often, the CEO would pick all the board members. They were less than independent, and they were often seen as rubber stamping, uh, what the CEO and the management team wanted to do. So in 2001, the Enron scandal happened and it led to an increased focus on compliance. So why hadn't the board seen what was happening at Enron?

And should we have the board really focus on compliance and their oversight? That worked reasonably well until 2008. In 2008, with the financial crisis, there were questions on why didn't the board see it coming? And as a result, there was a lot of pushback on board members that were not seen to bring an independent perspective. So if the board was really bringing that independent perspective, wouldn't these financial services companies have recognized some of these risks earlier?

And so what you then saw was board composition that focused very much on independent perspectives. So in 2014, uh, Toys R Us started leveraging up. They had a PE owner and they started leveraging to be able to do rapid expansion, although by 2017, they were bankrupt. And so one of the things that, uh, came out of that period of time was the notion of, gee, we should have had someone on the board who recognized the major changes that were happening on the market.

And Toys R Us did not have anyone who was an expert in e commerce and the type of changes that were happening in the retail market. So that leads us to this notion of a value enhancing board. And what we're gonna be talking about today is taking the next step between value enhancing and a catalyst board. And what we mean by catalyst, and I don't see many boards operating at this level yet, is boards are pushing the management team to set a higher ambition.

And to do that well, boards need to understand the current trends in the marketplace, the current trends in technology. Not to fully be able to code AI themselves, but rather to be able to push the management team on how they're using AI in the business and to provide the relevant oversight for that. And so that's really where we come to and and kind of the context for this conversation.

Thank you, Celia and Aamer. Maybe we could set the context for governance involving AI, and you could give us a quick tour describing what AI actually is and where it came from.

First, it might seem there's a lot being written about AI, and it's all happening very recently, but AI as a as a technology is actually fairly old. The new forms of AI are obviously new, but AI has evolved over the last six decades. In fact, AI was in invented in the 1950s. So there's a good chance that most of us on this cast, uh, were not around when AI AI was invented. And AI has also evolved quite a bit, and we use AI sometimes without even knowing it.

Machine learning, which is the type of AI that does predictive and prescriptive analysis, is been has been used a lot. Think of Netflix recommendations in in our daily use. In the business side, advanced analytics, machine vision and speech recognition are all part of machine learn, uh, machine learning. That led to a branch of AI that was called deep learning, and then that led us to what's called generative AI, which generates content and synthesizes content. The technology was also predicated on actually predicting the next word.

And we also, many of us are used to using, uh, some of the chat bots in our personal life for question and answers. And in businesses, we're also now used to having virtual experts, content generate, generation, and for many of us dealing with customer service, many times on the other side, there's a chatbot that is answering questions back to you. We're now at the phase of, uh, of AI evolution, which is all it appears to be about agentic.

Agentic is the next wave, and there are four discrete things in, you know, in large part that agentic AI does, building on generative AI. Agents, which are pieces of software and tools, can plan, can learn, can remember, and they can also act on your behalf. And that's where the word agentic comes from. They have some agency. We're talking about agents in, uh, in the back offices of many organizations now. We're talking about agents being digital assi- as assistants, and also in some areas, digital employees.

And this, over time, is gonna give way to what will be physical AI. This is AI in the physical world. The part about physical AI that is perhaps most relating to us is when we see what is called a humanoid, a robot in human form that can pick up things and move things, walk. But that's not the only part of, uh, physical AI. You can have AI in plants, in stores. This is the bridge between the virtual and and the physical.

And Aamer, a question related to this pace of change. It does feel like things in agentic in particular, but even some of just the new GPTs are evolving at an increasingly rapid rate, and, you know, some of them are writing their own code now. So as we think about the six decades that you just spanned, are we also seeing an exponential increase in the speed with which these things continue to evolve?

Great question. And and absolutely right, Sean. Speed seems to be the one common theme across how, uh, AI's evolved, particularly, you know, in the last decade, but it's really accelerated since 2023, since the launch of ChatGPT. We have a saying that says time has really collapsed. What used to be done in, uh, in years in the COVID era was done in months, and now with AI, it seems to be done in in weeks, sometimes.

And actually, if I could chime in with a with a real world example, um, having recently come out of, uh, Microsoft Nuance in healthcare. The benchmark to deploy technology in healthcare had been the deployment of the electronic health record, which was actually mandated in healthcare. And it most experts would agree that it took between seven and eight years to get to the point that all, uh, hospital systems and care delivery, uh, assets had some form of electronic health record.

The ambient clinical documentation solution, which was not mandated, was actually the doctors adopting it themselves. Actually went to a very similar degree of adoption in a matter of 18 months. So I think to Aamer's point, like just time collapsing and the speed with which this technology has been and and and will continue to get adopted and evolve is is vastly different than what we've seen historically.

And this must have far-reaching implications for board members as well.

Well, we're also at a moment where companies are asking about where AI adds value, right? And um, since 2023, when it evolved, we thought it was a very passionate affair. And like many affairs and relationships, uh, that that start off with, uh, a a lot of hoopla, uh and intensity validated by a lot of opinion leaders, uh, saying this is going to change the world. Now we're at a point where reality has hit and we say we need to talk.

And the question the of the day, of the month, and I think of 2026, is going to be, where is the value? Rebecca, y- you've lived this in both your board roles and also in your, uh, executive experience. What are you seeing?

I, I would say it's three major phases that we've gone through, and I think we're entering the fourth at this point. And I love the way you phrased this first blush, uh, the way one of my clients put it to me, uh, while I was at, uh, Microsoft, was he called it death by pilot. People were so enthusiastic about this very easy to use technology that only required natural language interface. You didn't need to be a computer scientist, you could actually just engage with this technology and try a whole new set of of capabilities.

What ended up happening is people prioritized deploying in the organization and having as many people use it as possible, and it just created a lot of activity. What ended up happening subsequently? So the second phase was at some point, the CFO in the organization pops his or her head out of their office and says, Well, hold on a minute. How much is this costing us? And is it actually driving value? And I think that's what you get from the first blush to the honeymoon phase, to say, okay, which one of these actually makes sense?

Which one of these is driving value? And then ultimately, uh, by the end of 2023, entering 2024, people started saying, well, even if it does bring value, can it actually scale? And I think on your chart it says reality hits, like, is this something that's gonna really matter? I actually think we're now entering a fourth phase, which is a much more strategic stance around. How does AI play in our business? How do we want to leverage it?

Where do we want to leverage it? And I think it goes back to Sean's initial question on what was the impetus for this article? We fundamentally believe, and having had this discussion, uh, at our boards and with clients. I do think we're entering a different phase where companies are really starting to think about, like, how do we do this in a much more strategic way? And I think the for the archetypes are are a great way to bring that to life.

It's a great segue to, uh, um, the paradox, uh, that we're seeing. The quote that we we often hear is, uh, AI is everywhere except in the bottom line, right? There's a saying that says, as of the sailor that's lost at sea, water, water everywhere, but not a drop to drink, right? There's water, there's AI everywhere except in the bottom line. And then reflecting your experience, Rebecca and many other companies. When we surveyed companies on where AI has been adopted, over 80% have said that, uh, AI, gen AI has been adopted in at least one business function or domain.

But when you ask them at the same time, where, you know, have you seen any real impact on your PL? Similar number, 80 said no real impact, right? And this is your question or comment on scale. So what we took away is while adoption is mainstream, impact is a niche, right? And then when we looked under the covers, uh, again reflecting, I think, your experience, but I'd love you to comment on it. Some of the reasons why this wasn't happening were reasons why the board, uh, really engages.

We found that AI was not directed towards area that areas that are really going to drive value. So it had a big value creation focus. Uh, the second was the question on scale, exactly what you pointed out, point solutions that weren't going to scale across the enterprise. The third was not a lot of emphasis on how this is going to change how work happens and how and what people do, the work talent, and then obviously about readiness and risk.

So when we step back and we said, hey, on one hand, we see gen AI not delivering value. But then when we look at the reasons why, the reasons were exactly the reasons why the board engages and challenges management teams. So that's another reason why this study came together. So love to hear your perspective on this Hi, as well.

very much echo that sense of it ultimately boils down to, I think, these these three fundamental questions, which is one, are you solving a problem that matters? And, you know, for me, when I led the business at Nuance Microsoft clinician burnout and the complexity of capturing that patient provider interaction in the most effective way was clearly something that was burdening clinicians. Given that for every one hour of clinical time with a patient, they were spending three hours on on administrative and documentation.

And so, yes, it was a problem that mattered. Two to your point is, can AI actually make a difference here? Does the technology actually work in the clinical workflow? And the answer is yes, it actually can capture the conversation, it can transform it reliably into what is a clinically validated note. And then the third thing is, can you actually deploy it and scale it rapidly so that it doesn't actually disrupt the most important part of that conversation between a doctor and a patient?

And that ability to integrate into the electronic health records, et cetera, was a critical aspect of how to, how to make it, how to make it happen. But it all boiled down to the fundamental strategic things that we knew were were vexing the the the healthcare customers that we were serving. Versus just, hey, it's a cool technology. What are what are places we could potentially deploy it?

And how about privacy? Rebecca, you were just, uh, speaking about healthcare. And one would think that medical records are probably some of the most sensitive data that you can have. How should companies think about privacy and adoption, and what protections, or maybe additional protections, they need to put in place? That boards should be asking about to make sure that their companies aren't taking on too much risk.

Rebecca, do you want to start on the privacy side? And then I'm happy to talk a little bit about how boards are doing it.

I, I'm happy to. And what one of the things that you you cannot lose sight of is that, yes, AI creates a a different set of risks. But you actually, as a board member, also have to manage the underlying risks that were already there. And what I mean by that is, you know, do you have a data protection policy? Do you have an IT risk view? Do you have a view into third party risk? Do you oversee and manage, uh, business continuity and resilience?

Like those are building blocks that you need to be able to have the right level of governance and oversight, in addition to thinking separately. And I think Celia, you'll talk about separately about what is the incremental AI risk. It's funny, Celia and I were presenting to a a board a few weeks ago, and interestingly, we got completely dragged into, like, cyber risk. Well, cyber risk is something you have to manage regardless. AI might bring it to a new level.

But I I do think as a board, really thinking through those building blocks you have to have in place. And then you have AI on top of it. So maybe I'll pass it back to you, Celia.

I, I totally agree with that, Rebecca, in terms of your 2nd question, Sean, on how boards are using it. I think that there are. It's a bit of a conundrum. So, um, many of the board uh platforms like Board Vantage, Diligent, et cetera, have some AI modules that they're starting to put in place. So things that can help synthesize, go back and forth across different dates of board meetings. You can look back at content from a year ago.

However, most companies have not adopted that, most companies the general counsel is concerned about the data being used, first of all, outside of the ecosystem of the company. And so while these board platforms are contained, there is some question about the AI either being able to listen in or have access to materials that are beyond what the board would normally get, and therefore opening up to potential liability. Now, I think that will be solved over time with the right guardrails, but I think we're still working on that.

And perhaps thinking about the real challenge, though, and there there are some statistics that two thirds of directors have limited AI knowledge. And by that, I mean, many of them are not using AI in their daily lives, have no interaction with ChatGPT and are still trying to understand what the technology does. And so certainly having a education on how to use AI in its most basic form can be helpful then in guiding the company.

And do you have any indication about the extent to which boards are overseeing or involved with the evolution of AI in their companies?

Yes, uh, we are so, in fact, just continuing on that thought process. Less than 40 of the Fortune five uh, Fortune 100 companies, uh, disclose that they have any sort of oversight. So if you look at their committee charters, they don't mention AI. If you look at the expertise of directors, they don't necessarily mention AI, they might mention technology. And if you think about what they're putting in their proxy about how they manage ethics, they don't necessarily have a strong point of view on AI.

In fact, one in three directors, if you can believe this, say, tell us in our surveys that AI does not even appear on their agendas. So they really not only have no oversight charter, but they don't have any discussion of AI and how that is changing their industry. And so that that is the real conundrum here. And we wanted to give both board members and management teams a way to discuss this, uh, and that's where the archetypes grew from.

Thanks, Celia, I was just going to ask you to talk about the archetypes and how they can help board members in their work. And in your article, uh, four archetypes are mentioned the pragmatic adopters, functional reinventors, business pioneers and internal transformers. And the archetype that will be most appropriate for a given company is really determined on where they fall on two strategic dimensions that you just spoke about. That source of value, optimizing internally or expanding strategically that AI is anticipated to bring to the business, as well as the degree of AI adoption across the enterprise either.

That's selective one part of the enterprise or holistic across the entire thing. So Celia was hoping you could talk a little bit more about those strategic dimensions to start. And again, we are linking to the article in the graphics in the show notes.

So the sources of value can be to optimize internally. So, for example, I get a lot of finance reporting, and perhaps instead of all the finance reports now, I can just go in and query, uh, what I want from the finance department and get those numbers and and trend lines on my own. There's a way to also expand strategically. So I'm entering a new business. Some of what Rebecca was talking about. And how do I then provide AI as a way to unlock some new value propositions I'm providing to my customers?

Then there's the degree of adoption, and we've seen two types of adoption out there, selective adoption. So let me try this in a particular area. And by the way, I don't mean a pilot, I mean a holistic way of doing it end to end, but in one business process, rather the than the company as a whole. And then we do see more holistic adoption that is more enterprise-wide on the important things. And so if you take those two dimensions, you can actually create a two by two matrix, uh, and lay out four different archetypes.

And so let me just quickly walk through those. So if someone is going to expand strategically and at the same time take a very holistic view across the enterprise, we call those business pioneers. So they're the folks that are going at it hard. AI is really the engine of growth and and reinvention in their industry. We do also see folks that are focused on optimizing internally, but also in a holistic way, and they're really deploying, uh, transformation to get there.

And so we call those internal transformers, and they're really thinking about rewiring what they do inside their company using AI. Again, they're putting forth the talent to be able to do this holistically. On the other side of the ledger, a more selective implementation of AI in a strategic way, we call those pragmatic adopters. So high on strategy, but more selective, so they are cautious adopters. But they're thinking about what Rebecca had to say. Where is the source of value?

And there, how? Therefore, how can I deploy AI in that area? We also see selective adopters that are focused internally, and we call those functional reinventors. So you might take my example of the finance function as a great example of selective, uh, and optimizing internally.

Super so, just to recap two dimensions that define the AI strategy source of value, optimizing internally and expanding strategically, and degree of adoption, embedding selectively in the organization or holistically throughout. And depending on where a company falls on those dimensions places them in one of the four archetypes. So now let's bring this matrix to life for our listeners through some examples of the different archetypes. Rebecca, maybe you could share one first.

Thank you. So the way I think about it is having worked for a hyperscaler like Microsoft, they very much made AI their business and how we were going to grow and scale in healthcare. And so they squarely, in in my view, end up in that business pioneers area, which has a whole new set, if you think of from a board perspective, many of the boards I served on, both public and private. If I step back about two years or so, most of them started as these functional reinventors.

It was very much about let let's look at this technology, let's understand how in in the case of the the the public financial service company I serve on is. Is it a way for else us to rethink our call center? Is it a way for us to rethink procurement? Is it a way for us to rethink how we, our coders, code? And over time, they've actually spread to the to the to other, uh, archetypes, and they're very thoughtful about which way they go in the in the matrix.

So I know, Aamer, you have, uh, some examples of the other two types.

Yeah, I was working with a global in industrial company that does a lot of manufacturing, uh, and distribution. When they looked at AI, they decided they are actually a functional reinventor because they decided there are only two processes or two functions that really make a difference to them. And they're going to put a disproportionate or of their energy into into that. So they didn't go whole hog for the whole company. They said these are the two places we're gonna do it, so they would be a functional reinventor.

Also have worked with another client that is a financial technology company that's seeing this as a new source of growth and revenue. They want to go from just delivering financial products to delivering data services. So they're seeing a bit big expansion of their target addressable market or TAM. So they would be in a business pioneer because they see AI changing every aspect of, uh, of their business. So it just shows that you can be in any one of these quadrants.

And if you are a large enough organization, you might have a business unit that picks, uh, one quadrant and another one that picks another one. But the important thing is to actually align on on your posture, so then you have a way of communicating with the board.

And for companies in the internal transformers archetype, what are some of the most reliable board level indicators that AI is actually rewiring the operating model? Rather than just sort of automated, uh, automating, excuse me, isolated tasks, for example.

Maybe I can go f go first if and and then, uh, Celia and Rebecca, please, please follow. I think if you're an internal transformer, you're really focused on cross functional improvement of your capabilities, whether it's processes or technology or policies and what have you. As a result, change management becomes really important, and there are lots and lots of examples where improvements are one time. So the big question for the board is to make sure that the rewiring that's happening across is structural and that the changes and the improvements are not temporal, that they last, uh, for a period of time.

Yeah, and I would just add talent, talent, talent. So I think to do either of these holistic archetypes, the internal transformers or the business pioneers, like you really need talent spread across your business. Uh, understanding the business. It's much harder than doing it selectively, where you can focus your talent that understands AI in your enterprise.

And the last touch I'd add is it's really about thinking about the value chain end to end. And and maybe to illustrate this, um, one of the boards I sit on is in the care delivery space. And one of the things that they have very capably rethought and reinvented is the ability for patients to reschedule their appointments and get matched with a provider. Now, in order to do that, you've got to think through two fundamentally separate things, which is how does the patient interact with the provider and how do they reschedule?

Et cetera, as well as what are the provider schedules and where are they located? And are they available? And do they have the right specialties? And are they of the right level? And so that ability to rethink end to end. And I think Aamer, your key point around is, it does it fundamentally change how scalable the process is actually requires a complete different end to end change management thinking in order to to fit in that box.

And Celia, in the article, alongside those four archetypes, you also lay out the related implications for the board's traditional role as a challenger to the management team. And and how that might look different for each one of these archetypes. Maybe you could walk us through those, starting with the business pioneers.

So we have laid out by each archetype the different type of challenger role the board should take on when dealing with that archetype. So again, if we start with the most holistic change, the most strategic change, the business pioneer, then, uh, the board should really be challenging management and having a debate on are we making the right strategic bets? Because remember, in the business pioneer, those bets are changing the company and your value proposition. And so thinking about how much capital put it to put in, when's the ROI, what's it gonna pay off, uh, when's it gonna pay off?

And also putting a highlighted attention to risk management, not in terms of just the risk of the technology, but actually the risk of the entire business plan and how that might change if regulation changes, if competitors change, kind of the normal big questions you'd have on a strategy. If I think about the um, more strategic and more selective adopters, the pragmatic adopters again, you still want to make sure that you understand risk and the strategy. But you may need to pivot quickly if your competitors are doing the holistic adoption of the business pioneer and you have taken the more selective adoption of the pragmatic adopter.

You may need to do a much better job of keeping an eye on long-term changes and then therefore your competitiveness, as you don't want to be blindsided by a competitor that leapfrogs you, uh, because they've taken on a more holistic, uh, pace of change. And then again, on the um holistic, but more internally focused, the internal transformer. Again, you'd want to see what Aamer talked about, that the rewiring is happening across the operating model. Um, that you have, as I mentioned, enough talent to reskill the organization, redesign those incentives and think about the change management that's necessary to get there.

At the same time, uh, that you're doing it at a pace of change that's really gonna matter. And then our last archetype, the folks who are more selective and more internally focused, the functional reinventor, thinking about the incentives to make that happen, thinking about, um, how not to over fragment the staff that is high skilled on AI. And not ending up in this death by a thousand pilots kind of situation that Rebecca described earlier. And so I do think, ironically, this more selective and more, um, uh, internal focus can be one of the hardest to get right.

Because you may end up feeling like you're not getting any value out of it. And so thinking very clearly about where you're gonna apply AI and for the board to really challenge on that is important.

Thank you, Celia. And what about the implications for the pragmatic adopters?

So for the pragmatic adopters, it's really about staying aware of the moves you're making, because remember, these are the strategic, selective moves, the moves you're making relative to others. And thinking about what AI vendors in your space, your industry are offering. And I wonder, Rebecca, if you want to talk a little bit about this, um, because I know when you were working at Microsoft, you probably had a lot of clients that fell into the pragmatic adopter range.

Uh, they did, and to your point of it being harder, um, really, we. The discussion often was around, okay? Which one of these things is going to truly unlock a a vexing problem? Help us understand how you, as a user of the technology, are are going to do this and how are you going to do this? And and really thinking hard about what is important to you strategically, what are the capabilities you need to control versus those that you can buy or borrow or partner with?

Um, the the pragmatic adopters is very much where that rubber hits the road. And thinking through what are things that, frankly, are better bought because others are able to build those versus things that you want to build or adopt on top of that. That are specific to what you're trying to to achieve in your very, very specific selective bets. And that that dichotomy between the do I build, do I buy, who do I partner with? Ended up being a a really important conversation, given the other point you made earlier.

The level of capital investment required, particularly when it comes to AI capabilities, in order to actually take it to a more strategic level.

In the board I sit on, uh, Rebecca, one of the things that we're doing as the pragmatic adopter, uh, for our biggest business unit. And we recently found that we leapfrogged in one core area of value creation, and so this was around being able to certify and get nurses out very quickly to hospital systems, leapfrogged the industry. And in fact, uh, it was both a dramatic improvement in our metrics, uh, and even analysts in Wall Street noticed that we were able to do this quite quickly.

And so we'll now take that back from the pragmatic adopter. Kind of what Aamer was saying, and think about where else in the other business units we can apply that same skill set.

And Aamer, it looks like you'd like to jump in here.

Yeah, I wanna go back to a point we made at the outset, which is time is collapsing. And, uh, Rebecca had given a very good example of of, uh, of how it had collapsed in one of her, uh, companies. So one of the risks to watch out for in being a fast follower or pragmatic adopter is to see if things are gonna move suddenly in your market. So you have to have signals that you want to watch for.

And that's one of the questions that the board can ask, mad, the management team, what are the signals we're gonna look for, right? And, uh, Celia covered a few in terms of competitor moves, but it could, it could also be something in an adjacent market that just comes to you. So that's just one thing I wanted to highlight, because, uh, things sometime happen very quickly and it's important to watch and pivot.

I I just to add a little bit of color to that. What you wanna avoid is what I initially saw a few clients do when I was at Microsoft, which is what I call the freeze in place. The it is moving so fast, I'm not sure what to do. That's what I was trying to illustrate earlier of, like, being very thoughtful about where you're gonna lean in, who you're gonna buy what from to actually move the the the problem forward.

Versus Tr. The risk being to wait on the sidelines and then, to your point, potentially being leapfrogged.

So Celia, maybe now you could take us through the internal transformer archetype.

So remember, this was a holistic view of the company and rewiring using AI, so more internally focused, hence the term internal transformers. And this does relate very much to the book Rewired. If you're interested in doing this, thinking about the whole enterprise and how to rewire the full operating model. So a couple things are important here, and the board should really push for end to end integrated solutions, not siloed. And the reason being, with AI, you can rethink the end to end business process completely, not simply optimize the process that you had before.

Um, to do that, though, does require eyes wide open in terms of the risks that you're taking on, um, in terms of testing as well. And I, you know, I believe the oversight structure structure needs to then be tuned for that. So having at least an AI fluent director, someone who can, uh, really push on the management team. You would then be s a board that would see it at the full board either monthly or quarterly, depending on what was going on.

And you might consider a specialized committee, either an ad hoc one or a permanent one, uh, for tech or risk. That would also have in its charter thinking through the internal transformation that AI is being deployed for. Now, Aamer, I think a lot of your clients are in this bucket as well. And I thought you might wanna just describe a couple of examples that you've seen.

Yeah, uh, thanks, Celia. A lot of, um, the clients I serve have enterprise AI efforts, so enterprise AI transformation efforts that, by definition, go across functions and regions and businesses. So this is this requires a, um, lot of muscle building in terms of how, uh, the company transforms because you're actually changing the way, uh, the business works. In some ways, you're actually, um, going to a very different cost curve, you know, to be able to be competitive.

And I would say, um, uh, the biggest risk here end up being not in technology, but more in change management and an execution risks. And sometimes management teams, uh, underestimate the, uh, ability of the organization to adopt and embrace this type of change. So, um, what we have talked about, uh, a lot over the years has been the importance of people in change management. That becomes really, really important in this.

Aamer, can I ask you? A number of folks are talking about having an org chart with agents in it, and I think it's more likely to happen in this archetype. Uh, do you want to talk a little bit about how folks manage the agentic part of this, particularly in the internal transformer?

Yeah, it's, it's, it's a great and hot question. Um, uh, I would say, um, first of all, that, uh, having agents on an org chart is, uh, is is also a niche, and it's it's not really, uh, you know, widely adopted, although in some tech companies, it's actually being being publicized. I think the core, uh, is not really just the org chart, but how work gets distributed between agents and humans. This requires a real reimagining of the business process, right?

Even if, uh, if you pick up a simple process like, uh, uh, like, um, updating an invoice or, uh, processing a claim, right, or updating a system, or, uh, creating a ticket of some sort. Usually, these are done through very well documented processes by humans. If you just put agents in there, that's going to make the process more efficient. And if the process itself was inefficient, you're gonna make an inefficient process better, and you're not gonna get better outcomes.

So it requires companies and teams to really focus on the actual outcome we're trying to achieve. And then be somewhat clean sheet about how the work gets done and really rethink the delineation of of work. Um, one simple um construct that I've seen is separating out the routine work with the highly complex work. Right now, highly complex, uh, exception work being done by humans and some of the routine work being done by agents. So that does divide up the work and it creates new a new coordinating mechanism.

But it does give you a much higher lift in performance than just keeping the process as is and having agents do the same process.

And what about people who find themselves in the functional reinventor camp, but they really want to be an internal transformer? Should they focus on more than one function at a time?

Let me start, and then I think Rebecca might also have something to add here. Look, I, I do see a lot of companies who are in this conundrum, and sometimes it's because not all of the management team is on board with the internal transformer. And so, um, the the management team has some, uh, leaders that are volunteering to be, uh, focused on their function. And so they see a lot of value in AI and they're deploying it.

And fundamentally, then they force the company into a functional reinventor if they can't get the whole management team to stack hands. Rebecca, what do you see on this?

Yeah, and uh, I. This is a lot of the evolution that I just I've now witnessed on on some of the boards I sit, which is it. It was easier to start as a functional uh uh reinventor, because, frankly, there were some use cases, it drove value and you could actually capture that in in reasonable time. Um, also, if you're a company that is a very high regulatory burden, these internal transformations, you have to bring your regulators along with you.

And so that is a lot of what I see helping them understand. To Aamer's point, like this is routine and lower risk versus this is complexity and higher risk. By the way, this touches customers and has, you know, compliance and other things attached to it. And being really, really thoughtful to to Celia's point across the management team is is where are we going to make a bet on an internal transformation? And what? And how are we gonna bring people along, not only inside the company, but I'd even argue for outside the company?

Um, because it has such a high, it has potential for such a high impact in change management.

And Aamer, before we wrap up, was hoping you could quickly share observations you've had regarding boards and how they're currently working with management teams on AI.

Great, thank you, Sean. Look, some of the typical questions and and and observations we have precipitated this dialogue. One is, uh, really, why are we not aligned between the management team and the board? And part of our effort here, humbly, perhaps, is to create a new vocabulary around postures, around archetypes, to then create a mechanism for, uh, management teams and boards to engage, right? That's that that was one question that really came up. I think there are other questions, which is wh- which come up from boards that Celia and and Rebecca can can also weigh in on, which is what should they expect from investors, their investors?

And our initial ongoing thought is because the boards are the fiduciary representing the investors in in many ways, the question.