C-Suite Perspectives Recording with Matt Rosenbaum and Erka Amursi - Transforming Organizations for AI

[00:00:00]

Steve Odland: Welcome to C-Suite Perspectives, a signature series by The Conference Board. I'm Steve Odland from The Conference Board and the host of this podcast series. And in today's conversation, we're going to talk about how organizations are transforming for AI, what strong AI strategy is necessary going into 2026, and how organizations can successfully redesign their strategies, structures, and processes.

Joining me today are Matt Rosenbaum and Erka Amursi, both principal researchers with The Conference Board's Human Capital Center. Matt, Erka, welcome to the show.

Matt Rosenbaum: Thanks for having us.

Steve Odland: You all did a great study that you've just published. And for our listeners, that paper is available at tcb—our initials, The Conference Board—tcb.org, and then click on the Human Capital Center under North America, and you can find their great study called [00:01:00] "Transforming Organizations for AI." But we're going to spend the rest of this podcast and we're going to talk about some of the highlights of this.

Erka, let's start with you. Why did you write this?

Erka Amursi: Interesting question. I was very interested in how AI would impact organization design because it's not just a technological update, upgrade, but it has a transformative power. So, moving forward with the same structures, with the same workflows and processes, with all this transformation coming in, seems very challenging. And a lot of organizations need guidance in terms of how to adapt with the design in order to be able to use AI and its benefits at its best.

Steve Odland: And Matt, this is different than just taking on yet another software tool. So it's not like, we're going to roll out Teams [00:02:00] or Zoom, because AI just has so many tentacles into the business, doesn't it?

Matt Rosenbaum: Yeah. I think one of the main differences is it's not just a tech implementation. That was one of the points that came up in our interviews with various enterprise leaders, was the need to treat this as a bit more of a wholesale change than just "we're swapping out some software."

And I think part of that is because one, AI use cases often cut across the organizational boundaries. So you have a lot of folks who are maybe collaborating more closely with partners than they may have in the past or in different ways, whether it be between HR and IT or HR and legal, or your risk management functions, lines of business leaders, depending on your organizational structure.

So there is just a lot of partnership required. And then also, AI changes what differentiates you as an organization. One of the folks I was talking to is a leader at a large education company, and she was saying because of AI [00:03:00] capabilities, they had to put personalization much more at the forefront of some of their offerings. So it just changed the products that they were offering to customers in the first place. So it really does affect every facet of the enterprise.

Steve Odland: So Erka, your point was that structures needed to change, and that's not something that usually happens when you adopt different tools. And so this is more than a tool, as Matt was saying. It's really a transformation. It's a strategic adoption, if you will, which then does require organizations to be aligned about that. And that's really what you're saying in the paper.

Erka Amursi: Yeah. As long as we have a transformative power, you better have a strategy. Otherwise it can be chaos, right? So, that's why having a strong strategy is especially important. You want to make sense of what is happening, and you want to be able to adapt better. And not only adapt, but also be more productive, be more efficient, innovate better [00:04:00] through this new, more-than-technology thing. And that's why strategy is helpful.

Steve Odland: And what do you see as the key components, the key strategy components for AI in a company?

Erka Amursi: So what we discussed in our report is this integration of top-down and bottom-up elements and finding the best balance between the two because you need the vision, you need to have some perspective on how you're going to implement AI. Are you going to build your internal tools, or are you going to purchase something? Or how you are going to design your workflows and processes based on it?

But you also need to experiment and have your own lab through bottom-up innovation. You have AI implemented in different teams in the organization, and you have those people coming up with ideas on how to better use AI, on how to improve efficiencies on specific teams or processes. So the integration of these two [00:05:00] directions, top-down vision and bottom-up experimentation and ideas, is probably the ideal path moving forward into designing these organizations.

Steve Odland: So Matt, if you're in a company, and you're trying to put together an AI strategy. As Erka was saying, what are the key components that you advise people to consider?

Matt Rosenbaum: Yeah. I would start with the "why" of, what are we trying to accomplish here? What are our goals as an organization? Do AI capabilities change that in any way? What differentiates you and your business. From all your competitors? And then looking at the "what," specifically the task level of, what are we trying to do as an organization? What is the work that we need to keep doing? What are some of the things we need to change?

And then you can look more closely at the "how" of, what are the actual workflows by which we accomplish that work? Is it a person doing this? Is it a person in partnership with machine? Are we [00:06:00] automating it entirely? But ultimately, the goal with the strategy is not just to be able to determine what you're going to say yes to, but what you're going to say no to.

I think a lot of organizations sometimes forget or misunderstand just some of the trade-offs that need to be made when it comes to resources for AI initiatives. So having a clear vision of "this is what we're trying to accomplish" and "this is what we're going to prioritize," is going to help you make a lot of those decisions and avoid wasted effort.

Steve Odland: But you don't do AI for the sake of doing AI, right? I mean, you do it for a constituency, and in today's business world, there are multiple constituencies, customers, employees, owners, community, and so forth. So you have to start with that: What is it that we're trying to accomplish in our strategy? It's not an AI strategy, it's a business. It's how does AI compliment and facilitate the business strategy? And so that business strategy is targeted to this constituency group.

And so what do we do to be more [00:07:00] innovative around our customer, to facilitate growth? What do we do around our employees to ease their—you know, yada, yada, yada. My point is that you start with, it's a business strategy that is really focused on those constituents. Isn't that right, Erka?

Erka Amursi: Yes. And of course, you need to align your AI strategy with your business strategy. So you need to see how AI will help you with your business strategy, but also AI has the potential to actually help you innovate in different areas, and maybe AI may help you develop new business strategies or focus on areas that you haven't considered before.

Steve Odland: Yeah. Matt, what do you say to those people who say, AI is just an evolution of software. It's not that big of a deal. People are really overreacting."

Matt Rosenbaum: I think there's some element of truth to that skepticism. Like there are certainly people who are overreacting. There is a lot of hype, there is a lot of noise. But fundamentally, [00:08:00] even if the technology does not continue to progress and improve, just integrating what we currently have available is going to take quite a lot of time and quite a lot of effort, particularly as people figure out new ways to use existing tools and systems, let alone if there is continued progress.

The world as we know it is changing in terms of who does work, what can be done by a machine that previously couldn't. And we've had pathways of innovation. We've had Excel, replace, or not replace, but get rid of some of the back office functions and the way spreadsheets used to work back in the '70s, for instance. That world has gone away, but we haven't necessarily seen the large-scale effects that some people are predicting from AI specifically.

And I think that comes down to the question of, we've had automation come for manual work and seeing the effects of that. And I [00:09:00] think now the question is, are we getting automation to affect knowledge work in a similar capacity? That is yet to be determined. But I think that's the big question that people are asking is, if we have the same effect on labor from knowledge work perspectives that we've had on some of the manual work that's been done in the past, what does that change for how we operate as an organization?

Steve Odland: Erka, when you were saying before that you are excited by the need for a transformation of structure and processes for AI, it's not for the sake of AI, as we said. It's really in order to make sure that AI is deployed effectively. But what is it about traditional structures, legacy structures and processes, that may hinder AI adoption?

Erka Amursi: Yeah. If we think about legacy structures, they have all been built around humans, and they have been built to deal with humans' potential and limitations. If we go back, for example, it comes to mind, all the [00:10:00] work that has been done at the early last century with industrial psychologists trying to understand human behavior at the workplace, how efficient they can be in a certain amount of time, how much tasks they can do in a factory, et cetera, or scientific management.

All of that work has been designed to understand humans, their capabilities and organize those capabilities into a system so that organizations could produce more complex products or delivers more complex services. But humans are limited, like each of us has a profession. It's one person, one profession. And work has been designed to integrate all of that with AI. We may have AI systems or tools or agents that can have multi domains. They can have the legal knowledge, the economic law knowledge. They [00:11:00] can be physicians, engineers, researchers. So we have all of these domains in one AI system, right? And all of this is a new capability that is added to the organization. And it will impact the way you organize your work around these new capabilities. It'll require new processes and new workflows because these AI systems can be the new glue that concentrate all these knowledges and points of view from different domains.

And if you put this complicated, this multi-domain integrated system into a siloed process, it'll not add much. It may help with automating some things, but the system will remain broken. While the potential of these systems are to integrate the processes, to integrate the workflows, to change them from beginning to end. So by using them into [00:12:00] siloed, hierarchical, old patient legacy systems, you are not gaining their full potential. You are just increasing some efficiencies, but the system will remain slow and broken.

Steve Odland: Yeah. And Matt, the reason that makes a lot of sense is cause AI is like really like nothing else that we've had before. It is on one hand, a tool, another hand, it's software, on another hand. It's the closest we've had to machines being able to quote-unquote think. They're not sentient. But this gives us power, great power analytically, creatively, innovatively. Agentic AI allows some level of replication of human interaction.

The reason is that it is so different. And it is so applicable to so many different areas that you can't think about just overlaying it on a current organization.

Matt Rosenbaum: Yeah. I think the things that stand out to me about what makes AI different, as it were, [00:13:00] is one, the capacity for some sort of autonomous action. It may not be much yet, but certainly, agents are developing that in the terms of the task length that you can expect an agent to accomplish without human input. That increases day by day. So that's one element that's very different.

And then as Erka was mentioning, the multi-dimensional capacity to have all of that bound up in a single system versus having to have a team of different people with knowledge domains that they can bring together. That's another element, as well. As one of the people we talked with said, in a way, the human has become the integration layer for the AI, in so far as you have humans in the loop kind of overseeing AI systems and nudging them in different directions if they're going off the rails. But ultimately, a lot of the work now is being done, or looks like it can be done by AI systems, and then humans will plug in and fill those gaps.

Steve Odland: We're talking about the [00:14:00] transformation of organizations for AI. We're going to take a short break and be right back.

Welcome back to C-Suite Perspectives. I'm your host, Steve Odland, from The Conference Board, and I'm joined today by Matt Rosenbaum and Erka Amursi, both principal researchers with The Conference Board's Human Capital Center. And they're co-authors of our latest report, "Transforming Organizations for AI," which is available on our website, tcb—that's The Conference Board—tcb.org. And you just click on the Human Capital Center, and you'll find their latest work.

So, Erka, let's come back to you. What leadership models are organizations using to guide these AI redesigns?

Erka Amursi: During our study, we conducted a survey. We sent the survey out to leaders in the organizations, and we asked them about the leadership model that they are using. From their answers, we understood that there is a variety of types of governance here.

Starting with [00:15:00] cross-functional project management teams that are assigned the role of leading these projects. Added AI oversight on the existing executive roles is another option, another variant. Task line managers with driving AI redesigns and adoption is another option that organizations have used. Delegating AI leadership to functional and business division heads is also another way of governing AI.

So there is a variety out there, with different kinds of settings. And some organizations, 32% of them, for example, said that they have created a new AI-focused executive position in the organization. So yeah, there is a variety.

Steve Odland: OK, so Matt, when you think about transforming organizational design, all that, it's like, that's human resources' job, right? But what I hear you saying is, this is so different, because it requires strategic transformation first, [00:16:00] and then organizations follow that.

So you start with what's the business strategy? How does AI integrate with the business strategy? And then what models of organization come out of that, right? So it seems like this really goes beyond traditional HR, doesn't it?

Matt Rosenbaum: Absolutely. I think this cuts across numerous different remits. HR obviously plays a key role, but certainly a lot of organizations are focusing this either in their IT organization, or maybe operations is leading the charge on this. Or, as Erka mentioned, sometimes they're appointing some sort of AI executive officer to oversee everything across the enterprise.

But ultimately, I think, regardless of whether you have it be the remit of one specific person who's overseeing all of AI across the enterprise, or if it's a team of people, there is a need for collaboration across the C-Suite and across the enterprise because of just how far-reaching the change is. And it should flow from that central question of, OK, what is [00:17:00] our business strategy? How does AI enable that? And then everything else will kind of trickle down from there.

Steve Odland: So the entire C-Suite needs to be involved here. And really, truly all business leadership needs to be involved to make sure that each of their areas is organized in a way to effectively deploy this.

Matt Rosenbaum: Yeah.

Steve Odland: Yeah. Erka, you can't just drop this stuff on an organization and just say, OK, have fun, right? There is training and change management that needs to go with that. So talk about how you do that in an organization. We talked about strategy. AI is a tool of strategy. You follow it with organizational design, but then you have to deploy it. You have to implement it. And that's where the change management comes in. How do you suggest people go about that?

Erka Amursi: One of the things that we are hearing over and over from the leaders is the engagement of people as early as possible in the process. Being transparent, having your employees involved in designing the [00:18:00] change that will affect them. Having them listening to their concerns and fears. Being as transparent as possible with your vision and the way you are going to implement AI.

A lot of studies are showing that trust is really important in adoption from employees, so engaging there. Engaging them early, being as transparent as you can. Providing as much data and information that will back up your decisions is essential. But also having them come up with new ideas on how they are using AI. Having them come up with suggestions is important. Listening to them constantly about fears that are emerging. Because fears can emerge during any change implementation process at different times because of different communications. It can come from internal implementations or communications, but it can come from also external things that are happening to other organizations. So being able to manage that, attitudes and concerns and [00:19:00] emotionality, is important.

And as well as providing guidance on how the roles are going to be transformed. What is going to be taken away from your current role, but what is going to be gained? What skills do you need in order to do this new role better? And what is the exact support that the organization is giving you in that regard? Is it a formal training? Are we providing you tools that you are going to learn on your own? Are you going to get a coach or a mentor or an AI coach to help you with that. So being clear on all the resources that you're going to offer in order for them to build the required skills.

Steve Odland: Matt, your thoughts on change management.

Matt Rosenbaum: Like Erka was saying, there's a lot of fear and concern, so I would just highlight the need to address that and assuage those concerns as best you can. But also what we were talking about earlier, the clarity of vision. What are you trying to accomplish with AI? and having that be clear to people [00:20:00] across the organization and consistent in the messaging is, I think, an incredibly important part of the change management.

But also one of the things that stands out to me from our conversations with executives is, yes, you want to focus on adoption, but adoption is just the beginning. Ultimately, what you're trying to accomplish here is the outcomes that come from using AI effectively.

And the challenge there is a lot of organizations don't really have a good baseline to measure, is this AI tool or system that we're introducing actually helping us in this regard or not? So if anything, I would say the change management component that also needs to be addressed is understanding the baseline performance of the organization as best you can and buttressing that where needed, so you can determine whether your AI initiatives are producing the results you want to see or not.

Steve Odland: So, Erka, both you and Matt have indicated this isn't something you can just simply drop on an organization. You can't just drop AI on them. You have to train AI. You can't just drop AI in the [00:21:00] absence of strategy. You have to really link it all for people. You have to build a case so that people understand. You got to really train people so that they understand, and you wipe out the fear and that people live in the state of possibility. All of these great things.

But that suggests that it's not something that leaders just throw against the organization. This needs to be an interactive process. Leaders and all the people in the organization working together to figure out how to implement, adopt, and make AI most effective, doesn't it?

Erka Amursi: Yes, exactly. And this collaboration needs to be long term, because AI will continue to evolve, and new tools and capabilities will continue to come up, which will need continuous collaboration. Which on one side will be positive for the organization, because that closer communication will help build trust across levels and will help with other [00:22:00] change adoptions, will help organizations move forward faster because they will be forced to have easier ways of communication between them. There is no other way.

Steve Odland: Yeah. So Matt, what's the best way to engage employees in this process?

Matt Rosenbaum: So one of the important things there is to start early. I think one of the most surprising findings from our survey was, we asked leaders and workers how involved workers should be in AI organizational redesigns. And to my surprise, 56% of leaders and 48% of workers said that workers should be fully involved, which means they're part of the kind of co-creation of what we're trying to accomplish as an organization with AI from the ideation phase, rather than the leaders coming up with what we're trying to accomplish, and then the workers are involved in implementing it only.

And I don't know how realistic that is for some elements of AI redesigns. Like I think the strategy is probably going to be set by [00:23:00] leaders, whether we like it or not. But ideally, there will be some input from workers on that component. But certainly when you get into some of the structural and procedural elements—how should we redesign this workflow? That is especially crucial to have workers be working hand in hand with leaders, rather than leaders come up with whatever they think is best and then telling workers to go and implement it.

Steve Odland: OK. And finally, Erka, talk about the incentives that have to be in place to make all this work.

Erka Amursi: The incentives. First of all, before going to incentives, organizations are doing a lot of work in terms of understanding what are the skills that they actually need, skills and capabilities, mentalities, behaviors, attitudes that they will need in the future when workers will work more closely with AI agents or tools.

So identifying those critical skills and behaviors will be essential in developing the [00:24:00] performance evaluation systems and the rewards that will be attached to those evaluations. So whole performance evaluation systems will need to change based on new priorities of skills and behaviors that you'll need to see in the future.

Matt Rosenbaum: Yeah.

Erka Amursi: And tie that to the reward system. So it may be the same rewards that we have, but tied to different behaviors, to different outcomes, to different capabilities and skills.

Steve Odland: Matt, last word. Any remaining thoughts that we want to share regarding the paper?

Matt Rosenbaum: I think the overarching point that I would end on is just there's a lot of change happening. There's a lot of change coming. But it's important to start where you are. You have to walk before you can run. And I think there might be some organizations that are saying, we'll just wait for this to all settle down, and then Accenture, we'll pay them a couple million or whatever it might be, and have them come in and tell us what to do.

But there really is a need [00:25:00] to experiment and learn as the technology develops, and learn what your organization can do well and effectively, and where some of the areas internally might need to be buttressed or developed in order to use these systems effectively, versus waiting to see how this all shakes out.

I think the important thing is to get in there, start that experimentation, and start developing that culture now.

Steve Odland: All right. Words of wisdom. And again, you can find this new paper from Matt and Erka on our website. It's called "Transforming Organizations for AI." It's at tcb.org under the Human Capital Center, the North American Human Capital Center. Just click on it, and you'll find this paper. It's really worth the read.

Matt, Erka, thanks for being with us today.

Matt Rosenbaum: Thanks for having us.

Steve Odland: And thanks to all of you for listening to C-Suite Perspectives. I'm Steve Odland, and this series has been brought to you by The Conference Board.