How to Build the AI Skills Your Business Really Needs

Diana Scott: Many organizations have invested heavily in AI tools and yet your research suggests workforce capability is lagging behind. Can you talk a little bit about the biggest misconceptions executives have about what it actually takes to build an AI-ready workforce?

Matt Rosenbaum: Yeah, absolutely. I think the two things stood out to us from our research, looking at organizations globally and also how workers are feeling about the skilling opportunities that they're provided. One, I think organizations are overly focused on the prompting itself, when if you do it right, the prompting is what matters the least of all the things you can be focused on in terms of AI training and AI capabilities.

And then the other thing that stood out to us is a lot of organizations are taking this seriously but I would say they're not taking it seriously enough in terms of actually being prepared to meet AI capabilities both now and in the future and integrate those successfully.

I think a lot of organizations are focused in 2026 on preparing people for 2024 technology. So basic chatbots, understanding how to prompt, basic AI literacy, things like that dominated a lot of our conversations with executives, instead of things like agentic AI or the surrounding components that make people able to use AI well, such as holistic thinking, business acumen, their understanding of the systems of the organization as a whole.

Those are the sorts of things that are much more valuable than do you know a prompting framework to organize your prompt. Because especially now, the technologies are getting so much better at interpreting vague prompts that really-- yes, there is value in being able to prompt well but unless you're at superstar level, the real value is going to be are you asking the right questions and looking for the right things in the first place to help the business move forward.

Diana Scott: Yeah, and I also noticed that the employees themselves are actually going out and learning a lot on their own, and yet the research shows that the organizations aren't necessarily the ones that are teaching them at the same time. And so there's really a disconnect there, isn't there?

Matt Rosenbaum: Yeah, I think sometimes it's because organizations, again, are often appealing to the least common denominator, so there's a lot of very basic training, just clicking through to say you completed it, and then everyone says, "Look, we had 89% of our workforce complete their AI training," and the reality is nothing of value has actually occurred in the process.

And the other challenge, I think, and maybe an opportunity for organizations is workers can go to other places to learn generic AI usage training. You can go to Coursera, you can read stuff on Reddit, you can follow people on X or whatever it may be. But what makes it actually valuable is seeing how it can be applied in your current work, and that's something where, again, I think some organizations are doing that well but a lot of organizations have room to improve.

Diana Scott: Looking at the research, if the employees are doing this on their own, I think that creates a certain sense of lack of trust in terms of what is the organization going to do for me? And then if you look again at some of the report's central arguments, it's that AI literacy, to your point, is only the starting point.

Can you talk a little bit more then about if it is only the starting point, and most of the organizations are just only focusing on that, then what separates the organizations that are simply teaching employees to use AI from then what actually does create that measurable business value?

Matt Rosenbaum: Yeah. I think a few things that stood out are, one, knowing what business value you're trying to achieve with AI and where you're choosing to focus your efforts versus not.

A lot of organizations couldn't tell you what they're not trying to use AI for, and it can be illustrative to understand, okay, these are the areas that we are focusing on, and the trade-off of that is we are not focusing on these other areas, whatever they may be for each individual business.

But if you don't know what you're saying no to, I don't think you really understand kind of what goal you are trying to accomplish, where you're really putting the bulk of your efforts. So that's one thing that stands out, is having a clear idea of what are we trying to accomplish as a business and how does AI fit into that?

And then the other thing that was important is, like I said, the components around actual AI use that make the difference, and is that use being directed towards those business outcomes, yes or no? A lot of organizations we talked to are using very basic volume sort of metrics for assessing whether or not people are completing trainings, things like that, but they're not assessing the outcomes tied to those trainings, and those can be quite difficult to determine.

But the incredible importance thereof is that you need to know what is working and what is not, and if you don't know how your training is actually tied to business outcomes, then you're not really completing that essential piece of the puzzle.

Diana Scott: Can you step back a little bit and talk about how you actually conducted this research? Because I think when you actually talk to some of the executives, you got some real granular detail about what was working and where people actually did dig in, which I thought was really interesting. Some of the use cases, some of the details of organizations that were doing the deeper dives, where it was working, where it wasn't working.

That part of the report was really interesting. You also did some quantitative research, mostly out to workers as well. Frame that a little bit for us so that our listeners can understand a little bit about that, because I think the frame is interesting.

Matt Rosenbaum: Yeah, absolutely. So what we did was combine two main streams of research, one of which was a survey of workers globally that we had out in the field in late April/early May of 2026.

That was just to understand how do workers feel about AI. How do they feel about the skilling opportunities that are being given to them? Are they taking advantage of that? What are some of the potential weaknesses that leaders may or may not be seeing? Because there's always the reality of this is what the leaders pitch and think of as their organization's skilling program and this is what actually is being perceived by workers. We wanted to understand from that perspective, so we used a survey for that.

We also conducted, I think, about 40 interviews with leaders at large enterprises around the world, primarily North America, Europe, and Southeast Asia, and understanding how they were using AI and how they were skilling their workforces for AI, and what were some of the opportunities that they were seeing, what were some of the challenges.

And again, I think some of the things that stood out were the fragmentation of the learning experience. Some people are far ahead, a lot of people don't know how to apply this at their job, so there's that unevenness within the organization's workforce just in general. And then I think organizations are often reinforcing that because of who they're giving people or who they're giving training access to and who has access to the best AI tools.

I think one of the things that was striking to me as we went through the process is how much of this is leaders picking the winners and losers of how AI affects people in their organizations, either implicitly or explicitly by saying, "These people get access to the top AI tools. These people don't."

I was talking to some leaders a couple months ago who were saying, "We're already starting to get people who are trying to apply for jobs in this organization, which is much more AI forward than some others, and we're having to turn them away because we're saying you don't have experience putting this into practice in an enterprise setting."

And so I think there is this compounding cycle of people either develop their AI skills relatively early or they're falling farther and farther behind in organizations because of the cost of tokens are having to make a lot more decisions about who has access to what tools and what they can or cannot do with them.

That is reinforced through the training opportunities that are or are not available to people because a lot of organizations are doling out training based on people's current roles instead of saying, "Where are we trying to get to in the future? What are we trying to accomplish as an organization and what do we think we'll need in terms of our workforce capability six to 12 months from now?"

I'll give an example. Technical teams often have much deeper training available to them than your standard associate. The problem is a lot of the time, organizations are constricting who has access to, let's say, an agentic AI system versus your basic chatbots, and that's often for security reasons or concerns about controlling what agents are doing in the organization.

But it leads to a situation where people are just fundamentally not developing core skills that are going to be essential for knowledge workers moving forward. And because the organization has chosen to limit their training opportunities, those people are fundamentally handicapped in their development.

Diana Scott: Yeah, and that creates quite a dilemma when you think about going forward. You're likely to have then potentially shortages of access to the kind of skilled workers that you're going to need and those people are not going to have the opportunities that they want to have. So it's really creating quite a dilemma.

The report really does introduce a very interesting idea of what you call an enterprise-wide learning and adoption ecosystem. I'd like you to talk a little bit about that because I think it is interesting, and you introduce this concept of some essential components of that ecosystem and you talk a little bit about much more broader ownership than just HR or learning and development alone, and why that is.

Can you just describe this ecosystem, talk about the ownership and why this is really essential across the organization? Because I think it's a really interesting concept.

Matt Rosenbaum: Yeah. So what struck us as we went through the interviews with enterprise leaders talking about AI is how many organizations were thinking about learning primarily as the learning opportunity itself, whether that be the formal program, we watch this video, we have people go through this course, they have some basic AI training that they have to complete before they get access to a tool, whatever it might be.

Some organizations are also layering in other key components. The social aspect is especially important. Seeing your peers use AI, understanding how they are using these tools to do the work that you have to do together is always a huge driver, and then also getting people hands-on experience. So sandboxes or opportunities for people to develop skills in that sort of way are incredibly valuable.

But then a lot of the time, organizations stop there. They don't necessarily think about all the ways that they are either reinforcing or undermining what people are learning in those learning opportunities, and I think that is honestly one of the most important levers. It's not easy to pull but it is something that if you can get the rest of the organization to support what you're trying to teach people in those training sessions, you're going to get a lot more bang for your buck.

I'll give an example. A lot of organizations do not yet consider AI skills in their promotion decisions. There is a verbal message or whatever it may be of, "AI is important. You should learn AI and here's all these little programs that we have available for you." But until they reinforce that by saying, "And also that's tied into who gets promoted or who doesn't," that it's tied into leadership pipelines.

That is going to undermine that message of the importance of AI, the importance of developing those skills. Another thing that stood out was we're talking to a large insurance company that said one of the most important things that they did was starting to emphasize growth mindset in the people that they were hiring.

So yes, you want to develop skills in your current workforce but also reinforce that through the people you're recruiting and what you're seeking there. And the person I was talking to said that has dramatically altered the environment and the culture of the organization in a relatively short time.

Diana Scott: The cultural issue is around, you want to have this sort of continuous learning, agility, openness, et cetera.

Matt Rosenbaum: Yeah. Yeah. It's all these other components that are not, quote-unquote, "part of the learning."

Diana Scott: It's not a technical thing, "program." Yeah. It's a mindset.

Matt Rosenbaum: Mindset, the ownership or lack thereof, the feeling of people saying, "I can experiment and test things and work in new ways," or being content to work in the way that they're used to working because they don't really see anyone else around them changing things either.

Diana Scott: It's not just the individual. It's the organization also has to be open to that mindset of test and learn and fail fast, et cetera. The organization itself.

Matt Rosenbaum: If the individual has that mindset and the organization is not open to it, that individual is going to be gone very quickly.

Diana Scott: So culture becomes very important all around.

Matt Rosenbaum: Yeah. No, culture is, I would say, one of the essential enablers, or it can either help or hinder. A lot of organizations have gotten to where they are because they've been relatively risk averse, very buttoned up, very centralized, process driven, whatever it may be, and that is antithetical to some of the behaviors that you want to drive in terms of adopting AI, experimenting with new ways to work, things like that, and you're seeing that clash play out in a lot of organizations.

Diana Scott: This is all really interesting. We're going to take a short break and we'll be right back with more of my conversation with Matt Rosenbaum.

Welcome back to C-Suite Perspectives. I'm your host, Diana Scott, Human Capital Center Leader of The Conference Board. I'm joined by Matt Rosenbaum, Principal Researcher in the Human Capital Center at The Conference Board.

So Matt, we've been talking about AI skilling and perhaps the most striking finding is that most organizations are focused on upskilling today and you talked a little bit about that. As role redesign and reduced hiring and workforce disruption are already beginning, what do you think CHROs should be doing now to prepare for the much larger reskilling challenge over the next two to three years?

Matt Rosenbaum: I think first and foremost, we need to make the case for why it's important to focus on reskilling because a lot of organizations are not. I think out of all the organizations we spoke with, only one was concentrating their skilling efforts on people who they thought were going to be left without a seat at the end of the day unless they were reskilled and redeployed. The vast majority of people are focused on how do we get people to use current AI capabilities to achieve their current workload and get through that.

The challenge is it's a much more involved need to reskill people and redeploy them than it is to upskill them in their current roles and I think a lot of organizations just do not have the infrastructure that they need in order to support that. So it's important to start now.

A CHRO I was talking to a few months ago said, "Do we even have a problem?" The unemployment rate here in the US is pretty flat. We're not really seeing this widespread AI job loss that might have been predicted, so how important is this in the first place? And I think the challenge is by the time it shows up in the numbers, you're already too late if you want to start to ameliorate a lot of the pain and suffering that could be involved in this or doesn't have to be.

The good news is by developing a strong learning and adoption ecosystem, building that continuous learning culture, all these things we've been talking about, that is going to help get people ready to reskill significant portions of their workforce. But I think the other thing that is important to recognize is a lot of organizations are not necessarily clear on the business case for reskilling in the first place.

It is very clear to a CFO or anyone in a similar position to understand why reducing head count is going to lower costs. It is not clear what value someone is going to get from embarking on an uncertain reskilling program in order to shift people over to other parts of the organization that may be new or are growing.

So that is the apparatus that needs to be developed in the first place. The infrastructure for that redeployment maybe exists in a lot of organizations but often does not. Things like the redeployment of talent require addressing the fact that a lot of managers are going to want to hoard talent, especially if they're top talent, for instance, or AI-skilled talent.

So you need to address that. You can't wait to address that until you have a redeployment problem on your hands. But you also need to develop that business case and be clear as an organization about this is how we do or do not approach reskilling our people and redeploying them because there's a lot of consternation. People can see writing on the wall of organizations that are not necessarily going to want to keep everyone that they have on board.

Some organizations I've talked to are taking a proactive stance on that of we're going to outskill people, we're going to help them develop skills that are going to be needed in the marketplace, even if we don't necessarily have a place for them here. But there's going to be that general fear and concern unless organizations start to put forth the effort to say, "We're going to reskill you and here's how we're going to actually make that promise come to pass."

Diana Scott: In the research, you talked a little bit about the impact that transparent communication has on the employees in terms of their building confidence with them. If they feel that they're being communicated to, that they can trust what the employer is saying and what the employer is committing to do, that actually creates a better environment. There's greater trust, there's more likelihood that they will be more engaged. Perhaps it improves retention longer term.

Talk a little bit about that because I thought that was an interesting point that you were making and that you saw coming out of the research.

Matt Rosenbaum: We know that people's attitudes based on their view of did AI affect my career confidence in a positive way or in a negative way is correlated with whether or not they have strong engagement, strong intent to stay at their organizations.

The folks who tend to have a dimmer view of how AI has affected their career prospects are also more likely to leave, they're more likely to be unengaged, they're more likely to not be putting forth their best effort in their current jobs. And some organizations have still chosen to take the path of we're going to pretend that everything's hunky-dory and in the back rooms of the organization, we're making decisions about who do we decide to let go and whatnot in the future, or who do we anticipate that happening to.

But I think the organizations that are doing this in an admirable way are being much more transparent about that process. Everyone knows it's happening. It's foolish to pretend that it is not. So be upfront about it and acknowledge this is what we're going to do to try to help. This is not a tomorrow thing. It's a year or two down the line, there's going to likely be some people who are going to be affected by our change in organizational structure or strategy or whatever it may be, and here's all the things we're doing in the meantime to help develop you all to ensure that those of you who can will have a place here and those of you who do not will be as best suited as you can be to find a job elsewhere.

Diana Scott: So you're still investing in and preparing them for the skills that they might need, whether or not they're at the current organization. You're making that commitment to them.

Matt Rosenbaum: Exactly.

Diana Scott: Given all of that, if you were advising a CEO and CHRO whose organization is scaling AI and has rolled out some of these foundational AI training to their workforce, what would you say are the three leadership decisions that they need to make over the next 12 months to ensure that AI becomes a competitive advantage rather than simply just another technology investment for them?

Matt Rosenbaum: Yeah. I think a few things that are going to make that difference or at least help get over the hump. One is that alignment that comes from senior leaders about this is what we're trying to accomplish with AI, this is where we're going to invest, this is what we want to be in five years, and how AI helps us get there.

No one else in the organization can make that decision and that is going to guide where you focus your training efforts, how you develop people, who you think is the right target. Is it focused on current roles or where you think the organization is going to be going in the near future?

All of that flows from that kind of leadership decision and vision. So that is, if you do not have that fundamental grounding, you need to start there. The next thing is going to be giving people training that meets their current needs, yes, but it's much more future-focused. So not tying people down into the roles that they currently have but being much more understanding of we're going to have people moving around the organization, and it's going to look different in a few months, in a few years than it is right now, so how do we start preparing people for that world?

To the point that we've been talking a little bit about already, if you're emphasizing the prompt and the prompt structure, like, that is not where people should be focusing. If you're talking about digital acumen or business acumen or analytical acumen, that's where you need to start. The things that you need are people who can understand the business, understand how the organization as a whole fits together in order to say, "This is where we're going to derive value from using AI," versus, "This is where it's not going to be as useful." And that is much more important than a prompt but oftentimes that is not necessarily a focus of AI training.

So focus the training on actually producing value and understanding what are we trying to accomplish as an organization and how does AI fit into that, and then how do we create a workforce that can get us where we need to go.

And then the last bit here is outside of that formal training, how does the rest of the organization either support or hinder the development of those skills and those capabilities? So like we talked about with the recruiting or with the promotion decisions, how are other facets of the organization and how employees experience the organization going to either reinforce what they're supposedly learning in the message that you're sending about AI or going to undermine that?

And that, I think, is going to be critical moving forward just in terms of helping people understand not only are we saying this is important but this is actually important and we're showing it in our actions.

Diana Scott: Great. Excellent advice, and thank you so much for joining me today.

Matt Rosenbaum: Always a pleasure to be here.

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