Why is employee AI use rising so quickly while measurable business value remains difficult for many organizations to find?
In this episode of Tech Talks Daily, I speak with David Martin, Senior Partner and Global Leader of People and Organization at BCG, about the firm's fourth annual AI and workforce report. The research surveyed 11,749 employees across 14 countries and points to a growing divide between companies that distribute AI tools and companies that give people a clear plan for changing how work gets done.
BCG reports that 74 percent of frontline and nonmanagerial employees now use AI regularly, an increase of 23 percentage points from the previous year. Adoption, however, is only part of the story. The report says 71 percent of employees receive little or no guidance about what to do with the time AI frees, while over half are not redirecting that capacity into strategic work.
David describes one of the research findings that best captures the problem. Companies where employees understand the strategic direction but rate their AI tools poorly can realize greater value than organizations with strong tools and limited strategic clarity. Better technology helps, but its impact remains small when employees do not understand which business problem they are solving or how the operating model should change.
The report connects clearer strategy with a roughly 25 percentage point increase in measurable business impact when companies redesign workflows from end to end or create new business models. BCG says strong tools without that clarity produce an improvement of roughly five percentage points. Companies that redesign workflows also outperform tool-only adopters by 23 percentage points on measurable business impact, 22 points on time saved, and 20 points on job satisfaction.
David explains what redesign looks like in practice. Giving software engineers stronger coding tools may improve part of a development task, but keeping the overall product lifecycle unchanged limits the result. A deeper redesign considers how research, product management, engineering, and decision-making operate together, then changes roles and processes around the capability of AI. The objective is a better business outcome rather than a faster version of the same work.
This distinction also explains why promising pilots fail when companies attempt to expand them. A pilot can prove that a model works inside a controlled environment. Wider deployment tests whether the organization surrounding that model works. David cites BCG research indicating that 70 percent of the factors determining whether AI scales with a return relate to people, organization, and process. Talent, operating models, cross-functional teamwork, incentives, learning, and leadership all become part of the result.
Measurement must also move beyond adoption. David argues that the final metrics remain familiar business outcomes such as conversion, competitive win rate, price realization, cycle time, and inventory performance. A pilot can use controlled comparison to test whether AI changes one of those outcomes. The missing management step is often accountability. If several executives share ownership but nobody is responsible for the return, the investment can continue without a clear test of success.
There is a case for broad experimentation because it can build familiarity and surface ideas. David warns that hundreds of isolated use cases can also fragment investment, increase risk, and save small amounts of individual time without producing company-level value. His preferred balance combines focused governance with structured opportunities such as hackathons, where employees contribute ideas but the organization selects which ones receive investment.
The workforce findings add an important human dimension. BCG says 67 percent of regular AI users report higher job satisfaction, while 41 percent also report higher cognitive load. David connects that tension with the effort required to assign work to agents, evaluate quality, and keep those agents operating. He refers to separate BCG research called AI Brain Fry, which found productivity rising as employees managed additional agents until a limiting point. In that research, productivity fell when workers moved beyond managing three agents.
Training remains another stubborn problem. The report says 72 percent of employees believe AI has changed skill expectations, and nearly half say their role is moving toward directing and managing AI. Only 36 percent feel they have received enough training, a figure David says has not improved despite new learning programs. His recommendation is in-context training that brings AI into daily work, followed by peer discussion about what worked, what failed, and how behavior should change.
The purpose of recovered time may be the most revealing management question of all. David says employees report saving an average of around eight hours a week, but many use that time to perform additional versions of the same tasks. That can become demoralizing if greater output benefits the company without giving employees room for learning, infrastructure improvement, experimentation, or new product work. Leaders need to explain where the capacity should go and why.
Clear communication also reduces fear. Automation targets introduced without an explanation of strategy can leave employees assuming that efficiency is a code word for job loss. When leaders explain whether AI is intended to improve customer experience, create growth, reduce cost, or change the business model, employees have a better basis for understanding what is expected of them.
If strategic clarity is producing greater value than better tools, should the next AI investment begin with another platform or with a decision about how the work itself must change? Listen to the episode and share your thoughts.
Useful Links
[00:00:00] The leading issue of agentic AI in businesses right now is ensuring agents act with compliance guidelines. And Denodo applies guardrails across your entire data estate. By aligning your company's data infrastructure under one system, these guardrails perform consistently across your platform. So start scaling your business and start with Denodo. Simply visit denodo.com to learn more.
[00:00:27] Why are employees using AI far faster than companies are turning their activity into business value? Well, today in this episode of Tech Talks Daily, I'm going to be welcoming back Hedda Barber-Dolson, senior partner and global leader of people and organization at BCG.
[00:00:53] We've spoken several times in the past, but today we're going to be looking at the findings from the firm's fourth annual AI and workforce report, where they surveyed 11,749 employees across 14 countries and found that regular frontline AI use is rising sharply. While guidance on how work should change remain limited,
[00:01:17] David will join us again to explain why strategic clarity can matter much more than just better tools, why successful pilots can often break when they meet the wider organization, and why employees need a purposeful plan for the hours that AI gives back. And we will also discuss workflow redesign, accountability, training inside daily work, and the cognitive burden created when one person starts managing too many agents.
[00:01:47] But enough from me. Let me introduce you to him now. So thank you for joining me on the show today, David. Can you tell everyone listening a little about who you are and what you do? Yeah, sure. And thanks for having me. My name is David Martin. I'm a senior partner at BCG, and I lead our people and organization business unit globally. So that is all of BCG's work on topics like organizational design, talent, culture, and change management.
[00:02:14] And in that capacity, I'm also a part of our global AI leadership team, which oversees all of our AI-related work globally. Well, over the last decade, I've had many guests from BCG on now, always have fantastic and engaging conversations with some pretty big stats in there as well. And before you join me today, I was reading BCG's latest research suggests AI adoption is actually rising faster than business value,
[00:02:41] which maybe is a familiar tale for a few people listening. But what is it that's separating companies with a strategy from those just collecting more tools? Yeah, so there was a funny individual stat as part of that research that actually showed companies where the employees say they understand the strategic direction of the company
[00:03:03] and have horrible AI tools are actually realizing more value from their AI investments than companies where employees say, we have great AI tools, but we don't have strategic clarity. Wow. And so I do think that this notion of focus is critical to driving business value versus just adoption. The best companies are infusing AI into the overall question on strategy.
[00:03:32] Where can AI help me drive better competitive advantage? And then setting up focused initiatives against those core parts of the business and being very focused. I think the companies who have said, let's deploy AI as broadly as we can across the organization, targeting hundreds of individual use cases, have really struggled to see the value come from those investments. And just to bring to life what we're talking about here, I don't expect you to name any names at all,
[00:04:01] but can you share an example of a company or client that has changed its operating model rather than simply adding AI into an existing process? Because many of those existing processes were designed for a completely different era. I'm just curious if you've heard any positive stories around that. Yeah, it's a great question because those companies with focus are thinking about their processes end-to-end. And so there's a lot of examples in the technology industry,
[00:04:29] and I've been working with a few of the major hyperscalers around how to completely rethink the product development lifecycle. And so whereas some companies who are struggling are giving engineers the greatest and latest AI tools to help their software development process, and we've seen some of those be highly effective, if they keep the overall product cycle the exact same, then they're seeing very small amounts of productivity improvement.
[00:04:56] The hyperscalers have said, how can we take an eight-person process that goes from market researchers, the product managers, and full-stack engineers, et cetera, down the chain? And can we turn that into three people who are doing many different things using slightly broader-scoped agents to improve our process? And so they've seen productivity improvements three, four times increased relative to what they had beforehand.
[00:05:24] I mean, there's also interesting companies and publicly known instances like a SpaceX or I think Jack Dorsey had a controversial tweet from Block around both kind of establishing this. So Jack's tweet talked about this. And then SpaceX is like rethinking their model completely, right? It doesn't go just end-to-end processes, but they're saying, you know, how does AI change our business model fundamentally, and how do we get different parts of our business actually more tightly integrated?
[00:05:54] And there's been a lot of noise over the last few years around getting return on investment from expensive AI projects. But, of course, the mantra in IT has not changed. It's been that way for as long as I can remember, and that is you can only improve what you measure. But the question is, I guess, is what measures are actually best showing whether an AI investment is improving things like revenue, speed, or customer outcomes? What are the metrics that you're seeing people focus on here?
[00:06:22] It's certainly not just AI adoption and what percent of my employees are using the tools. And it's not token maxing or even saying how many tokens am I consuming and, you know, the higher the better. I think maybe a boring answer to that, but the right one is they're not changing their, you know, top-level KPIs. They're looking at those as the ultimate measure of value. So if, you know, you're looking at reimagining a sales and marketing process,
[00:06:49] you're looking at conversion rates, competitive win rate, price realization. Has that improved or not? And I think as companies roll out these new solutions, normally they're piloting it, you know, with the intention of scale. And so the benefits of pilots are you can actually see, you know, an A-B testing as an example of, am I seeing incremental change to those core business metrics? Is my conversion increasing if I'm, you know, doing something in my supply chain? Is my cycle time improving?
[00:07:17] Are my inventory levels being more efficiently managed? So there are the fundamental business outcomes. I'd say the one additional add to that as it relates to metrics is companies thinking about AI have not put in place really well-established accountability for those metrics. And so it's shared across multiple leaders of the company. And I think being able to establish, okay, who's ultimately accountable for driving this return?
[00:07:44] And then are we measuring that value and holding those leaders accountable is a very simple step that I think a lot of companies have missed to this point. And another big phrase we've heard over the last few years is pilot purgatory. So why is it that many pilots look so successful in a small team, but are so often struggle when they reach the wider organization? Is it the bias within that team or is it something else?
[00:08:09] Yeah, so pilots have done a good job of testing whether the technology works in a controlled environment and actually scaling tests whether the system around the model works. And so our research has consistently shown the ultimate drivers of whether or not companies are scaling this with return. 70% of that is on the people, organization, and processes components. So do I have enough talent in place?
[00:08:35] Is my operating model set up to empower leaders and support more cross-functional teaming? And pilots aren't able to capture that. I think talent scarcity is also a big piece of that. I just alluded to it quickly, but to build on that, pilots normally take the best individuals in the company and put them in one place. And one thing that really inhibits scale is companies just don't have enough AI fluent, AI capable talent right now.
[00:09:00] So it's very difficult to capture scale without the right people on the team. And AI and the shiny next big thing in technology have always been the focus of many business leaders. But a question I've got to ask is, where do leaders underestimate the work required around things like data, process ownership, and employee behavior? These things still seem to continuously keep going over the radar.
[00:09:28] Yeah, and those are the things that get in the way. I think one point of failure that companies have had is they're viewing those three things in isolation. So you might have a CIO or chief AI officer on data, and you might have the business lead of whatever use case you're deploying, owning the process. And then you have someone in HR who's probably not properly empowered based on our research, who's thinking about employee behavior.
[00:09:54] And unless you're an incredibly mature company with great AI talent across the board, companies have really had to do a more purposeful job of building what we would call an AI hub. And that's where those leaders come together in kind of a centralized way and manage those three things in concert with each other. So the data investments are directly related to how do they support process and those leaders working together on making sure they're utilized properly.
[00:10:24] And employee behavior, like it's interesting because AI adoption, even though more abundant than return on investment has still been something companies have struggled with, especially in some of those more knowledge work jobs like software engineering or marketing. And the drivers of adoption actually span. I mean, it's human behavior related mostly. It's how do you motivate the employees? How do you incentivize them?
[00:10:49] How do you change the training model and learning and development to be more in context learning? And because those have been fairly consistent across functions, the employee behavior side of it, you really get some good scale and impact by doing that from a centralized AI hub and making sure that those best practices are deployed with each of your AI deployments and that you're not trying to handle those in a piecemeal way.
[00:11:15] So I think just tighter coordination on those three things has been proven to be highly successful. And I think traditionally there are some executives, especially in younger startups and SMBs that favor broad experimentation, especially around things like AI. And it helps them move quicker, while others in maybe mature enterprises, they want a tightly governed portfolio. And I guess the truth is somewhere in the middle, but what are the trade-offs in either of these?
[00:11:43] Look, so broad experimentation and even the decentralization and democratization of digital tool usage has been a trend for 15 years across companies.
[00:12:22] And that was great. But you can't realize tangible business value from that unless you're finding ways to better restructure the work to unlock productivity. And so experimentation, while fostering creativity and helping to build AI fluency, so really important, needs to be balanced, to your point, with focus. And I think companies who have been focused.
[00:12:46] Actually, we see the ones who have been leading on AI return on investment have done fewer things and more governed from the center, which was kind of a fascinating result. So centralized governance, I think, is really important. It's also important from a risk standpoint. I think companies have probably dangerously pushed experimentation at the edges at the risk of losing intellectual property or cybersecurity, introducing cybersecurity risks.
[00:13:15] So I think just the risk that comes with AI also would point toward more tightly governed experimentation from the center being a more effective way to realize value and reduce risk.
[00:13:28] You know, one thing that folks have done to help still capture some of that creativity at the edges is host things like hackathons and to try to get the broad enterprise engaged on AI ideation, but not necessarily to go invest in all of those different ideas. When you do a hackathon, there's a winner. You don't let every member of the hackathon, you know, have an unlimited token budget and go pursue every idea that came up. And I think that would be a watch out for companies.
[00:13:57] A hundred percent with you. And also, if we have a quick scroll down our LinkedIn news feeds, we'll see a lot of AI theater from enterprises there. So how do you see or how should a board challenge some of those ambitious AI claims that the maybe marketing team are posting without slowing the useful work to a whole day-to-day stuff that really moves things?
[00:14:20] Yeah, well, so back to the metrics point earlier, there's one kind of lucky benefit, which is you're ultimately going to measure this value from investment by your core business metrics. And so the board can't lose sight. And I've seen very few boards lose sight of holding the company accountable to hit their, you know, level one KPIs. I think boards who have behaved badly either have been too AI illiterate.
[00:14:45] Actually, it's funny because 84% of company employees, whether they be leaders or frontline employees, have stated that the company would need substantial process change to deliver on AI investment. Boards, on the other hand, 42%, I think was the stat, actually suggested it would need that magnitude of change. So boards can build their own literacy.
[00:15:09] I think some bad board behavior gets into approving or rejecting individual use cases. I don't think you can micromanage it. I do think focusing on those business outcomes is really important. But I think the other place that they need to be pushing CEOs and the executive leadership team is to make sure they have the right leaders with the right culture in place.
[00:15:30] We've published a few pieces, maybe thought papers on, you know, what a CHRO of the future looks like and a CIO looks like. And I think that change needs to happen from top to bottom and thinking about the leadership and do they both support culturally what's needed to deliver on AI, as well as, you know, do they have the right skills and capabilities in place? I think the board could do a good job of assessing senior leadership talent and perspective.
[00:15:59] The last one, back to metrics and maybe the board level one is, you know, again, we talk about token maxing. I think a lot of boards are confused and a lot of senior leadership teams are confused with how to measure return on tokens. And we have seen a new metric emerge that's interesting to me, which is thinking about it as return on intelligence, which is the combination of your AI technology investments, including tokens, with the labor related to those processes that you're tackling.
[00:16:28] And to look at it in a more unified way than just looking at how are these AI investments decreasing this type of labor costs. I think that's a very large watch out and boards can hold leadership teams accountable for return on intelligence as well. Oh, wow. That certainly sounds like something worth keeping a close eye on. And this year, of course, everything's being around agentic AI and agent integration into workflows more than doubled since 2025.
[00:16:56] I think it was another stat in the report that 61% believe agents could actually do half their job within the next three years. But governance and accountability still lag far behind the tech, of course. Anything, any takeaways for you around all things agentic AI and AI agents? What did you take away from some of the findings there? There's a lot of fun data points in there.
[00:17:18] One interesting one, I think, that might dispel some rumors or myths is we saw employees who were using agents in their workflow to have 67% job satisfaction, which was higher than those who weren't using agents. And so you did see satisfaction go up.
[00:17:37] At the same time, which I think is the big watch out with agents, is those same employees also, 41% of those even happy employees talked about an increase on their cognitive load. So more challenging or stimulating intellectually and more demanding. And a lot of that is because of the ability to distribute proper attention on, am I giving the agents enough work?
[00:18:05] Am I able to accurately assess their quality? And can I keep the agents running? And we had a good paper earlier this year that was labeled AI brain fry. And it talked about, depending on the number of agents, the employees were overseeing the increase in cognitive load. And the increase in productivity up to a limiting point.
[00:18:28] And I think after they were managing three agents, once they got into four plus, their productivity actually decreased and significant relationship to cognitive load. So I think agents are good. Agents can help take the toil out of an employee's job, which is great and can lead to satisfaction. But you have to watch out the demands it's putting on the cognitive load for those employees. Yeah, a couple of other stats I remember from there. One was, I think, 72% of employees say AI has almost changed their skill expectations.
[00:18:57] And nearly half are saying their role has shifted more towards managing and directing AI rather than doing the work itself. But even with only 36% feeling they've had sufficient training. I'm curious, when you look at the whole report, was there anything in there that surprised you? Because obviously this isn't your first rodeo. You see these kind of stats all the time. Did anything take you by surprise? It's, I don't know if I'm surprised because I see it with my clients.
[00:19:24] And I saw it to a point with BCG, but I think we've actually cracked the code now, which I've been excited about. It's sad because that 36% number, which is around, do they feel like they've had sufficient training? Or that there are sufficient training programs in place to actually address the amount of disruption that's coming up? That stat has not changed.
[00:19:48] And I think one sad part about that is everyone at the company knows that skills are going to have to change. And a lot of companies have even put in a substantial amount of new training curriculum. The L&D department for HR has been incredibly stretched to try to think of all the different ways to improve AI fluency and familiarity. And yet the number hasn't changed.
[00:20:13] And it's because we see this day-to-day that training in a classroom and even training in demos and training and kind of go code your own agent in a controlled environment has not been a successful driver of better use of AI. And what is necessary is what we call in-context training.
[00:20:37] It is taking AI into your day-to-day work and having purposeful focus on for each thing you're doing that day, can or how would AI potentially help support this? And then spending time with peers later in the day talking about it.
[00:20:55] So companies have done four hours in the morning of take your work and think about it in an AI-first way, four hours in the afternoon talking to your teammates about what worked and what didn't, and doing that over two to four weeks. Because habit building is an important part of this too. It's not just about knowledge. And the amount of improvement companies see in that have HR departments wondering, should I be doing this in an always-on way?
[00:21:21] Like, is this not just a training curriculum, but this is literally the way the job changes because of how fast the AI tools and capabilities are changing? And not enough companies are doing that or thinking about that in a good enough way. And consequently, you see that 36% number, despite investments and despite the awareness of that, just not move. And that's sad. It really is.
[00:21:45] And as this is a human-focused enterprise tech podcast, are there any warnings around what could happen to employees when maybe that strategy is unclear, but automation targets continue to grow? Any observations or warnings there? Yeah, well, so, you know, you see also in the report, and this has been consistent, that employee fear is at an incredible high. And, you know, there's some obvious fear of job loss.
[00:22:14] And companies who aren't communicating their strategy with their AI investments, whether it be to realize productivity savings through labor, or in many cases, companies are looking to drive improved customer experiences or improve revenue. And they don't tell their employees that, and so they're still scared. And leadership behavior has been huge, and communication of clear strategy has been huge to reducing employee fear. And then setting automation targets on top of that obviously only exacerbate the situation.
[00:22:42] I think the other piece about automation targets is we had a stat that was asking employees about how much time have you saved? And it was neat, because I think the average employee is saved something like eight hours a week, they're saying, in their jobs. And then you ask them, what have you done with those eight hours? I think number one is, well, I've done more of the same, which is incredibly demotivating. And I've heard that from software engineers, which is like, why am I going to use the tools to double my productivity if I'm just building twice as many things?
[00:23:11] Like, you know, the company's benefiting, but I'm not getting paid more. And that's demotivating. And so companies who have then been purposeful, even if it's not like business strategy or business model innovation, but the strategy around what to do with unlocked capacity needs to be clearly articulated. And so with engineers, it's like, spend 20% of your time improving your core infrastructure. Spend 20% of your time thinking about new products that you could offer and go tinker around. Spend 20% on learning and development.
[00:23:41] And so it's like, you need to be purposeful with how you're using the time you're freeing up or else it's demotivating. So increased fear, less motivation, not a good recipe. 100%. I think it's been 30 plus years now. Technology has been promising to give us all so much time back in our lives. But of course, as you said, we just end up doing more and faster. But so I will include links to the report that we've referenced today.
[00:24:09] But for all things BCG, people want to connect with you or your team. Where would you like me to point everyone listening? I think LinkedIn's great. We publish a lot there and especially links to our latest reports. That would be the place to go. And it's either me personally, David Martin, or we have a few channels through the BCG network as well. So I think we have a BCG on people and organization as one example. And we could follow up with more specific links if you want to include them in the post.
[00:24:39] And certainly appreciate any of the links that you want to put in about our research. We've been prolific this year in publishing, which I've been excited about. And there's a lot of neat factoids that we've covered some of. But as much as you can publish the links to the papers, that's appreciated well. And I think folks will enjoy some of those takeaways. Yeah, 100%. I will add all the links that you mentioned and a few others to the blog post associated with this at techtalksnetwork.com.
[00:25:05] And big takeaways from the one report we referenced was this isn't just another AI adoption story anymore. The company's doing best. They're not just giving employees more tools. They're actually giving them more clear direction on how work should actually change. A big, powerful moment to end on. But thank you for explaining all this in a language everyone can understand. Appreciate your time today. Perfect. Thank you so much. And thanks for all you do. Enjoyed it.
[00:25:30] I think one of the big takeaways today is AI value does not come from just distributing tools as widely as possible or celebrating rising token consumption. Companies need to decide where AI supports their strategy. Redesign the work around that decision. Assign accountability for the result. And start telling employees what to do with the capacity that they recover.
[00:25:57] And it's also important to highlight that human consequences matter here too. Yes, AI users can report higher job satisfaction while experiencing greater cognitive load at the same time. Especially when they are expected to supervise several agents without enough training or even direction. Remember, you can follow David and BCG's people and organisation work in the links in the show notes.
[00:26:24] But if AI returned eight hours to your team this week, I want you to be honest with me here. Has your organisation explained exactly how that time should be used? Let me know. TechTalksNetwork.com I've got about seven or eight different tech conferences I'll be going to before the end of the year. If you'd like to meet me in person, give me a shout. Other than that though, time for me to go now. So I will return again to your podcast feeds tomorrow. Bye for now.
[00:26:58] Bye for now.