The AI Daily Brief: Artificial Intelligence News and Analysis - BONUS EPISODE: Enterprise AI Trends and Predictions (You can with AI.)

Episode Date: August 16, 2025

NLW recently collaborated with KPMG on a 7-part enterprise AI-focused series called You can with AI. On this Saturday bonus preview, we share episode 7 of the series, all about the trends shaping the AI-ready organization of the future. Featuring Steve Chase, KMPG Global Head of AI and Digital Innovation.Learn more about the series: https://www.kpmg.us/aipodcasts

Hello, AI Daily Brief listeners. Today we have a little bit of a bonus episode. It is, of course, Saturday, which is normally the one day we take off every week. But this week, I'm excited to share one episode from the new You Can With AI series that I collaborated on with KPMG. Basically, the idea of the You Can With AI series is to create a primer for a full breadth of conversations around enterprise AI. That means we get into everything from data readiness to leadership changes, to how you build an ecosystem around AI, to what they were seeing in terms of different industries and how they're changing. There are seven episodes that run that gamut, all of which you can find at www.kpmg.com.com.com slash AI

podcasts. Obviously, that link will be in the show notes as well. But today I wanted to share one of those seven episodes, this one with KPMG AI leader Stephen Chase, where we zoom out and talk big picture about how AI is changing and what it means for the enterprise. So enjoy this special bonus episode of You Can With AI. And again, go to www.kpmg.us slash AI podcasts or subscribe to You Can With AI. on Spotify, Apple, or wherever you get your podcasts. Welcome back to you Ken with AI, a seven-part series on artificial intelligence from KPMG that is part exploration, part blueprint, and focused on helping you and your organization fully leverage the AI opportunity.

In this, our last episode, we zoom forward into the unknown to discuss the key trends and possible futures that will be shaped by AI and agents. To discuss the key trends, we're joined by Stephen, Steve Chase, vice chair of KPMG's artificial intelligence and digital innovation organization. All right, Steve, welcome to You Can with AI. You're closing us out. I'm excited to be here, excited to do it. I mean, listen, this has been a great series, I think, you know, as we were conceptualizing

this, we wanted it to be something for big conversations for people who are very much in the thick of AI and agentic transformation right now, but also a place for people to start who are just trying to wrap their heads around this. And I think across the course of the conversations that we've had, we've had a chance to sort of, you know, both go broad and go deep. But this is our chance to kind of zoom out, contextualize ourselves in the market that we're in and look forward. And when we say look forward here, we're not talking about, you know, a whole bunch of predictions, although I'm sure we'll do some amount of that prognosticating. It's more about what are the questions that we, that enterprises, that leaders should be asking now and over the course of the next, you know, a few months. which is obviously an eternity in AI time to help ground, you know, real strategy and thinking

through the challenges and opportunities that face us. So as we dive in, I would love for you to just provide a little bit of the day-to-day experience that you're bringing into this conversation for people who aren't familiar. Yeah, sure. So Steve Chase, as you mentioned, I'm the vice chair of AI and digital innovation at KPMG. And what that means is I run our enterprise innovation team. And in particular, one of the things I'm running is our AI program to bring AI into everything we do from our services to how we sell services, how we build those services, how we deliver

them, and also just how we run the firm. And I'm part of the global organization that does that and run that globally as well. Yeah. So you have, I think, makes sense that we're doing the zoom out with you. You sort of inherently have a context to take this sort of zoomed out view and, you know, that spans a lot of different parts of this transformation. I guess with that in mind, let's start by trying to understand, you know, where we are now and how what organizations are experiencing, what they're thinking about has changed, call even in the last six months, you know, what's the state of conversations today that are different or the same as they were heading into 2025?

It's a little cliched to say right now by this point in the year, but it's still very much about like people start wrestling with what agents mean. We have something we pass around. It's Jan Brady, instead of saying Marsha, Marsha, Marcia. she's saying agents, agents, agents, because, you know, it's just such a part of the dialogue right now. But there's also a lot of lack of understanding about what that means, where that is, et cetera. And we are seeing this breakaway between the leading organizations that are, you know,

they're piloting these things, are starting to get those into production, they have spent a lot of time preparing for that. And the ones that are still, you know, they look at their populations and they don't even have AI in the hands of a bunch of users. So, you know, I think we talked about before this idea that, like, the leaders are starting to break away from and materially break away from the laggards. This is like jagged frontier, if you will, of how invested are folks in this? And it's interesting because no executive executive has any doubt that this is the most transformative technology that they'll ever, that they'll bring into their enterprises.

So what leads to that? I don't know. But in terms of where the dialogue is with the leaders, it's about what's my agent strategy look like right now out of that, sit into my business strategy. What am I trying to accomplish with that? And how am I bringing my people along to have that mindset to be able to think with AI first mindset and AI first, including the scale opportunity with agents. Because really, I think about this a lot. Agents is this opportunity to bring intelligence at scale horizontally across the organization to accomplish a variety of objectives that might not have been possible before. Yeah, I think that one of the things that the introduction of the agent conversation, as perhaps confused or ill defined as some parts of it may be

brings, is it moves the AI discourse from one that is strictly about productivity enhancement to something that is much more structural, fundamental, that's more about business redesign, new product or service design, new business, you know, revenue opportunity. And it feels as though it just shifts the tenor of the conversation almost definitionally in some ways. I agree entirely with that because what we're talking about during this phase is rewriting fundamentally how businesses operate potentially in the way that maybe we were doing with the internet. But even more so because it's not just how supply chains work or how customers get access to products or whatnot. It feels even more fundamental than that potentially. We're having legitimate conversations about, hey, do we have the right board charts right now?

Like, is this going to survive this? I do want to step back on the productivity thing because actually I think there's a piece there too, which is we're really unlocking some of the most value, potentially really unlocking some of those valuable employees in an organization. And that productivity gain isn't a little bit or incremental. It's like massive gains for those, you know, what we call those 10x employees might be 100x employees as well. So you might not have called that productivity, but I know, I think there's something in that we can explore. And this isn't the kind of thinking that anybody's really.

paid to do very often is to, hey, I need to rethink like how we work, what the actual business model of the firm. Will it survive in its current context? Not, hey, how do I make finance? How do I make finance more productive? How do I think about generating my month in close quicker? We're really talking about, I don't want to be cliched, but like, how does finance procurement operations come together and to make potentially just a different kind of an organization altogether to deliver organizational goals. Yeah. So this is really interesting, and I want to pick up on this thread. This is something that you and I have talked about offline separately, the potential for almost overcorrection to just considering agents

and sort of, you know, dropping the ball on sort of some of the more human aspects of this. Yeah. One thing that's been interesting watching a variety of different data sources, but, you know, the KPMG Pulse Survey is an example of this recently, is how quickly it feels like on the one hand, co-pilot usage is becoming table stakes, right? There was a huge jump that showed up in the Pulse survey between Q4 and Q1. I think it was 22 to 58 percent jump from, I think it was daily usage of these sort of copilot tools. And on the one hand, I guess an interpretation of that might be, hey, look, people are adapting naturally, you know, that's a trend line that's just going to continue. But based on what you're seeing and where, you know, you're sitting, what are the

conversations that people need to be having, how do they need to be thinking about the continued sort of need to support their people even as we move into the sort of the agentic era of AI? What are the sort of unresolved challenges that relate to the people and kind of human side of AI transformation? Great question. One of the things we talk about internally and one of the things we talk about with clients is that this, you know, developing an AI first mindset is really difficult, right? like what does that mean and helping people understand what that mean. We've seen some really interesting top-down messaging coming from certain CEOs and whatnot about what that might look like.

I think that's actually been kind of helpful to get into the dialogue, you know, this idea of don't take the traditional approach, think through how could I solve that with AI as even like a mindset shift to take. I certainly believe that is one of the things is if you don't ask yourself to include AI in your workflows. People are like, oh, you've got to have it materialize. in my workflow. When it materializes in your workflow, you probably, I might not need you to do that piece in the future, right? Like, I've heard you talk about like what agents are going to do. Like actually, that kind of thinking in your workflow, well, the people who can push the button on the agent, like I don't, that's not what that's not what value adds going to be. And that part. Like I need to be thinking about the part where I'm adding the value. And so, so I'm thinking about, okay, how could I do that thing with AI? And that includes the agents that can go do the deep. research for me or whatever the item is and I'm going to manage that and likely get myself ready to be managed by those sort of processes too. We don't like to think about it that way, but clearly to the extent that I'm managing a set of agents, there will be a set of agents managing some set of things that I need to deliver back to them. And we probably haven't quite hit this topic

very many times, but there's a group of people who are choosing not to use it. I don't know, like in every organization, there's whether it's 10% or some organization, 30, 40,000. 50% who even when faced with here's the tool, don't use it. And you can only run as fast possibly as the slowest member of the team. And I think we're trying to see some mind shift change on that front too, which is, hey, I got to focus on the leaders in the pack so that they can go further and faster. I mean, I'd suggest that we'll leave the others behind. But like at some point, you know, how backward compatible do we need to make the things that we're talking about?

I can't dumb down every we're moving on to what the future enterprise. looks like. And I'm super fortunate to be a KPMG where you have a lot of people who think that way, but, you know, organizations struggle with this kind of thinking. And I do believe in a world where the pace of change is getting faster the approaches to bringing the people along. I mean, you know, those are got to, those got to change as well. And I think that's why you, you started super intelligent in the first place, right? You did it flagged this issue. Yep. No, and listen, I think that you're correct to identify as well, that you're almost, starting to see a bit of a bit of a shift from leadership discourse around this. An example that I can think of recently is a Norway sovereign wealth fund made a bunch of headlines because they went from highly encouraging AIUs to it is just mandatory now, right? You flagged some of these CEO letters that have sort of provoke conversation where you're starting to see more companies say basically, you know, almost have soft hiring freezes. They're not saying we can't ever hire a person again. they're saying, we have to be confident that we can't do that thing with AI before we hire

for confident. That's fine. And you're seeing more and more of that. Again, if we're thinking about maybe how we are in the midst of an inflection point, this does feel, you know, the actual usage of these tools feels like one of the shifts from sort of optional to it's just part of the day to day now. Yeah, because you're going to use agents. Like those are going to show up and you're going to use them.

I mean, whether you'll use them well or not, that we can explore like what it takes to get like agentic thinking into the mix. but I also need to be using all the other things too, right? And it won't be, there's a, I saw it really interesting discussion around personal agency in a world of agents in AI. And like it meant I can't wait for it to come to me. I've got to be using it. People ask me all the time about what it means to be a top performer and be a top performing company. How do I do that?

And part of that is I need to show that agency. I need to drive it. What you just said about how companies are shifting, that's the conversation we're having as well is the couple of years. we've all been on running the big change programs. Those programs need to continue exist. The content in that is shifting, you know, to the more techniques that we're using to solve problems, the way that that's fitting into the, this is how we're doing those things, getting more consistency in it.

Agents certainly help with that, right? But, you know, we're talking about we're into new skilling, not, hey, this is an optional thing you should, you know, try out. We're not in the experimentation phase on that. Yeah. I think that in general, it feels like even when we're. still doing experiments. It feels like the experimentation phase, at least from an intentionality standpoint, has sort of gone, right? Like with agents, yes, all organizations are basically in some version

of piloting or experimenting with agents, but they're not doing it with the idea that maybe they won't use agents. It's just to figure out what they need to do to be prepared to deploy these things more broadly. And there is a lot of difference it feels in that in that intent shift. I think that's right. I mean, I was going to say on agents, you know, one of the challenges is The ecosystem around agents is still clarifying itself, right? So I do need to be. I think that explains a lot of like I got almost every, you know, was it 65% of companies now are experimenting in our Pulse survey.

Yeah, the full two thirds, yep. Yeah, have experiments running right now, POC is running right now, which you would assume many of those are going to go to production. The holdup into production, which we're working with plenty of companies on this topic, we have these things ourselves. some are pretty easy to bring through. The real game-changing stuff is a little more difficult. I don't think the control frameworks in general have been in place.

We work with a lot of companies on that. How do I feel comfortable with semi-autonomous agents moving around? How many of those do I really want? You know, it is like custom software that needs to be maintained. Everyone is almost like a product in of itself. So I don't know you spend a lot of time with clients on the same topic. So I think some of those things are holding up. It isn't up because they're going, oh, you know, this isn't going to work.

we're not going to do this. It's exactly what you said is we know we're going to do this. And I know I'm going to work with my enterprise software solutions on the agents, their delivery. And I need to be good at that. And I know I'm going to, I'm going to have certain things, but I need to be thinking through what that user experience is going to be like across a sort of heterogeneous set of capabilities. You know, I've got the new architectural components that what does MCP allow me to do? What, you know, so what is agent to agent going to open up? that's where we're seeing clients sort of trying to figure this out right now, not am I going to do this? Yeah, absolutely. And so it's interesting. This is actually perfectly segues to what I wanted to ask you next, which is on the one hand, there's some part of these, the sort of issues and things holding full deployment back are being raced to be solved by the sort of the market right now. You have agent infrastructure companies. You have the protocol standards that you just mentioned, A to A to A MCP that are sort of racing with extreme haste to, to sort of.

solve some of these questions. I think it is extremely notable that if you look at standards wars in the past, things like, you know, even email standards were decade-long battles between, you know, turfs and factions. And when it comes to pretty much everything so far for agent infrastructure, everyone has just piled on to whatever the leading standard was because it's just the opportunity to get to the other side of actually deploying these things is so much bigger than the the turf war of who created the standard to get there. But, you know, so that's the external side. But, you know, and you started to mention this, but what are, what are the types of

conversations that companies are going to, are having right now and are going to be having over the next six months to get them prepared for sort of more full deployment of agents? You know, we talked about the human side of that. You sort of have referenced maybe some of the governance issues. But, you know, what are the key things that people are working through, you know, as we speak? Well, there's a bunch of them. We talked about ecosystem and tech. making some key choices around that, leveraging the investments they've already made.

Being on a much older system is a real holdup right now. You know, because the agent capabilities aren't being deployed in those. You need to be in. So there's those things. Plus, you need to think through, do I have the right ecosystem and me be willing maybe to be a bit long on some ecosystem decisions like piloting things. And so there's that. On the human side, for many years, the part of a project that,

tends to get cut a little bit. You know, oh, we'll just do that ourselves. It's change management, training. You know, if you're rewriting business processes, you're going to spend more time in training. You need to be thinking about what the training organization is, what your partners around training, where the content's going to come from for there.

And I honestly think that may be, Nathaniel, I think that's still a high hurdle because those deployment costs are actually quite high. Everyone wants the silver bullet, but the silver bullet requires a set of things to be a silver bullet. Otherwise, it's going to look like, you know, I think there's a trough of disillusionment coming where people were like, oh, I didn't get any benefit at it. Yeah, surprise. You didn't do the things actually that we know are required and we actually cut that out of the budget, right?

I mean, so that I'm really concerned about. Data, I mean, and I'm not even going to call it data. Like, actually, I think data is, there's some really awesome solutions coming in the data space. And some of them are already there. Many probably people listening going, I've implemented some of it. It's gotten a lot. It's gotten a little bit easier. But there's still expensive investments. You know, if you're thinking about what am I going to do with AI? Data is the most expensive area. Knowledge is way more valuable right now than data in this like, what do I know? And this is often, you know, you put a chat bot in front of something.

You figure out like, wait a second, we don't even know what our policies are not as clear as we thought they were. No wonder, you know, we have poor results coming out of, for example, our customer engagement, because the people were training and we have a high turnover because they're frustrated. So definitely knowledge and investing in that is a key component. And knowing where you have good data and knowledge is a great place to be doing your POCs and your enterprise class. And there may be some places where you don't have that, but it just is super critical. You're going to have it.

So there's a few, Nathaniel, there's a ton more. I also think a clear vision about what you're trying to accomplish. One of the mistakes I see is, hey, there's this thing. let's light a bunch of brush fires in the organization and see what sticks. You know, like I see what really go. And I actually subscribe to that because I like the idea of distributed innovation with a way to harness that and then bring it back to the court. But without that and the vision for it, no one ever sees the opportunity to scale that.

You get trapped in these POCs that never sort of go anywhere. So I'm a strong believer of that too. Strong central vision. Mandates in many cases, I think you're going to end up being a part of it because that's how some organizations work, but other kinds of organizations, they operate differently, but strong central vision,

what are we trying to accomplish? Why? What's the time horizon? And a really strong, setting the expectation, very strong clock speed. It's a little bit longer, but I probably could go longer,

but those are a few of my thoughts. No, you know, it has me thinking, one of the sort of subtexts that I think is worth making explicit here is I think that you and I both have a sense that things are excelled,

not decreasing, which is hard to imagine in some ways because of how fast things have moved for the last couple years. But is that your belief? Is that what you're seeing? And how is that impacting how organizations have to be making these decisions? Yes. In a word, yes, it's happening faster. It was with a client last week where we were talking through this topic of how much is shipping in terms of real material capabilities over the course of like mapped in quarters who will be more than they probably have delivered to their users that really changed their day-to-day user experience, then they maybe did, this would have been like five or six years worth of stuff coming out, right?

And I mean, forgetting whether or not they can deploy all that, which they will be able to, they'd never actually ask people to consume that much change at any time ever. And they're looking into the future and they're going to be able to deliver more than that. And you can see it even with, listen, You can see it with the ecosystem partners. I mean, we're working with technology. I can't think since maybe like the late 90s that this much technology is starting to, we're working with it kind of at release date.

And, you know, it's pre announced and you're waiting for the features to show up so you can put them in. And this stuff is coming. I think you've had a podcast where you're talking about like models showing up daily and whatnot. Yes, that's happening. Like you can't even wait for their big, you know, like five conferences a year. Can't wait for them because this stuff just keeps shipping and it's too good to. wait for. Enterprises that are built to be stable and take advantage of opportunity are not usually built to take immediate advantage of this level of change occurring at this pace.

And not all of it do you need to do. Some of it's pretty critical. One of the things we've, you know, I think we've learned, everyone probably knows is how much drift happens when that much change is happening, like alignment drift can happen. I can't go test all that stuff manually if I put that much new things in, I'm going to have to have test harnesses, frameworks, et cetera, for the way that I'm doing my evaluations. And I can't just do it on the new stuff.

I've got to do it on all the old stuff too, right? So you're making the right now into legacy really quickly. I guess I'm going to say it that way. So all really difficult challenges, absolutely moving faster. And the things I think organizations can start to do is, how do I start to create an organization that thinks like this, that thinks on a different clock speed.

And what can I do with automation and agents themselves to actually manage the agents, the changes that agents and AI are going to drive into that ecosystem as well? And you don't hear as much of that. I mean, I think some real leaning organizations are doing that. You know, Nathaniel, if you're not in it and doing it,

you actually never really think about how important that is. If you are in it, you realize pretty quickly, oh, this is like, we're like the most important investments and I probably am underinvestment in it, you know? Yeah, no, I think that there's a natural, when you think about it as software, your decision set is which are the software that I use to replace what processes that I have currently. Whereas the implications of this really involve organizational change and new infrastructure

for adaptability. I mean, you know, even just by one small example, knowing the difference between the jump from Google Gemini 2.5 to 2.5 pro, how much do you have to care about that as opposed to the jump from 40 to 03, right? Like the 40 to 03 jump fundamentally unlocks completely new use cases. The jump from 2.5 to 2.5 pro optimized for coding a little bit more, right? These are two fundamentally different types of knowledge sets, but most organizations don't have a structure where they could even internalize that in any real way. Clearly. And so what's your, so you have to have a good signaling and research organization. We use a lab environment. in to continually bring that kind of insight back in and turn it out.

Because the other thing is the new models, they obviously cost more, they're more, you know, and I don't need them for every use case. And so the bias isn't that they're better for everything. They're better for some things. And so I was thinking about that is also an important piece. But you don't want to overthink that because the way it works is it's expensive for a little while and then the prices come way down really quickly, which is a, I mean, don't you feel like that's been like one of the really interesting changes?

I mean, how fast price declines have occurred, which have then made us have a completely different mindset towards this, which, you know, there were entire companies set up to do, like, oh, we're going to have to do all this metering and figuring out with us local. Right. Who's doing that now? Yes, I need to be worried about it in the beginning. But like that, like, ultimately these things are all, I don't want to say they're headed towards zero, but they're, they're, that's what ends up happening.

So, I don't know, it's a really like, and again, do I want my, every one of my users trying to figure out? No, I need, I mean, my systems need to figure that out for them, too. I need that insight into what they're trying to do. It needs to use the right models. And that's one of the things that we're really building into the agents that we're building and whatnot is that some of that situational awareness. Yeah. Listen, appreciating that there's going to be huge variety between different companies and different organizations. Yeah. Do you think in general that organizations are better suited or are better positioned now, heading into this agentic change in terms of the sort of decisions

they need to make, the conversations they need to have, then they were when chat GPT hit, thinking about sort of, you know, the first co-pilot phase. Are they, you know, are there structures that they've built out over the last couple years that they can rely on? Basically, are you optimistic about organization's capacity to keep up with the change, you know, in the coming months? Let me stare at KPMG for a second. We certainly are, right? We've used this period of time to really change the way we think about this entirely about like how it fits into the work we do, the products that we offer to clients, the way we do our services, the way we train our people, etc.

I don't think everybody's in that situation. And I think there's a number, I talked to a number of folks who are like, you know, we feel like we've done enough and then we'll decide when we're going to bring our capital online. I worry about that because like you're talking about catching up to a speeding bullet. And so I hope they've made the right no regrets decisions around that. I do think that agents are pretty different. On the one hand, if you just think about it as like enterprise applications will deliver me a set of pre-programmed agents that I'll be able to use.

It'll make things like resume screening easier or some set, you know, that will absolutely happen. And you'll get that kind of as long as you've made good SaaS investments and other things. You'll be able to take advantage of that. That's not what you and I've been talking about. That's table stakes, right? That's happening. What we're talking about is the things that are fundamentally changing.

the way work is really occurring. We're probably talking about some version of a universal agent that's able to look across all of those things and process every single thing that an employee would potentially need. That's not being delivered probably by one of those systems because this needs to work in all of those systems. Everybody starts fighting for it. I don't think the majority organizations right now are that prepared for that agetic thinking. I don't think it's, I mean, that's not a, it's not pessimism as much as to say there's time, it's where we are right now, but I want to resist putting my arm around people and saying it's going to be okay.

There's, you know, time is running out on this topic. I think you feel that, I think I feel that. And since I believe this is a disruptive technology, disruption changes the winners and losers in industries, new ones show up. And I think you can see decisions that some of the big tech companies have made. If you want to see what could happen to an industry, you can see that. that happening in the tech industry right now. Winners and losers starting to materialize.

Once you would have thought are impregnable having real difficulty with their AI story, right? And others that have really jumped in headfirst been willing to remake, like how they think about their certain. And I just think that's what we see over there is coming for every industry, I believe. Yeah, perfect segue to sort of a final kind of wrap-up question, which is in a world where everyone has access,

theoretically to the same tools, the same models, the same technology. What's going to differentiate the winners? I think they'll go back to that vision point, Nathaniel, and that willingness to be bold and go after that vision at pace. We have this phrase internally. We use bold, fast, and responsible. I actually think it's a group of those things with really solid vision. And then you cannot think that it's a separate group that's doing this for you. It is every one of your leaders, you know, kind of buying into that mission and pushing. And by the way, it's not just AI. It's like it's bringing AI into the business strategy and the things.

Like if you just think, oh, well, we'll do this with this, this AI thing. AI first mindset really important, but it includes all the other things that are really important as well. You know, like it needs to come into the other things. So like if we're working on our delivery model, well, that includes AI, but it's not just that, right? But I think vision and that commitment to, I'm going to use our phrase, bold, fast and responsible, but something like that because of the changes in clock speed, the pace of this thing and whatnot. I'm in agreement. I think that to the extent that you're starting to see differentiation, a lot of it is just

willingness to engage with the world as it is, you know, at the speed that it's happening, with the sort of breadth and depth of change that's happening and really make a concerted effort to do what it takes to do that, even if it means surrendering some priors. So I think that that totally resonates. Can I ask you a question before we wrap up? I've been quoting your Dr. Strange example quite a bit. I actually think that's quite mind. It's a good mind experiment, right? Like so that you're a good thought experiment to do is if you horizontally scale an agent that can do some sort of thing and can go out and work in that way. And I know you've, you've a couple of examples. Thoughts on how we bring that into this discussion? We haven't really talked

about that idea of the real horizontal scalability, how we bring that in. What are your thoughts on how we ought to be bringing that more into the conversation? So I think that if we are trying to kind of understand where we are right now, right, and thinking look ahead, a framework that Microsoft shared recently, which I think is fairly useful is this sort of trajectory towards everyone becoming an agent boss. So their ideas, the end state is us, managing, you know, big swarms of agents that do things. And where we are now is more figuring out how to collaborate, you know, we've spent some time figuring out how to collaborate with, you know, co-pilots as assistants. Now we're trying to figure out how to collaborate with agents

as actual coworkers, right? It's sort of a variation on the theme. And I think that's where a lot of agents are currently. But the mindset expansion here, and sort of the Dr. Strange example, is if you were not limited by sort of ability to onboard people or cost of people and you could have not one person doing a job, but 100 people doing the same job in different ways, would that change how you did it? And maybe in some situations, the answer is no, because you just need a person to do the right thing. You don't need sort of big scenarios. But the example that I had used before was sort of marketing, right? If you're experimenting with campaign themes or copy, Well, why not have agents that imitate the 100 greatest writers of all time, you know, write,

write tweets on that basis. And then you could build a team of review agents to sort of make synthetic audiences and get feedback. And I think that in general, maybe the mindset shift is less the specifics and more. What happens when you really have nearly unlimited intelligence in infinite variations? Like, how would that change the work that we can do? And what would the process change? Would we look at, you know, are we going to all just be curators of the different work that

might have happened, picking the best work, you know, from what did? Again, I'm not sure that that's how it plays out for everything. But I think that it's a, for me, it's a sufficiently different way of seeing the world that I think it's useful, if only understand, in an understanding the magnitude of the change that we're about to face. Because if you go through and do that, and as you go into process redesign or something Or do you think about outcome redesign is the way I've been sort of talking about it, right? Like, how would I get to that outcome in a different way?

You think, well, why do we sample that thing? We should just do 100% of it because we have that, that availability to do that. Why would we do that in batch? We should just do it real time because we now have the ability, right? Like, it changes your mindset about, like, how I would accomplish that because actually the way I'm accomplishing it was a set of accommodations based on restriction led to that. That's really hard, thank you. Even in small ways, we're seeing just the ability to do more.

You know, we have an audit product to help companies figure out what agents they should be deploying. And it's based in a very, you know, like a standard sort of professional services consulting type of idea. You interview people about what they do to try to figure out what changes you might want to make. Except because we're using voice agents, we can interview everybody comprehensively and at length and process it all, you know, in hours, not months. months and you just never would have thought to interview every single person in an organization or department before because it'd be ludicrous, right? And, you know, it doesn't necessarily like, you know, listen, sampling and statistics are valuable tools. They can often get you, you know, a lot of the way there. But there's just all these opportunities that were never even considerable before that,

that are now. Yeah. I would say the other side of that is people can get too infatuated with tech and start to free people. And, you know, at the end of the day, One of the things we think about is so that I can be more with my clients so that I can be more with my people, right? Like, because that's still super important as well. It's like that, that piece that's showing up being there, being able to do that. Like, that's what this is giving us and to be able to be more effective in following up and making sure and everyone knows what happened. Like, that's the thing we're trying to get to. Well, this is absolutely.

There's this other dimension of this, which is in addition to just thinking about how differently work can get done, there's the other dimension is thinking about how differentiated the opportunity, like what the product or service on the other side could actually look like. Yeah. And I think customer service is maybe a really useful example here where it's so natural that the first phase of this is, you know, the customer service agents have some instant benefits over humans, you know, the agentic versions, that they don't get impatient, they don't get bored, you know, they're all, all these things that are useful where they're available 24-7.

But my strong guess is that customer service in almost all areas ends up becoming closer to a hybrid of AI-agentified customer service plus luxury Ritz-Carlton concierge-style human customer service at the top end because it's going to be a prestige position instead of just sort of, you know, something at the bottom of the totem pole because people are going to differentiate on how great their customer service can. be because it's now, they have the opportunity to. And I think that'll be a really powerful shift as that starts to happen. And there are a ton of examples where that's true. I mean, our family buys our cars from a place that is like that is the most, there's the best customer service because of that, right? Like, yep, like, 100%. So anyway, well, listen, I, what Nathaniel has been an awesome conversation. As always, Steve, you know, what I'm looking forward to is, is just how radically different this will be if we redo this again in six months. But I think, you know, what I will say is for people who are listening, look for the fundamentals, sort of the big

principles underlying this, because I think that they're going to be more instructive maybe than the specific details. Like, yes, you need to get your data ready. But I think, you know, broader mindset shifts, thinking about actually articulating vision, that's where a lot of the, a lot of the big changes are going to happen. So hopefully this series has been to help. Steve, I think this is a great way to close it and appreciate you taking the time. Yeah, we've been really pleased and proud to be able to do this with Nathaniel. And we know you're one of leading voices in this. And, you know, we're kind of proud that we're becoming a leading voice in it as well. Love it. I think this will help. And hopefully all the, all the listeners agree.

All right. Great. Thank you. Thanks. This has been You Can with AI. Check out all seven episodes and learn how you can better integrate AI into your organization today.