How CEOs are Preparing for AI in 2025 – Futurum CEO Insights

So I head up our CIO practice. So chief Information Officers, those who have to, uh, deliver very large and complex, uh, IT systems and keep them running and keep them upgraded, um, and move everything to the cloud and roll out AI and do a hundred other things. So, chief Information officers, um, are the ones with their, the, the bullseye on their back.

Uh, things don't work or they don't, they happen fast enough, especially, right? Today's, they won't happen fast enough. CIOs on the front line.

Now, normally for a very big technology, uh, uh, rollout or a major new tech trend, you would see the CIO really leading that, at least having ownership of it. It might be a big thing on the business side. It could be like, you know, something to do with customer experience, but really the delivery is still in the end, gonna be done with, with it, because they have staff, they have the enough people, they know how they connect all the systems together.

They have the support, they have the operational chops, they have the whole thing, you know, end to end. I'm a, I'm a tech guy, enterprise architect. I've rolled out really big systems.

I've still got probably about a half a million lines of code in production today. Um, so I know that there's a big disconnect between C-suite leaders and a lot of technology staff. And we can actually use this to kind of help Deconflict, because here's the, here's the kicker.

59% of CEOs, those are the ones that we talked to. And these were, you know, CEOs with over a billion dollars in revenue. Uh, these are global, so it's North America, Europe, Asia.

We talked to CEOs everywhere. Uh, and we, and on top of that ones we sampled, we also did in-depth interviews to really climb inside the heads of, of, of these, uh, organ of these folks. So these are all deals are like in the top three or 4,000 organizations in the world.

Uh, so this is what they're facing, and they think that they're leading AI in their organization, not, not implementation, but they think that they're leading AI strategy 59% too. And, and that's because the board, I I, I've talked to over 250, um, CIOs since, uh, generative AI came out. And almost universally they say the board is breathing down their neck to do something before a competitor derails what they're doing with, with ai, that's the big fear, is that it's, it's a powerful enough technology that you'll be able to offer your products at least a lot cheaper, or you'll be able to offer breakthrough products that are much better.

And it's gonna happen in every industry into every single company. And they know that the board, that boards are, their job is to protect the company long term. The, the guide and steward it.

The CEO moves the company in the future and everyone else runs it. And so it's this big stack, and then all this stuff, you know, it, you know what flows downhill to the tech, um, teams. We have to have to cash the checks that are being written by actually making this work.

And so this is what CEOs are thinking. They were much more, um, conversant with AI than we suspected they would be. They did not talk technology at all, except maybe their Microsoft partnership or something, or Google Partnership or something like that.

They're, they're really trying to figure out how, how to put AI in their organization and then what they're gonna do long term, and then how to protect against it. So this will, we'll walk through this whole thing. We have a lot of data for you.

I've tried to pare it down and make reasonable. You can get the report if you need it, if you really want the, the, the, the whole thing. But let's, let's go ahead and dive in.

Uh, and I will take questions. Uh, so he just, he's gotta speak over me so I can hear you. So if you have a question, please jump in.

Do ask them as we go along. So pretty much universally, uh, CEOs don't look at AI as just a tool. It's not just this technology.

I'm gonna roll out like analytics and understand my business. They do understand it's a strategic new way of operating their business and that it's gonna affect every level of their organization. And maybe not now.

And right now it's only in corners of their organization, but it is in, it is in corners of their organization, but they view it, it's going to be everywhere. And they're right about that. And they know they have this choice.

If they don't get figure out how to disrupt with ai, they will be disrupted. And that's the concern. There's a, there's a tremendous amount of value and a great weight on their shoulders.

They feel to do something. They don't know what it is. I mean, they have a sense.

They know in the short term, they have a good sense long term. They, they, the things they wanna do with ai, AI is not ready for yet. We'll get into some of what that is.

Uh, and so the, our, our study, uh, which is all done just in the last, you know, 60 days, we just finished it. We're trying to, things change so fast. You know, if this information's more than just a, a month old, it's outta date.

You know, things like deep seek happen and our all the rules change suddenly overnight. So, um, so, you know, we try and measure the real readiness. So what are the issues?

The what, what's the expected impact? What challenges, and we won't get into all of it. 'cause I can't give away everything in here.

You can get the whole a report for all the details, but I, I, we have some juicy stuff for you. Um, and we did this in, in partnership with the, uh, the management consultancy, uh, ney, um, who offered up some of their CEOs, uh, in this. And so bottom line is the future belongs to those who act decisively.

And it's very hard to act decisively in a fast changing environment. Um, and so how, how do you do that? So let's dive in.

Um, so I'm sure you've heard all the Harvard Business Review, um, headlines. Here's a few of 'em, you know, that are almost certainly true. I've taken the most conservative ones, believe it or not.

So AI's gonna inject $20 trillion into the global economy easily by 2030. Everyone wants a piece of that. CEOs are, their responsibility is to get a piece of that for their organization.

And so they need to, they need to infuse it in their organization, find the right places to do it without blowing everything up at the same time, right? Because AI has real risks, although most of 'em think it's manageable at this point. Um, every dollar spent on AI can deliver up to $5 in business value.

That gets everyone, this is why you get the CFO is gonna give the CIO the money. 'cause the CEO is gonna go to the CIO and say, alright, give me your best ideas. Um, and I know it's gonna cost also, gimme tell me how much it's gonna cost, and then, and what are we gonna get for that?

And that's, that's what gets the money, is what, you know, what are you gonna get for it? And in our study, um, 78%, so virtually all of them believe AI will drive competitive advantage going forward. You have to have AI in what you're doing.

And, and many of 'em are starting in the obvious place, like customer service and things like that, but they're doing other things too. We'll go into that. Uh, but only 25 think that 25% think they're ready.

So they know they're not ready. They know that. They know that.

And, you know, with things like talent are really a, a big issue. But there's other issues too, we will talk about. And it's not just automation.

It's about being able to make decisions better and faster in your competition. At the end, that's what competition really is to say, I, I know what to do with my company, the right direction to take, and I can get it, get there ahead of my competition, beat them. Um, and then I enjoy outsized advantages.

So having the competitive advantage gets you outsize everything that you want. Outsized market share, outsized profits, outsized revenue, all those things you want. So the race is on.

So are CEOs leading? Are they really lagging? Are they, are they helping?

I mean, should they be leading 59% think they are, should they be leading? Right? Well, Can you go back to that for a second?

Yeah. Yeah. So is this the opinion of CEOs or where, what is the source of these, uh, these statements?

Your great question. Almost I re stat. The first two stats are pulled up from relatively authoritative government studies and things like that.

The every other number that you'll pretty much see on this will be from what CEOs self-report. What do they, what do they believe? So you what this, this study crawl helps you crawl inside their head.

It doesn't mean it's accurate, is what they think. Uh, and that's, and that's really important because they're in charge of the entire organization. So, so you have to take everything with, with from that perspective.

Understand, it's not, it's not gospel. It's not necessarily reality. Although CEOs to be effective, they have to have a pretty good connection with ground truth.

Okay, because I, I question the first two as well. Where did those come from? Yeah.

Uh, where did those come from? And it's, I, uh, what are they basing it on? I, I can tell you I've already done my own Mark.

I have my own numbers. Um, so for example, uh, the global labor markets, right? Around $50 trillion.

That's what it is. And so if you can, if you could automate a significant percentage of that using, let's say, AI agents, right? So, uh, I'm, I'm amazed.

I'm, I'm talking to people. I'm doing another study of AI agents right now, and I'm talking to these early adopters in like highly regulated industries that are like, uh, be able to do things that were impossible just a few months ago with AI agents in terms of making drug recommendations. That, that, that screen all of the counter indications, uh, that, that, um, uh, comply with all the regulatory compliance for that person in that state, or even in that city.

And they overlay all the AI agent overlays, all of it that makes recommendation. And the doctor just has to check all the documentation, make sure it's, they're okay with it. And then that, that prescription's issued, uh, it is cutting for complex healthcare situations.

They're using a agents to cut the time to, uh, to issue a prescription to somebody who has, uh, you know, uh, multiple major, um, diseases is cutting it, uh, a week off that time. They can get, they can do it like in two minutes. So it's that kind of massive automation of very complex as we, we have a very complex economy.

So, but it, if we can tackle very complex situations like that, now I estimate of that $50 trillion global economic pie, AI is gonna be able to do a good 2020 5% of that. And You're not saying that it's gonna displace 20 trillion, you're saying that it's gonna add 20 trillion 'cause those people aren't going anywhere. Yeah.

'cause we're, well, we're in most industries that are driven by technology, we're constrained by labor. That's the real issue, right? There's not enough developers, not enough testers, um, not enough anything, uh, just, and that's just in, in, in technology.

So we're, uh, every CIOI talked to, I think the last survey was 90% said their talent constraint. They're, they're fundamentally limited by, uh, how much of talent they have or even can hire. They just can't hire.

Uh, and a lot of other industries, uh, high tech industries are also labor constrained. You can look at aerospace or anything, tons of other things. And so if AI can automate the bottom third of those activities, um, that's easily 20 trillion per year.

Per year. Um, and if you look at, um, two thirds of organizations are looking at rolling out agents by the end of the year, right? So that's, that's other data.

So a agents are coming really fast, and I, and I think it's actually quite a little bit dangerous because I don't think we know how to put guardrails on 'em yet. Um, if you look at like open AI operator, so what we're seeing, what CI is reporting is AI is just coming and over the floodgates. There's, uh, they're trying to stomp it out because coming into every tool, uh, everyone's using unsanctioned ai, their own devices.

I mean, it's just everywhere. Uh, so yeah. So we're, will we see that, that 20 trillion, I think we will remember this is five years from now.

So both at that, that labor market will be more like 60 or 70 trillion. Then you add AI to it, how much more you can generate with it. So yeah, it holds, I I said I probably chose conservative numbers.

Your people are saying 30 or 40, which I think is crazy. Hope that helps. But yeah, these are not numbers we just, I'd leave throw around.

We're, um, these are conservative numbers given, uh, you know, what we're seeing, like what the advanced models can do is truly astounding how much of IP they can generate completely working IT systems or, or a new drug molecule, or it's, um, Google, just as an aside, because I, I'm trying, trying to technically ground some of this stuff for you guys who don't spend all all day trying to think about these big numbers like we do. Um, okay. Yeah, please move on.

I, it's, I was, yeah, yeah, no, no, it's, well, it's really like, uh, so just very alpha fold. Um, Google's AI can throw off new, uh, drug molecules faster than they can test them. That's the kind of economic acceleration we're talking about.

I mean, way faster than they can test. Okay. So important sidebars, just make sure everywhere we're all on the same page here.

Uh, so we saw some, you know, there's, there's not many double digit gains left in the enterprise or so we thought, you know, but because we've applied automation everywhere, we, you know, we've either coded it up or use low code or RPA or, or BPM or whatever it is, uh, to do all these things. But we're still seeing, you know, uh, uh, one bank we talked to said they were able to dramatically cut fraud because it can much more intelligently spot phishing and other, uh, other exploits because it can read natural language and figure out when, when things are not not correct, they're able to really substantially high double digit reduction in bank fraud, for example. So that was specifically about phishing and that kind of thing, or, because when bank brought to me could be the security, Not specifically fish, uh, phishing as just an example.

So the grounded, oh no, they were talk, they were using it broadly across all, uh, uh, customer communication to, Right. So that kind of is a little scary because I would like to know more about that number because that could also be they, one of the, you know, algorithmic justice kind of situations where they decided to implement things that, um, adversely impacted people that maybe should, shouldn't have been impacted. So I'd like to know more about how did they, I'm and, and I'm in the, you know, also the media business, so I understand, uh, all these, when you boil stuff like this down, you lose all the nuance, right?

So is it, is it over, uh, triggering? Uh, yeah. Uh, is it over flagging scenarios?

Uh, I would guess probably yes. We don't, but we don't know. But they're very happy by, by the fact that they're, they're seeing substantially, uh, lower fraud levels by applied generative AI to fraud detection.

Also, um, when we, when we summarize to your point at a very high level, like, you know, bank cut fraud by 60%, um, do you believe there's tracking of insider threat that could be included in this because an insider threat would actually know the system that's being deployed? So do we, do you think there's any, uh, timeline that you've, uh, come across for what A CEO is looking for their external threats and there's the internal threat. Um, so when they think of fraud, it makes me think it's external, not internal.

Um, but, uh, I'd be curious if there's any, uh, follow up research that you might be doing on internal. Well, These are from conversations I actually had with CEOs, uh, for this study. Um, and so I can tell you, they, they, uh, the, this particular CEO was, uh, looking primarily at, at, at external fraud, but inside of the threats are the, one of the biggest remaining, uh, as you know, um, vectors in the, in the organization today.

So I believe this is mostly all, Can I ask 'cause there's a smaller bank, because if you take the large banks like the, you know, the Bank of America, JP Morgan Chase, they've been doing this pretty efficiently before LLMs came along. So I was wondering if that bank really was not a tier one organization that had invested a lot of ML into fraud? Very possibly.

Uh, uh, we did talk to a lot of regional organizations. Um, you know, we were, uh, didn't necessarily always have the, you know, CEO of JP Morgan on the, on the phone. Mm-hmm.

Uh, certainly, yeah, there, one of the stories that came out is that things that were hard to do, but possible before became much easier. Regenerative ai. So those, those fraud detection technologies might come at a very high price or very hard to implement.

Whereas regenerative a ai, they were, uh, able to much more quickly, the speed especially really came out. There was a lot of these stories all happened, like in the last few months and all happened within the last year. Mm-hmm.

Um, there, there, the rollouts, I'm hearing consistently weeks even for IND industries, which I thought was very interesting. You don't hear that. You just don't, I mean, you don't, you don't hear anything in regulated industry takes two years, right?

So, Clear. These are you, you're saying ai, you're not saying gen ai. So, to the point about we've been doing, we, We, we generally prepared all the conversations.

This thing we're really talking about the new stuff, generative ai. Okay. Because just getting back, I mean, I, I personally talked to a Canadian bank 12 years ago.

Yeah, That's what I'm saying. Detection, using machine learning, uh, same, same with, uh, you know, predictive maintenance, uh, and things like this. So, so I guess it's the new shiny thing, right?

And maybe it's encouraging people to do the, the right thing. Um, well, no. So for example, um, says one CIO, um, who was on the call with, uh, with this CEO, they were explaining how they could do, like, it was the, the example I'm talking about is the BA to prescribed drugs in a very complicated scenario.

They could do this before, um, uh, using other technology. It just was very unreliable when it kind of got to the doctor. It just wasn't as accurate as the gen ai, um, which was able to correlate, uh, and overlay multiple different regulations that it had norm.

It had to look right down regional level what, what the, what the regulations are, then overlay everything on top. And it could do that, it could read all the regs in real time and make sure that, um, that they're, they're being compliant, uh, that you, that, that worked with, uh, you know, certain drugs have to, you can only, uh, prescribe generic drugs in certain areas, for example, things like, like that. And it would be able to under, through everything and overlay on top, uh, even if the regulations were unusual or are, or strangely stated or whatever.

And previously they would have to go and, uh, process all those regs and make a machine readable themselves, and they could never keep up, said is this, could do it for it. So it take a week to get a person, their prescription, the gene AI can just easily zip through all of that, uh, just process all like 10 or 10 or 12 documents. It would have to then overlay on top of each other, come up with a recommendation.

It was pretty impressive stuff. So, yeah, that's different If it's encouraging people to really do, at the end of my story, in fact, was the CEO really loved this idea. And then we talked to middle management and the quote said, we don't need any fancy math to solve this problem.

So, uh, maybe the gene AI bandwagon, they'll, They'll do it. Well, just keep in mind this is, a lot of these things are above their pay grade. So they're, they're reporting what they, they been told is the, the benefits that they're seeing.

But for example, one insurer, um, this is the other example here. They're, uh, they're, they do workers' comp and, uh, they, uh, there's all this, you have to be the entire case, case file to be able to answer some of these questions. And normally they'd have to make someone wait in the line while the the person's going through the person's case file trying to understand everything where it happened and where it was.

Whereas the gen AI could do a lot of that and only ca much less, uh, deflection to an agent where it says, you know what? I just can't make sense of this case file and, and where you are in your journey, uh, I'm gonna kick you to a human. It was able to navigate, you know, long term, like five year long, uh, workers' comp cases, process, all of that.

They said they couldn't do it before. Previously, the automated systems can only process case files that were very short or simple. And now they can do these big.

So interesting. It's, um, uh, so again, these are, these are grounded and these are from, these are gen AI specific examples that I'm giving here. So, like, I didn't think there was 40% more to come out with predictive maintenance, but I'll concede that a lot of these were, were organizations that couldn't afford big AI teams.

Almost all of 'em said, we, we only now are really hired, uh, somebody who knows ai. Um, before we, you know, we, we were too, it wasn't important enough for us to hire someone any AI lead to do this for big companies. You already have, you already have big AI teams.

Yeah. For That second bullet. I mean, uh, you know, there's, there's, There's all the software players in this space.

You know, you have the, the, the SAPs, the, the, the IBM maximos, the Oracles that have got, you know, specific offers, um, in this space of predictive analytics for people that have large fleets, deployments of equipment, you know, components within set equipment, um, because they pioneered a model and that model that they've pioneered and that data they've been, you know, holding onto for decades in some cases, um, 40% of big number. So I'm, I'm kind of wondering, is, is, uh, it, it, I didn't expect to see it either, except the, these CEOs were very acutely aware of these numbers because their bonuses are Bonus tied this Absolutely, yes. So they, they have the numbers right in their back pocket.

I was surprised how often they could just bring that right out. Oh, yeah, I just gave $2 million in our customer care unit. So, good point, Diane.

A lot of these probably have nuances, but maybe a specific situation where it's reducing maintenance downtime by a certain percentage, not all downtime by 40%, I think. So CEOs say they're ready. Um, and, you know, they claim some confidence, but only 19% are really trying to do any kind of genuine true transformation, which I think it's early.

It's not time to do major AI transformation because we're so, the vendors, uh, the AI vendors are not mature or not ready. Um, and so our, our finding is over, there's overconfidence at the top which masking, uh, execution inability below. Um, and what's very interesting is that, 'cause we asked these in separate questions, high performing firms, um, decentralize their AI leadership, um, 59% of, um, so it basically, I'm trying to make sure I state this correctly.

'cause uh, I, I, I see that I've got a, a, a wrapped audiences, uh, dialing into every, every detail. Um, so the higher performing firms decentralize their ai, whereas the ones that that micromanage, um, they have, they have a significantly higher failure rate for their AI initiatives. They self-report that.

So we say, how much are you involved in leading ai? Are you, you know, involved in, you know, week to week, day to day? Uh, and those CEOs tend to have less, um, a smaller, uh, uh, percent of successful efforts than those, uh, that that set the vision and said, go out and I want you to execute AI transformation every part of the organization.

Uh, those tend, those, those CEOs tend to report higher success rates with ai. Yeah. And is this a, a chicken or egg problem?

I mean, the more successful companies, you know, started out with a single micromanaging AI and then diverged into a mold more decentralized approach. And, uh, the ones that are still in, in a micromanagement, they're just too early in the adoption cycle to really get effective use of it. And it could be, and that, that could be useful analysis.

Uh, but I can tell you I'm studying, uh, um, tech adoption for many, many years. CEOs don't understand technology adoption, and they don't understand that even if you automate some particular big thing in the business, there's still a thousand processes and a thousand apps that aren't effective. Um, and so like with digital transformation, we saw the exact same pattern.

Those that, that broadly decentralized digital transformation, they set a common vision, common goals, and then they said, go to your part, your division, your function, and execute digital transformation. So it can happen in more places rather than if we just visually transform a couple big things. But most of the organization is still antiquated.

So we saw the exact same pattern CEOs because they're weren't as involved in digital transformation or tech adoption. Uh, don't necessarily understand that. But one of our, one of the big insights was, uh, don't overly centralize AI adoption, decentralize.

And we have the data that says that they tend to have more success from doing that. We'll go to the next slide. Any, if there's any questions, we can keep going.

So the, um, unsurprisingly well-managed organizations that have more rigor, you know, and I see this with, it tends to be larger organizations that have a professional project management office that, um, and, and IT departments that actually track results. So 48% of high achievers, uh, uh, rigorously track, um, AI versus only 17% of the organizations that reported they, they were doing struggling somewhat with ai. So not surprising.

And it just, it's, the message that we've always seen over and over again is, you know, you, you can't manage what you don't measure. And, um, and many organizations are used to, if you're, if, if you're have a very powerful technology, you expect to be able to throw it over the wall and, and get some results, right? I'm gonna get something good, right?

Um, and we know that's just not the case. You just can't be sure you're hitting your target actually improving what you're really trying to improve. So we saw that high achievers, uh, uh, and we separate the questions deliberately so they don't know that they're answering that.

Uh, so that's a key part of the study. And so, uh, it, AI without ROI tracking is just not worth doing. Um, you know, and yet you're gonna have that, you're gonna say, oh, well, copilot sounds good enough, I should roll it out across the organization.

And some CEOs did psych that said, well, we're rolling it out for ai. Productivity is one of our big initiatives. Um, you know, are we having trouble tracking it?

Well, those, those ROI numbers are hard to track as opposed to aiming AI at some very big and core process, uh, in your company. So What are some of the metrics that they're trying to hit? Yeah, So, you know, uh, their, their biggest numbers, of course are the top and bottom line.

There is, is most important thing for, uh, A CEO. But, you know, of course then if they're in specific businesses, you know, customer lifetime value, uh, efficiency measures, uh, a big one is, um, you know, time, um, uh, you know, quote to cash, how long does it take for me from a, from the time that someone places an order to the time that I actually get paid for that. So do you see the CEOs, uh, looking at a specific, uh, like, like how do they measure that?

So like, if they're implementing ai, that could be a huge, ah, that could be a huge number of things. If they're implementing generative ai, do they track, like, do they start by tracking one specific area and see how it hits all of those metrics that are typical for the bottom line? And, and or do they are, is it just kind of like, if if they, they get some AI only An ROI that we, we consistently heard from where they had numbers in their back pocket is where they, they tackled and put AI into a very specific function like their, like tier two or tier three customer care where that's been resistant to automation up until this point.

'cause um, they had real specific savings like that $2 million savings from that workers' comp insurance company. They said, you know, we had our best customer care people tackled our toughest cases, and now we only need half those many people. That's right.

Generative AI before that was automateable. And so I, I went to the board and they said, I just saved $2 million. Now gimme more money to go do this.

So that, that it's when they, uh, the only numbers we really heard, um, in the survey, we can't hear, we can't hear it, but we did, that's why we did the in-depth interviews. So we were specifically asking, where'd you use ai? Where did it work?

Where do you use generative ai? Where did it work? What do you see?

And sometimes they couldn't say what they they saw, but, but a lot of the time they had specific numbers. So The specific numbers were cost savings from eliminating jobs and Almost all cost savings. Yes, exactly.

Wow. Yeah, there's not, not not too many sales boosts. Um, the, uh, customer retention is a big one.

We had, there was one ink manufacturer who make most of the, most of the ink in Europe. They make most ink from newspapers and magazines these days, but their, their business is being eaten by companies in Asia. And, um, what they're now competing is, is on the quality of, of customer service.

They can, uh, they can answer questions in natural language. You can say, where's my, where's my ink, uh, shipment? Um, do, how much outstanding do I have with you before you cut me off?

How much credit are you giving me? They can ask simple questions, get very simple answers, and, and they're, they're retaining customers that they believe, based on their conversations is, I know I can get it cheaper, but, uh, they're much harder to work with compared to you. So we're working with you guys for now.

Right. So, so they're using generative AI to prevent, uh, customer defection, um, and, you know, just frankly, to stay alive until they find out some breakthrough product. Yeah.

So you mentioned the LTV and there's the cac, and then there's like the churn. And so some of these numbers, it sounds like they're, they're tie in whatever was in their original dashboard implementing ai. And if they're successful using these other metrics that you're, you're sharing that they, they did it the right way, they're able to probably brag about this to a board, to investors, and that's why they're able to trumpet a story.

Yes. Correct. I i, you're, you're accurate.

Okay. My son firmly in shape. Did any of the CEOs comment on the Harvard Business Review story about how CEOs could be replaced by ai?

They did not. That did not come up. It's all those other people.

Yeah. Well, well, um, what they, one of the things that came out, I'm trying to remember if we have a slide for it. One of the things that came out is that they, in the short term, they're doing these cost savings.

Uh, and these, you know, are, are, are addressing an urgent problem, a difficult problem, um, that's targeted, but they're, they want, like that ink company says we want, they wanna do like alpha fold. They, um, we talked about that. They said, yeah, we want a breakthrough formulation.

We want like a new super cheap quick drying ink that doesn't require uv. You know, something like that. We, and, and he says, I actually have a lab effort to, uh, try out these ink formulations.

We have an advanced, we have one of the best ais and they're reformulating ink so that we can create a breakthrough product. And it hasn't produced it yet, but I have hope. So they are thinking like this kind of bigger thing that, that was an, tends to be an exception.

We only heard that a couple other times. But, um, it's very interesting how they're thinking. Uh, they, they're hoping this will give them the, you know, the be able to break through the, whatever the barrier where they're kind of stuck in product development now.

Um, and as a result, as you know, I mentioned, uh, the, uh, 95% of the companies go surveyed and interviewed focused on these quick wins versus like, our, alright, how do I create breakthrough products and services? But they talked about it. They, they know it's coming.

They either don't know how to do it or they don't think the AI's ready for it. Uh, so I, um, uh, do I have it on the slide now, but I'll just talk to it. Um, uh, so I talked to the head of a global audit firm, you know, the CEO of a global audit firm who says, I'm going, my audit business is gonna be dead in three years.

There's no question about it. It's already been so extensively automated by ai, um, that I, I'm just gonna be, I'm gonna AI enable it. It will do, AI will do, you know, the majority of the dot 90% of the audit.

It'll be overseen by auditors who can check everything. And then that's not my core business anymore. My core business is now saying, now that I have this in my model, uh, I can help you make strategic decisions.

I can recommend cost savings, I can recommend, you know, uh, uh, you know, kinds of useful financial instruments you can use to improve your finance or whatever they want. I can find risks for you and things like that. Um, so they, they expect total business model innovation.

And we also saw this in staffing firms said the same thing said staffing is, is completely changing. Um, how we screen, how we interview pre-screen, pre-interview, um, how we help companies on board is all, it's all be being AI powered in a, I don't do what I'm out of business. So because all our competitors are doing it right in the middle of doing it right now, you know, so it was interesting, but it's still, right now, the majority of all their focus and spend is on these all.

Let's prove that it does something that we couldn't do before. Right? So, yeah.

So it's very interesting. Um, and of course, I wouldn't be surprised there, there's no talent to do this and it needs people. There's, or they're all using a, a mix.

Um, most of 'em are avoiding the big management consultancies by the way. They say every management consultant is pitching AI transformation. And every single one of 'em has reached out and said, we'll, we'll do your AI transformation.

And the majority have resisted that so far, thinking they're not ready for large scale transformation, but they're working with usually smaller firms that are experts in their, their specific industry like hospitality or whatever, where they say they already know our business and we, we, they've proven they know our business, our specific business, and now they're, they're, uh, uh, dealing with a lot of the cost and the expertise and talent issue. So we're just gonna have them do a lot of our initial work. So that, that was common.

So when, when, um, what did they see as the lack of experience? Was that for, like, from a data scientist or just like the ability to, to an, uh, engineer prompts? What were, what did they see as the, the biggest, uh, skill gap?

Um, they were that granular. Um, they were, uh, a half of them were hiring an AI lead of some kind. Um, you know, uh, bigger companies were hiring, you know, an someone to lead a big AI effort or, you know, CTO they were hiring.

So we're hiring our first CTO and it's gonna be an, you know, we've hired someone from the tech industry in Silicon Valley. We're super excited. We don't really never do stuff like that.

Um, but, uh, they're, they want AI literacy across a lot of different roles, right? So not just data science. And, um, they're, they're upskilling.

So some of 'em are, um, some of 'em, everything from, uh, they do, they've done all hands meetings and then they've had extensive education programs. Uh, these tend to be in like areas like management consultancy and things like that, where this AI's, you know, front and center for they, they have to be offering that. So they, they, they were doing, they're doing broad education across the entire base, you know, of, of firms.

So, hope that answer your question. Yeah. Thank you.

Uh, Diane and Mitch Ashley did, yeah. Did the topic of developers and how we develop software or the AI skills that we need in IT and technical organizations come up with their thinking about that looks like in the next two to three years, or is that kind of too far down the pipeline for them to, Uh, I believe that's below their pay grade. Uh, uh, but what was interesting is that they're not so much I didn't hear that they were leaning on.

Uh, they consistently did not hear that they're leaning on their CIO. They're more, uh, trying to hire someone who's really going to think about AI in the business and okay, you know, might, might be A-A-C-T-O or might be, we know we're just hiring an AI lead, but it seemed to me like the CEO is like doing this hiring outside of the IT loop. So it's, it's something that we'd like to press on.

We're gonna, you know, follow up this study here and, and find out how things have changed because this is just, you know, this is first full contact with the CEO, really having to grapple with a transformative technology. Awesome. Thanks.

None, none of them mentioned developers. Just, okay to answer your question specifically, um, now if you're, if you're in management theory, you understand, uh, you know, there's, there's innovators, you know, that are way out ahead, then there's that big gap, you know, uh, and then there's, you know, the, the early adopters and, and everyone else. Uh, and so a lot of companies want a fast follower approach, the small of the company in general.

They want a fast follower approach. Meaning as soon as they see someone having a certain success in their industry doing something, they're gonna adopt it right away. Say, but they're gonna let, let, all, you know, the companies that make mistakes and don't do anything, they, they can say, well, that, that didn't work, so we're not gonna do that, do that either.

And so they let the, the people out front take the arrows. Those people out front can get outsized benefits if they, if they guess correctly, they bet correctly on, on what the approach is gonna be. But over half of CEOs that we surveyed and talked to prefer as fast follower approach, they're gonna run, right?

They're gonna see what works, and they're gonna run right in and do that. And yet, what was interesting is when we asked about how, how aggressively they were, uh, rolling out ai, and then we then separately in a different question much later on, we asked, how's that? A, how's it all going?

Um, rushed AI rollouts had a, had a higher failure rate, which is what you would expect, right? So if you're not really paying attention to your data, you're not really being agile, not listening to what you've learned, um, and genuinely learning from, from your early experience and then adjusting. It's not assuming that you know what the outcome is gonna be.

You know, that's what Agile tells us is, is that, you know, you, you know how much money you're gonna spend, but you can't be a hundred percent sure what you're gonna develop. 'cause you won't know. You don't know what you really need to.

You build it. Um, that's, um, so it's very interesting. Um, and so 58, so almost two thirds of struggling firms pursue aggressive AI strategies, which I think is right.

If you do this too fast, stuff is not ready. The the products are not baked. They're not mature.

We don't know who the top vendors are gonna be yet to make the bets on. I don't know if they, they're gonna be able to support this properly or if you're even gonna be using the right ai. So, um, so, uh, steady wins the race right now, uh, because this is, this is gonna happen for another five or 10 years, right?

So, um, and, and so we, we are able to get that from their, from their impression of their AI portfolio versus how fast they're pushing. We're able to, uh, extract, uh, a signal about, you know, don't, don't push too hard, And then the power of letting go. And this is something I've seen in a lot of my other, I've done a lot of work about, you know, technology and business leadership and, and how it all fits together.

And we see that, uh, leaders that can let go and allow the organization, the experts closest to the problem. Um, you give them the support and the training and the education and the best practices that from across the organization so far. And then you let them innovate in their corner of the organization, and there's real power in let it Go.

Uh, and that leading firms that reported the most success and we'll able to see this, and you can see this more in more detail in the report, uh, really see, uh, distribute AI leadership across specialized teams and cross functional teams in particular, especially early on when we're asking the questions, trying to find a signal, we're saying, well, where did you really see, um, real results? And then when they brought cross-functional teams together, especially, um, they, they seem to be reporting of better results. So I'd love to see that as, I think that's the right answer.

And we see this with like open source software, you know, decentralized development beats centralized development. Not every time depends what you're doing, but a lot of the time, and they're getting consistent governance as well. I mean, for things like, oh, we'll get to that.

Is that the next? No, uh, uh, well, let's, uh, make sure we get to the governance piece. Um, uh, yes, we had some, we had definitely asked about that.

We have, we have some data to share at the, um, work workforce fears about AI are real. And they're, and they're, and the CEO's aware of that. Um, and, um, 41% of 'em said they're not sufficiently address addressing the impact on jobs and, and processes inside the firm.

Um, and the, um, is that, you know, the, the message, you know, that I've often, the way I've phrased this is, is, uh, when we tell people that AI will take humans and make them superhuman, that's, that's a better message than saying that, you know, we're gonna automate everything in the business that can be automated. And then that, where does that put the worker? Right?

So, um, one of the messages that came out, those were the, those who have fearlessly embraced their ais and told their workers, don't worry about your jobs, but we're going to, we're gonna use this, but we need your help to do it, and you'll be part of this story. We need you please be involved. They, those are also the cohort that reported higher success rates.

So creating a fearless culture, a key part I have that we have a, um, we have a playbook. We've extracted from the most successful company companies reporting the most, that, again, it's self-reported. We understand the limitations.

Um, no surprise here, uh, what the clear, a very clear signal, 60% of failed AI efforts. We said, why? And they go, well, our data wasn't right.

We couldn't get the data. It was bad or it was late. Or, um, you know, we had, there was some, there was, it always boiled down to some data.

And this is what we would expect because most organizations don't have sufficient data management to participate in a lot of these very integrated solutions says AI works. Generative AI in particular works well because it integrates, it's able to integrate data from so many sources with a very variable effort. Uh, and that a lot of the, the bang for the buck we saw when we talked about what was your pilot generative AI pilot that was so successful had often was bringing data in real time from multiple sources and sorting through it.

Uh, but, but the organizations that had failed efforts side of that, that bad data, um, infrastructure, um, in some form, here's your, here's your governance slide. So most of 'em, I not surprisingly worried about AI bias and ethics. They think they've got guardrails, though, which I think they're very optimistic about that.

But only half have a governance framework. So there's a big gap. Um, and I, I imagine the governance framework is very immature as well.

But we did not in this, in this, uh, version of the study measure, the, the maturity of it, we're trying to figure out a way of doing that. But, um, Do you have any, uh, did you, were you able to dig in, in your interviews to understand if the concern is really about bias? Or is this, uh, positioning that they believe they have to be concerned about bias because everybody else is concerned about it?

Some of it is lip service. Um, they are, um, more worried about, uh, some AI doing something wrong and, and, and having shareholder or public black backlash. I think the, the, the fear is more there is a shareholder of public backlash.

'cause AI just is really more, uh, you know, and, and there's, yes, they care about bias and ethics, or at least they say they do. Um, but when we ask, we ask specifically ask them, but are you worried about a shareholder or public backlash? We're like, yes.

Oh yeah, that's a real, that's a yes. That's why we want guardrails and governance. So we didn't ask them that up front.

We asked that as a follow up question just to make sure that, because we, we wanted 'em to bring it up. And they didn't bring it up until we brought it up. But then they said, oh, yeah, that's our big concern.

Um, that is unfortunately what you'd expect. Um, the, um, and of course non surprise finance, healthcare, insurance, uh, how the regulated firms already had strong oversight. They already have a strong muscle for that sort of thing.

They have a place to put it, that, where that strong muscle is accessible. So not a surprise that everyone else lags significantly. Uh, that's not a surprise.

Um, the, um, they're planning to put in governance more better governance councils. Most of the CEOs we talked to said they wanted to do that. Um, I'm trying to see if you guys would, now that I understand a little bit what, uh, this audience's interests are, um, the, um, this is an aggregate, this 52% faster re uh, response time.

It's, it's in our study. Um, but in terms of decision making, 'cause that was such a, a unexpectedly loud signal where they basically said, we want to make decision making do decision making in real time. If we can faster, I could, we get a big basket of stocks.

So we have to process an order that we think is high risk. We want to be able to like, make the decision if it's like strategic to our company. There are some, some industries like aerospace where ability to bid on something or take an order is can make or break.

You might only take three orders a year. Uh, or like you're a financial services firm, you're managing sovereign wealth. You might only have one chance at taking a, so, you know, uh, uh, you know, $10 billion from a sovereign wealth fund.

And you have to decide, can I turn money? Can I turn a, a good return on this money and I've got, you know, two hours to decide whether I can take this $10 billion, right? So that's, they, they're very hungry for being able to make decisions in that, in those scenarios.

Us, um, they broadly expect, they didn't never said the word a, a agent based ai, um, the, um, but they all said they are want to take and put AI in charge of actually doing, actually doing the work of the business. So we, we talked about it, and this is that global audit firm where they expect that, that the actual AI will be on the front, will be the front line talking to the customer, uh, to do a lot of the audit process. That, that, that they'll, there'll be auditors too.

Senior auditors overseeing everything and still humans talking. But a lot of that will be is now go talk, there's my, our audit system has a bunch of questions for you, you know, type of thing. Well, In our agentic future, then the, uh, person that they're talking to is actually an AI agent as well.

And so we can have my agent, AI agent talk to your AI agent, and that, oh, yeah, As I'm doing it, I'm doing work on agent based ai. Yeah. So agents, using agents, that's gonna happen very, very quickly.

I mean, it's already happening in some of these products. So it's very interesting. Um, and it's not as much as CEOs need to learn about the ins and outs of an IT project.

Um, you know, and, and now we've kind of left, left. There's, there's more, there's more details, uh, lots more data in the study. Uh, I just, you know, we just don't want to give it all away here.

Um, but you know, it was clear that long. I always said, what are you thinking about in three years? And they all had very different answers than what they were doing today.

And it was all about, I need breakthrough products and services. I need things to compete. I, I need to do things my competitor isn't thinking of.

Uh, but our customers want. And, um, a lot of 'em just felt like AI's not ready to do that yet. And they're right about that.

Um, but they want to get there. Uh, and they are, you know, the, the, the, the top cohort where there were definitely CEOs, they were thinking ai, AI first. Um, we have a heat map on what the top concerns are, which is very interesting.

It, it varies. What, what companies are worried about is heavily dependent on, in on industry, you know, so, um, some of 'em are not, not worried about their public brands so much, um, are risking it with AI and others are incredibly, um, sensitive to it. But I do wanna get to this part right here.

So we summarized, we took everything from all the high performing, uh, the, well we believe the CEO's cohort that was the highest performing and extracted what they did what, um, to get to where they are so far, the ones that we use this by ones you were able to us, right? Uh, what they're, they were seeing results that were very positive and had doubled many cases, double digit numbers that were not expected to be there. Um, and not so many things you would be find very surprising.

We have the decentralization, uh, thing in here. But the, uh, the fearless culture, AI culture thing was very interesting is they, they, uh, really quelled jobs, security fears upfront and said, and made them said, you're, you're part of this story with us. Uh, tho those c the CEOs who report a lot of success tended to do that.

Uh, so they didn't have resistance, they didn't encounter resistance. Or a lot of people whose jobs are be automated, they can install an automation project by just not revealing how how was actually done. Right?

You know, it's very interesting. I've been involved in a lot of automation projects over the years and, and there are three to block them if you have the right cohort of people. Uh, but the good CEOs really, um, really focused on on not doing that.

And just to wrap up, 'cause we wanna make sure we're, we're sensitive on time. Um, so strategy, talent, data, and governance in that order is what really matters. I understand what you, where you want to go, what you want to do, making sure you have the talent to do anything at all.

Having, getting your data story straight and then making sure you don't blow your own leg off with this stuff, which is, um, you know, I'm waiting for the first really big bad AI generative AI incident, self-inflicted. Um, the, um, That's actually consistent with one of the predictions at another event I went to, which was all things open in Raleigh at a, an AI meetup. They had a panel from Fidelity, duke University, a startup Flash, and then they also had someone from IBM.

And uh, that was actually a common prediction amongst that entire panel. Someone's gonna do it badly at scale and be in the news for it. Yeah.

It'll probably seriously hurt the company. I mean, we're gonna see, I think we're gonna see a, a Fortune 50 company collapse because of a failed AI project. We saw it happen.

We've seen it happen now with cyber security. So, and, and because AI will infuse how many things, uh, yes, I think it probability is very likely The equivalent of a data breach, right? But well, Yeah, exactly.

It's self-inflicted. So, um, but if you really have a really important takeaway is this short-termism, which is unfortunately how our financial markets are structured. Uh, it's all due to do something now, now, now.

And, and don't give that much thought. They're all feeling some like vague thought to do, I wanna do these things in the future. Um, but a few of them were making any serious investments to get there or even to study it.

Um, so it was interesting. So I hope you found that useful. Um, that's our study lots more.

com. You can go and request your copy and we'll send it to you. Um, um, but yeah, uh, it's, it is top of mind.

It is top of mind to expect lots of focus this year, uh, and all the other data from all the other sources. I've been tracking everyone else's data. Everyone's saying, despite reservations about all these things, about talent, about the right platforms, everyone's forging ahead.

Two thirds of organizations are forging ahead with major plans, and it's the biggest growth, uh, item to the budget other than cybersecurity, which is still often the biggest growth. Uh, but I, AI is the biggest new growth that it budget. We see that in our CIO study as well.

One problem they can inflate and you spoke about the CEOs telling their employees, you know, you're with us, you want be better. You want to use AI to be a better, more successful company. If they do start firing people, well that gets out.

And once that gets out, it, it, every other CEO you know, kind of loses that ability to say, oh, oh, don't worry. Right. So there has to really be kind of a commitment, um, for the CEO to follow through with that.

Yeah. It has to be real. Uh, or that, or that people will leave the sinking ship and guess who leads the best ones.

Right. You know, they're like, um, so that's the challenge. Uh, yeah.

So interesting things, but lemme tell you, it is top of mind. We had no, we had no CEO that said, oh, we're not, I'm not spending any time thinking about it. They're like, no, I'm, I'm worried and I am concerned and we're doing things that, we're trying things.

And so it's, it's real. It's, it's gonna be one of the biggest transformations of our lifetime.