Randy Bean on AI Investment ROI and Data Leadership in 2026

Jim Merrifield (00:00.915)
Well, welcome back to the InfoGov Hot Seat. My guest today is Randy Bean, author of the 2026 AI and Data Leadership Executive Benchmark Survey and advisor to Fortune 1000 organizations on data and AI leadership for over four decades. Randy, thanks so much for coming on.

Randy Bean (00:23.736)
Jim, it's my pleasure. Great to see you.

Jim Merrifield (00:26.743)
It's great to have you here for everyone listening. This is the 15th annual benchmark survey that we'll be talking about during this podcast. It's invitation only with participation limited to the most senior AI and data leaders at Fortune 1000 companies and leading global brands. And this year, you had senior AI and data executives from nearly 110 companies.

And 96 % identified as C level C level execs or equivalents. That's pretty amazing.

Randy Bean (01:01.39)
Yeah, thank you. As mentioned, I've been doing this for 15 years and it's really kind of become a pulse in terms of at the highest levels what Fortune 1000 and some global brands are doing in terms of their data and analytics and now AI investments.

Jim Merrifield (01:20.487)
Yeah, that's amazing. you know, what's one thing that you're trying to measure every year with this benchmark, Randy, because you've been doing this for 15 years.

Randy Bean (01:29.998)
Yeah, it's really progress because, you know, that's what's tracking each year. Progress in terms of implementing data and analytics and AI initiatives, progress in terms of the role when first started doing the survey in 2012. Only 12 % of organizations had a chief data officer in place. It wasn't until five years into it, 2017, where over 50 % of organizations had a chief data officer.

And this year, 90 % of organizations reported having a chief data officer and another close to 40 % had appointed a chief AI officer. So we've come a long way in those 15 years.

Jim Merrifield (02:14.729)
yeah, absolutely. That's That's a big gap in in percentages. So you could definitely see the positive trend up for these roles. Was there something in the survey for 2026? What surprised you with the results? Anything in there?

Randy Bean (02:34.094)
Well, I mean, I'd say the most important thing is that data is now recognized as critical because great AI depends upon great data. I believe it was roughly, wherever the number is in here, it's roughly 96 % of organizations that now say that making sure that data

is it has as good a quality as possible is a prerequisite for for strong AI. So I think that's probably one of the most encouraging trends. And then another one is the level of companies that have moved from experimentation to implementation of AI at scale. That was a pretty sharp increase over previous years. So I would highlight those two things in particular.

Jim Merrifield (03:28.871)
Yeah, very good. I mean data is certainly king these days. Absolutely so. Randy give me a plain English version. What is peak AI look like inside?

Randy Bean (03:42.511)
Well, you know, having lived through previous periods of excitement and bubbles and then successful implementation in the context of the internet, you know, if you go back to 1990, 2000, 2001, just before everything collapsed, there was this euphoria.

or intoxication about what the internet was going to be and whether it was going to change everything. And the term that I use these days in terms of, I think it was very true for the internet and it's true for AI is that many organizations and individuals are overestimating the impact that it will have on the short term.

underestimating the impact that AI will have in the long term. I've just been reading over the past day or so this paper from anthropic founder Dario Amodei where he talks, he wrote basically it's close to a 60 page paper that he posted just this past weekend on what he sees as the future of AI and it's scary.

on the one hand, because part of his thesis and he's operating this on a day-to-day basis is that AI, as soon as within the coming year, will have the equivalent capacity of 10 of a thousand Nobel Prize winners in a particular scientific discipline in the same room at the same time.

You know, he's really focused on how do you channel AI to achieve the greatest positive impacts and benefits and mitigate the very significant risks that can be posed as well. ultimately, you know, AI is a tool, you know, like any other tool, if you think about it historically, you know, the hammer, the screwdriver.

Randy Bean (05:56.579)
the printing press, railroads, airplanes, the internet. So things can be used, essentially they're neutral, but they can be used for positive benefits such as in healthcare and life sciences, improving life expectancies and the quality of life. Or they can be used in a detrimental fashion. And often what you see in social media these days is some of the dangers in terms of misinformation, disinformation.

selective communication of information. So, you know, these are the issues that we'll have to face as business leaders and as a society and as citizens in the years ahead.

Jim Merrifield (06:39.783)
Yeah, I'll have to check out that paper. For sure you have to send me a copy. didn't get a chance to read it, but it sounds very interesting. know, nice takes.

Randy Bean (06:48.406)
And it is posted on, I posted it this week on my LinkedIn feed. So that's another place to go.

Jim Merrifield (06:56.955)
Okay, I'll have to check it out. So here's a question, Randy. What's your fastest smell test for whether a company is getting real value?

from AI.

Randy Bean (07:11.078)
It's really simple from my perspective and that is are they building things that deliver value to their customers or help the organizations grow or are they just building things for the sake of building things? So for me it's, you know, what I often tell audiences is that if you're not getting measurable business value or have a clear path to measurable business value from your data and AI investments, maybe you should go back to the office and shut them down this afternoon.

Jim Merrifield (07:43.631)
That's wise advice there. let's talk, speaking of investments, let's talk money. Your executive summary says investment is basically universal. 99.1 % says investment in data and AI is top organizational priority. And about almost 91 % say their organization is increasing investment.

When almost everyone is spending more, where are leaders still wasting money?

Randy Bean (08:15.63)
Well, you know, I go back a little bit to the previous answer that I gave, and that is that I think it's really important for organizations to see how they're using AI, for example, in terms of improving the customer experience, improving customer service, improving the ability to acquire customers, retain customers, grow those customer relationships, because I think that

Most businesses, their franchise is their customers. So the more that they're doing to use AI to improve the customer experience, I think that that's where you can see measurable value. Things that are behind the scenes, yes, clearly there's operating efficiencies to be gained, but I think ultimately all of those activities need to serve.

serve the customer because they're why these companies are in business.

Jim Merrifield (09:19.111)
Yeah, hey listen, it's all about the customer. You don't have a happy customer, it's tough to increase that revenue.

Randy Bean (09:25.91)
Absolutely, yes.

Jim Merrifield (09:27.699)
Yeah, so you also note that integrating AI into mainstream business processes really took off after Gen. AI arrived in November 2022. for CIOs listening to this podcast, what changed operationally after that?

Randy Bean (09:48.237)
Yeah, it was really about starting to work with unstructured data as opposed to structured data, know, numbers and so forth. And it was really the development of language models so they could take this unstructured data and people could speak and ask questions. So that's really what made AI accessible to the masses through things like ChatGPT , and it's, you know, subsequent

capabilities like Claude and others. that's really what democratized the popularized AI within organizations and more broadly within society.

Jim Merrifield (10:31.537)
Yeah, mean, you say great AI depends on great data. We talked about it earlier. And 92.7 % report that interest in AI has led to a greatest focus on data. So here's a question. If you get dropped into a company as the new CDO, Chief Data Officer, what's the first data foundation you'd fix so AI doesn't wobble later?

Randy Bean (10:59.096)
Yeah, there's a couple of ways to answer that because organizations have been struggling with their data for as long as I've been in the industry. And one of the things that I've heard over and over, over the years, be it five years, 10 years, 15 years, 20 years is when organizations are faced with these issues, they say, no, not another data project because be it data warehouses, data lakes, data fabrics, data meshes, all of them have

in many respects, fallen short of what is the ultimate goal. And so what I urge organizations to do is to focus on what are the critical business questions that you're trying to answer and what are the data that you need to answer those critical business questions, because it's not that you have to make 100 % of your data perfect. It's more like the 80-20 rule. Sometimes it's a case where 5 % of your data

answers 99 % of your most critical questions. So my answer would be identify the most critical business questions, what data you need to support those questions and focus on getting that data right and worry about the other data as you need to over time.

Jim Merrifield (12:13.043)
Yeah, that's a reasonable approach 80-20 rule, right, can be applied to many different initiatives and projects. think, you know, AI projects is no different for sure. So, you know, this was a big headline in the paper, AI production at scale increased from 4.7 % to 39.1 in two years. And a limited production grew from 

24.5 to 54.5 percent. I mean that's that's incredible growth So yes for so for a CIO for a CDO. What's the real difference between? You know, we have models in production and we're in production at scale

Randy Bean (12:48.738)
Yes.

Randy Bean (13:01.27)
Yeah, you know, the kind of the question over the past couple of years was, okay, enough with playing with this stuff. You know, let's plug it in to our core processes and see where we can deliver business values. So that's what organizations have really, that's the big leap that organizations have been making. In other words, it was important and necessary at the beginning to experiment. But at some point you have to deliver.

a return on investment. Organizations, particularly public companies that have to post quarterly results, they need to show that these investments in AI are delivering a return. And so that's what's really driven organizations moving from experimentation into full-scale production.

Jim Merrifield (13:49.651)
Yeah, and you also in the paper, you showed a broad shift, right? Two years ago, 29.2 % had any AI in production. This year, it's 93.6. I mean, that's unreal. you know, here's a question. What breaks first when teams try to jump to scale before the operating model is ready? Because a lot of them say that they're in production. There's almost 94 % of AI tools in production.

is there's probably a risk, right? To, to, push it to production too fast.

Randy Bean (14:27.138)
Yeah, I mean, ultimately, it's all about organizational readiness or culture and people. you can't force new capabilities, new technologies, business processes down people's throats. You have to engender their buy-in so they see what is the benefit to them in terms of their role and their mandate.

So the biggest challenge that organizations will face, you know, right now and over the next three, five, 10, 15, 20, 25 years is really around organizational adoption from the perspective of people, the employees, their skills, their buy-in, their incentives, their understanding. So that's where more more organizations are coming to the recognition that

You know, there's not a limitation or a gap in technology. The challenge is, is people and culture and how do you align the people with the culture of the organization to achieve the business impacts that you're hoping to realize.

Jim Merrifield (15:39.923)
Yeah, and I think you saw that right during throughout the survey. I 93%, I think say.

Randy Bean (15:50.167)
Right. Yeah, there's no shortage of technology. But it's it's you know you can only absorb and synthesize so much and have an understanding and be ready and see what the applications are. So because otherwise people and organizations that become overwhelmed, you know, you have to balance that, you know, right now there's a lot of, you know, formal fear of missing out or chasing the shiny object.

and what organizations really need to do is take a long-term view. Where do they want to be as a company three years, five years, 10 years from now? And how can they best leverage artificial intelligence in their businesses to accomplish that rather than doing things just for the sake that this capability exists? The more thought through from a long-term perspective an organization approaches this,

the greater likelihood, higher probability that they have of success in the long run.

Jim Merrifield (16:55.015)
Yeah, because you're really changing behaviors and that spans throughout the whole organization.

Jim Merrifield (17:07.571)
change agents can be the change person within the organization. I've seen a few titles out there in firms and organizations.

and change. I don't know, that's a pretty big role there, you know.

Randy Bean (17:22.67)
Yeah, I mean, change management transformation is the big opportunity and challenge that organizations face now, but it's never an easy role and it combines a lot of different functions too. It combines HR and legal and compliance as well as IT as well as business function. So it's kind of a case of

every all hands on board and everybody coming together to think through what the issues are and what the outcomes are and where how AI can be used to have the greatest business impact. So it's really, it's really a complete organizational commitment. It's not just the chief data officer or chief AI officer.

will not be effective unless they have the active participation of their colleagues across the lines of business and the core functions.

Jim Merrifield (18:20.667)
sure, 100 % agree. So, but who owns AI? You you flag another flaw that companies are naming chief AI officers, right, but around, you know, 38%. But there's, there's really no consensus on where that job should report. And it seems to like the, the reporting lines are split, you know, to the CDO function, to the business, to transformation. So, you know, first question on that,

topic is why is it still so messy?

Randy Bean (18:54.466)
Well, because it's a new role and, you know, the chief data officer role is only 15 years old. And they, you know, there used to be a joke about the chief information officer, the CIO a generation ago. And the joke was that CIO stood for career resolver. And I remember telling this to, and telling that story to an audience.

in London about two years ago, and somebody raised their hand and said, well, I know what chief data officer stands for. And I said, well, what is that? And they said, career is definitely over. So the point there is that these are new roles. They evolve significantly. mean, the chief data officer originally started as a risk and regulatory and compliance role. It's evolved more into a business role. Now it has the AI component in some instances. In other instances, AI is a

parallel organizational structure, but my personal belief is that ultimately these functions are business functions and should report into the business and should be aligned with the business goals and priorities and objectives and mission of the organization.

Jim Merrifield (20:04.861)
It makes sense. all about the business. mean, and every, there's probably no, you know, one size fits all approach. It depends on the organization, the personnel.

You

Randy Bean (20:26.326)
Yeah, you know, I was talking to speaking with somebody earlier today and they were talking about organizational transformation. remember a few years ago meeting with the new chief data officer of in this case, it was a regional bank and I had worked closely with that regional bank over a couple of decades and I knew the pace at which they moved and the resistance to change.

And when I first met their new chief data officer who had had a track record of great success, I was listening to him and I was thinking, you know, this is a very impressive individual. And then they just start describing to me the blueprint they were going to bring in to change the organization. And as I listened, I started saying to myself, yeah, it's not going to happen. Not within this organization. it happened within organizations that they'd been in the past, it could happen.

at other organizations, but for this organization, it's such a massive cultural transformation. And sure enough, you know, that person joined and initially everybody was on board and in the room, they have 100 people meetings. And, you know, I was watching this because, you know, I didn't believe that could be successful in this context. And then six months down the road, you know, half the number of people were in the meetings. And then a year down the road, a quarter of the

number of people were in the meetings and then after that, you know, nobody was coming to the meetings. And I was just waiting for the day that I got the call that the person was no one within the organization, which was about after a year and a half. understanding the culture of the organization, one could see from the outset that the outcome was going to be inevitable, barring a miracle.

Jim Merrifield (22:11.645)
Yeah, well, hey, well, thanks for sharing that story there. So here's a question for the CEOs out there. What's one question you wish every CEO asked quarterly?

Randy Bean (22:18.754)
Thank you.

Randy Bean (22:32.162)
I think they need to balance the risk and rewards. So they need to understand and appreciate the great value that AI can bring with an organization, but they also have to have an appreciation of the safeguards and guardrails that they need to put in place because, you know, if you're a legacy company that's existed for generations or over a hundred years or longer, you know, your customers or your franchise and

There's the old saying about a reputation, it takes like a lifetime to build, it can be destroyed in a moment. So same thing with customer franchises and there's been organizations, even during the internet era that were, they were at top of the game 30 years ago and now those companies don't even exist anymore. Sears and Roebuck for example.

Jim Merrifield (23:26.467)
wise advice. So listen, we've come to the moment of the lightning round, Randy. So one sentence answers, right? So don't overthink it. So here's the first one. What's the fastest proof your past peak AI hype and into real value?

Randy Bean (23:31.511)
I'm

Randy Bean (23:47.982)
I would just go back to customer experience if If your customers are having a a better experience as a result of AI.

Jim Merrifield (23:58.355)
Perfect. Here's another one with 99.1 % calling data and AI a top priority. What is still underfunded?

Randy Bean (24:09.454)
I'm going to go back to people and process.

Jim Merrifield (24:14.387)
So AI pushed data focus for 92.7%. What's the first data fix you'd make?

Randy Bean (24:26.694)
focusing on just the data that you need to answer your questions, the 5 % or 20%.

Jim Merrifield (24:33.223)
Perfect. So 93.2 % say culture changes a blocker, like we talked about. What's one change move that works?

Randy Bean (24:42.926)
I'm sorry, what was that again?

Jim Merrifield (24:44.755)
So 93 % say culture change is the blocker. What's one change move that works?

Randy Bean (24:55.566)
Aligning your technology objectives with your business goals.

Jim Merrifield (25:01.064)
Here's the last one. So CAIO reporting line your default vote and why

Randy Bean (25:08.907)
Report to the business leader, like in the case of JP Morgan, the chief data and AI offices sits on the 14 person operating committee reporting to Jamie Dimon

Jim Merrifield (25:21.661)
make sense. I love real life examples. listen, Randy, thanks so much for taking the hot seat today. You've been a great sport, and especially with the lightning round is a fun time. And if you're a CDO, CAIO, or CIO listening, the investment is nearly universal. Production is happening. And the hardest part is still organizational design and us humans. So that's it.

for the InfoGov hot seat. See you next time.

Randy Bean (25:52.29)
Great. Thank you, Jim. My pleasure.

Jim Merrifield (25:55.699)
Thank you.