David: You conduct research on a regular basis among the community of CDOs, which provides some rather interesting and useful insights, one of which is a health check on the domain of data as a focus of investment for major organisations. So, I wonder if we could start by you sharing some insights into what that looks like at the moment.
Randy: Yeah, absolutely. AI has really elevated the question. I’ve been doing a survey of data leaders for the past dozen years and 96% of the respondents hold the title of CDO. The 2025 findings will be published soon, but just some quick specific data points on investments. When asked if investments in data and analytics are a top organisational priority, 90% said yes. And 83% stated that their organisation is increasing its investment in data and analytics. So, investments in data are only increasing.
David: I’m sure everyone in the room is delighted to hear that, because there are pressures on most budgets. Our own research shows that there is a general upward trend for 2025, but there is still increasing investment elsewhere, led by technology. One thing we’re talking about a lot today is AI and generative AI. I wonder to what extent that focus organisations are taking on AI and generative AI – is that going to draw a lot of funding because it is not a cheap undertaking? Do you see any tension emerging between those two areas, whether the role of data and the role of AI could lead to somewhat of a power struggle?
Randy: Let’s talk about the good news: the demand for data and AI leadership is only going to increase over the coming years. The bad news is that the CDO has been a challenging position for organisations to fill and for people to be successful in. I can speak primarily to the US, but the role of the CDO really came into existence after the financial services crisis of 2008-09 and at that time it was largely mandated, particularly for major banks, as something to get the data under control from a risk regulatory and compliance perspective.
So, this means a couple things. First of all, it is still a relatively new job. If you look at the creation of the Chief Information Officer role a generation ago, the joke at the time was CIO stood for “Career Is Over”, so it should not be a surprise that the CDO role has a lack of expectations. There’s a lack of concern. What is the scope of the job? What does the job report into? And then it has consistently evolved over the past 15 years, moving from largely a defensive role to more of an offensive role and integration of the analytics function – and now you have AI.
What we are seeing is a number of organisations are appointing chief AI officers that appears of the chief data officer and other organisations. For example, I wrote for two years for Wall Street Journal, and currently for Forbes and Harvard Business Review. In July, I wrote a case study about Capital One, and Capital One has appointed their first ever chief AI officer. I wrote about Cleveland Clinic; a month later, they appointed their first chief AI officer. And then a few weeks ago, I wrote about USAA, and I asked their chief data officer, “do you have a chief AI officer?” They said, “no, I own AI.” There is that tension where some data leaders think strongly that AI should be part of their responsibilities, and some organisations think that it shouldn’t.
David: Well, it’s good that you pointed out that the role of CEO emerged out the financial crisis, and then the need for banks to get their arms around their liabilities and their customer base. We’ve seen the US government mandate chief AI officers across every federal agency, but what we saw with the CDO was that it rolled out into the commercial sector. Do you think the same thing will happen with CAIO becoming commonplace everywhere?
Randy: It’s literally impossible to predict, but it is very much in play, and I’ll give an example. Last month, I went to the Wall Street Journal Tech Live in Laguna Beach, California – about 200 or so people. And last year, the keynote was Sam Altman from OpenAI. They asked Sam: “artificial general intelligence? How do you define that?” And he said, “that’s when a machine can do any task as well as a human being.” I then asked, “when do you think that will become available?” And he said, “if you’d asked me two years ago, I would have said, the next decade. If you’d asked me a year ago, I would have said the next five years. If you’d asked me six months ago, I would have said in the next three years. And if you’re asking me tonight, I need to go back to the lab and check.” Things are moving so quickly and at such a pace that it’s really hard to predict how things will play out.
But one thing that’s for certain is that large organisations tend to adopt technologies and tend to change and tend to transform slowly. For example, back during the pandemic, I was speaking with the CDO for Net Life, a large insurance company in the US, and he said, “we’ve done more to execute on our digital transformation strategy in the past six months than we had in the previous 20 years.” So sometimes it takes something that’s a critical factor. In other words, you may lose your customers, or you need to find a better way of serving your customers. The technology may be there, but unless there’s a compelling need to do something, most organisations won’t move that rapidly.
David: Back to the territory of the CDO. You’ve expressed some quite strong views about how CEOs have often been set up to fail. How does that happen, and is there any way to avoid it?
Randy: That’s a complex question, because the responsibility cuts both ways. I’ve lectured organisations on the need to set clear expectations and to have a better sense of what it is that they expect out of their CDOs. For the past four years, I speak to the cohort at the Carnegie Mellon Chief Data Officer Program, and I spoke to them a few weeks ago. I said, “help me understand something. All of you are going to this class to get your certification to become a chief data officer. The average tenure is two to two-and-a-half years, and now it’s even shrinking. Are your eyes open to this? Are you sure that this is what you want to sign up for?”
And I thought I was delivering a tough message, but I got all these messages afterwards saying thank you and that it was helpful in terms of grounding. So again, the demand for data and AI leadership is only going to increase it tremendous opportunities, but what shape and form it takes is unclear.
David: I wonder whether the use of chief and officer within this space was actually the right thing to do. Has it implied a certain level – C-suite – within organisations that hasn’t always been open and available?
Randy: Well, I should say there’s a lot of reasons why CDOs have been unsuccessful. Aside from all the issues with the businesses not having a clear direction, one is the use of jargon. I don’t know how many businesses have said to me, “I have no idea what they’re talking about.” Another is that there is that CDO is a fundamentally change agent and transformation agents, and no organisations, despite whatever they tell you, really want to embrace change.
I don’t know how many instances where I’ve seen where somebody go into an organisation say, “here’s all the transformative things that we’re going to do,” and there’s a lot of enthusiasm, and the organisation is excited. But then as they see what that requires, in terms of the change in the transformation that takes place in the organisation, the enthusiasm wanes. You can start to see at the first meetings, there will be hundreds of people from the second and third meetings, and then by the time you get a year into the tenure, there are just three people showing up into the room, and you can see the writing on the wall.
I did an event in Boston last December, and the first day was billed as the 10th Chief Data Officer Summit, and the second day was billed as the first Chief AI Officer Summit. And there was the first day, there was a lot of talk about the challenges being faced, and asking why don’t the business leaders appreciate us and understand us better? On the second day, it was all green and rosy; a beautiful future and so forth.
I said to the CDOs, “I have the solution. Just call yourself a Chief AI Officer and in the next two to three years, you’ll leave and then you’ll be held up as an example. Nobody will bother you until they reach that point that and then you have to figure out what job title comes next.”
These roles are not easy. I’m going to go back to the point about the critical need for data and AI now more than ever. Great AI depends upon outstanding data, so the demand for data leadership is going to be more than ever before.
David: When you were talking to those students thinking about becoming CDOs, and to a community like this, what suggestions do you have for people in that role, or looking towards that role, as to how to ensure they are future proofing themselves and keeping the people in the room that they need to be talking to in their organisation?
Randy: I say, look at things from a long-term perspective. In other words, adoption of generative AI and adoption of other AI capabilities, isn’t going to happen overnight. Think in terms of a three-to-five-year plan with a replace out over ten years. With a replace out over one year is really beside the point.
When I started my career in banking, I had a mentor who said to me, “nothing ever happens in this industry in less than 10 years.” And I thought that was incredibly cynical. This was about 35 years ago, but I’ve seen that play out over time. So, despite all the excitement about generative AI and all the excitement about AI in general, how it plays out in terms of being integrated into core processes, how it plays out in terms of improving the customer experience, that will tend to be more gradual.
Ultimately, there will be a huge transformation, and it will change how we work – it will change our lives. This is a revolution, but the adoption will be evolutionary.
David: The language you just mentioned is all about transformation, about huge disruption and change. Yet, most of the instances that we tend to see are relatively straightforward and a lot of focus on process efficiency gains. Do you have a sense of that timeline? What is the horizon when we’ll look back down the road and then state, “wow, we don’t even recognise ourselves from X number of years ago”?
Randy: The survey that I have been doing for 12 years, one of the questions was “what is the major benefit you’re going to get out of generative AI?” I think I had listed productivity gains, things regarding human in the loop and improving processes, and several people pointed out to me that I had nothing in terms of growth. It was all cost reduction, efficiency improvement of processes, but I didn’t have any answer in terms of how they were going to use generative AI to transform their businesses in completely new ways.
It highlights the point about things moving so quickly, both while I’m saying that things play out over time. At the same time, a lot of these capabilities are coming along so quickly that one of the things I tell audiences is that the answer that I give next week is not the same as I give today. It all depends upon learning and interacting with various groups, because things are happening quickly.
David: So much of that pace of change is being driven by the super six developers of the foundation on the frontier models, which is also where regulation is focused and the concerns about safety and responsible AI. To what extent can the organisations in this room have any meaningful impacts on that direction of travel?
Randy: If you look at the Fortune 1000 in the US, 90% of Fortune 1000 companies are legacy companies, meaning they’re older than 50 years. That 90% of the Fortune 1000 don’t have to compete against the big six, if you will. They just need to compete against the other 90% so their rate of adoption and their rate of transformation is going to be more gradual. It’s going to be more uniform, and that is going to depend upon the safeguards and guardrails.
They are going to take a more conservative and gradualistic approach, which will play out over time, versus the big six or that 10% of the Fortune 1000 that will move at a different pace.
David: Thinking about the CDOs and how they maintain their own momentum, you talked about the difference between defensive capabilities and the offensive capabilities. The defensive side of understanding the governance seem truly relevant to AI, but at the same time, you must demonstrate that offensive purpose of adding value, returning on the investment going in?
Randy: I think the most important piece of advice is if you’re not delivering business value, you’re not doing your job. Bluntly, you’re here at the behest of the business leaders. So sometimes I say to audiences, in part to provoke discussion, “if you’re not delivering business value from your data analytics and AI investments, you should shut them down this afternoon.”
It’s all about delivering business value. That’s why we’re here. Delivering business value in terms of improving the customer experience, retaining customers, acquiring customers, all of those type of things. That’s the justification.
To get involved with unique member only discussions that examine topics such as AI investment and securing data change, click here.
To showcase your AI transformation and earn the recognition you deserve, enter the DataIQ Data & AI Awards.
