Loading summary
Transcript33 lines
- [00:00]
A
Foreign. Welcome back to another episode of the AI to RRI Big Story Edition. I'm Ray Reich, founder and CEO of Benchmarket. And joining me as always is my co host, Peter Buchanan.
- [00:21]
B
Yep, I'm Peter Buchanan. I'm the founder of New Plant. It's great to be here. And today our topic is one our listeners have been waiting for. The rise of the Chief AI officer.
- [00:34]
A
That's right. This week's Big Story is focused on that rise, the Chief AI Officer, which by the way, I'm just excited to see how this rolls out over the next one, two, three to five years. But we're going to get a little help from IBM's Institute of Business Value because they conducted a research of the that included almost 600 plus chief AI information officers, or maybe I should say chief AI officers across 22 geographies and 21 industries. The headline that AI leadership is becoming non negotiable for enterprises. So hey, let's dive into our top five takeaways from this week's edition. Sound good, Peter?
- [01:14]
B
It does. Sounds great.
- [01:15]
A
Okay, so the first one we've kind of characterized under the chief AI officer gap is costing companies real money. So let's talk about what are the stats. So 26% of companies currently have a chief AI officer. And let's be fair, that's 26% of the 600 plus companies that IBM actually conducted a survey with. Now the good news is that that is up from 11% in 2023, so almost 2.5 times more. So the number is growing. But it also means that about 3/4 of enterprises are still trying to operationally AI without a dedicated executive at the helm. And to me, Peter, that may create some issues. What do you think?
- [02:05]
B
Well, I mean the first thing is there's basically a price gap or business penalty attached to it. So companies with a chief AI officer get 10% greater ROI out of their AI spend and that ROI goes to the bottom line or it goes to increased revenue or it goes to increased productivity. And they're 24% more likely to report outperforming their peers. So the ones that have gone to chief AI officer early are getting a competitive advantage against their their rivals. It's sort of like the difference between using AI to win a cool science fair experiment, competition in high school versus actually getting a meaningful competitive advantage to put your company ahead.
- [02:57]
A
Well, for the AI to ROI podcast, it's really good to see this data. It sounds like maybe we named it right. And I really appreciated how because you write a lot of the newsletter and this one from Monday, March 2nd. I love how you framed it that most companies don't lack AI ideas. What they lack is the ability to turn those ideas into projects that transition to production scale initiatives to deliver measurable roi. And often what was found in the research is that chief AI officer was the executive in charge of fixing that. In charge of fixing what hasn't yet turned into roi.
- [03:41]
B
Right. So now the question becomes who becomes a chief AI officer? Where do they come from? So Ray, let's bust a myth here. These aren't techno geeks that have been in the lab coming up to go home to their families. And you never see them otherwise. When you hear that, you picture that. But they're actually coming from a different place, It's a different hiring area. So where do they come from?
- [04:13]
A
Well, it makes sense that 73% of existing chief A officers from this research had more of a data focused background. Data science, statistics, analytics and some level of AI research. But it was clear that the most effective chief AI officers are truly hybrid leaders, equal parts business strategist and data scientists. So that can, that's where it translates AI capabilities into that measurable business value and be able to communicate that value to the board and honestly to the investors. And, and to me that's really important. It is.
- [04:59]
B
And here's the part that's really interesting is 57% of the CIOs in this category were actually promoted from within. They're not coming from outside. And so what enterprises are realizing is that cultural fluency and institutional knowledge really matter as much as technical expertise because a lot of these AI projects have succeeded, have failed to generate roi, or have crashed and been shut down because for business and cultural reasons as opposed to technology reasons. So the CAIO who doesn't understand the business is at a real disadvantage and it really hurts the enterprise.
- [05:46]
A
Yeah, you know, I have an analogy here. Everyone's talking about how vertical AI is going to be the early winner. We can really encapsulate subject matter expertise, domain specific knowledge, like for the healthcare industry versus financial services, those are going to be the early winners. And it sounds like here people who have institutional knowledge. Right. Of a company, because the way that Dell operates is very different than the way PepsiCo operates. So you need to understand the unique domain of a Dell. And then you also need to have skills such as change management. Because one of the biggest issues in any traditional transformation initiative, things like digital transformations from 10 years ago, it was those change management skills, including the communication skills to communicate the Vision the goal to not only the employee base at large but but to the board, to the technology team who may be building your agentic AI internally and familiarity with AI governance. Because AI governance is not just internal governance and policies, it may be industry regulations and coming soon. And we've covered this with a previous newsletter. Governmental regulation, whether it's state based or federal based or state, you're a multinational company, you need to understand AI regulations in the UK versus Japan versus the us so they're much more than a technologist. Peter.
- [07:20]
B
Oh totally. And also where they report in the organization really matters. So it's not just about whether you have a chief AI officer, it's where they sit and whether they have, whether they're basically anointed to act. Know whether they can have some teeth in the organization to get things done.
- [07:41]
A
Yeah, you know what, that's really important because think about you have project offices and large Fortune 1000 companies and they're project managers and their job is to implement what the functional executive or the executive team. Here it's almost ambiguous in what the findings are of what's leading to success of AI initiatives. Number one, that that chief AI officer exists and then reports directly to either the CEO and even maybe a dotted line to the board of directors themselves because that proximity signals that AI is a strategic business priority to entire company and not limited to being a technology initiative. And when we think about the power of agentic AI and it's going to be how we fundamentally transform and increase productivity of end to end business processes that are going to cross multiple functions and having the CIO also reporting to the CEO that also gives more control of budgets, it helps to drive enterprise wide adoption. And at the end of the day, just like our podcast is named deliverable, measurable ROI from those AI investments.
- [09:00]
B
Right. And in fact 61% of the CAIOs and IBM survey control the AI budget directly. So if the chief AI officer doesn't control the purse strings, they're essentially an advisor, they don't have an authority to act. So there's a structural model that really works. Appears to work well in IBM's research. What's that? Ray?
- [09:26]
A
Yeah, it's really more of a hub and spoke model. When I first was reading this, it's like oh, I don't want anything that looks like an airline. But in this case what they're really meaning is you might have that central AI executive and function, but they partner very closely with individual business units. And it's interesting, I recently conducted a finance maturity models research and what we found was the highest maturity companies were those who had the finance function distributed throughout the business business units and the functions. Same thing here. When you have CIOs who are really partnered with the individual business units, they may even have AI business partners or AI business analysts embedded within the business unit or the function, you know, within supply chain, within, within manufacturing, within distribution. And this model achieves a 36% higher ROI than those in decentralized structures where supply chain makes their own AI decisions, but without the benefit of having that expert who reports into that centralized AI organization and allows you to get governance and accountability in a centralized model. So you, your chief legal counsel is very happy, your compliance officer, your chief information security officer, they're really excited because now you have some centralized accountability, but with the agility to execute at the business unit level.
- [11:02]
B
Right? So that, that leads us to determining what makes the chief AI officer successful or not. So research identified three critical success factors for the caio and I want to walk through them, us to walk through them because they're deceptively simple. So Ray, why don't you take the first one?
- [11:25]
A
Well, I appreciate you giving me the first one because it is basically who I am and what I've been doing for the last five years with the business. And that is measurement. And let's be clear, measurement equals metrics. In my world, do we actually know what success looks like? And this is where so many AI initiatives fall apart. There was a great vision. I'm going to fundamentally transform the efficacy of my marketing organization or my sales team. But how are you going to measure that? So in a lot of common mistakes is, well, I'm going to have some upfront measurements, I'm going to increase productivity. And my marketing resource and demand gen can do 1.5 times more campaigns or they launch a campaign 50% closer. But those are nice leading indicators. But a year later it's like, hey, did I increase my efficiency as measured by more work for less cost? Did I increase the efficacy for every hundred dollars I put into a marketing campaign, am I getting more back as measured by pipeline? So this chief AI officer's job is to ensure that before a single AI product is deployed or a single line of code is written, if it's an internal build, that those defined success metrics are tied to business outcomes, including those short term leading indicators I just talked about and the near to long term measurements that impact cost, impacts revenue, true return on investment. Hey Peter, you know, so hi, here I am, I'm the metrics guy. But that's only the first kind of pillar. What's the second one?
- [13:07]
B
Well, the second is in my world as a consultant to CEOs and boards of tech companies and that's teamwork. So are the right people aligned? So the Chief AI Officer is, he's kind of like an offshoot of a Chief Product Officer or a head of product, you know, ahead of product. As in a tech company doesn't have a lot of direct reports, but they have to go to the technology team, they have to integrate with marketing, they have to support sales, they have to make sure that their products can be onboarded, they have to get feedback from support. And the CAIO has a very similar role in that there are relatively few resources that report to them, but the impact can be absolutely tremendous. They go first of all across the C suite. Their relationships are cto, cio, Chief Data Officer, Chief Human Resource Officer, the ciso, the cfo. Those are all people who are interested in the overall effect of AI on the business. But critically, the business unit P and L owners are where the rubber meets the road because that's where you're deploying the AI use cases that are supposed to transform the company. And so if you're not able, we mentioned in the section up above the hub and spoke model, this is really how this actually works because the C A I O would be at the hub of the AI function, but they're touching 25, 3, 30, 50, sometimes hundreds of people to get their job done in eight or ten functions inside the company. So being able to navigate that, to get these use cases deployed and productive and generating roi, that's their job. Teamwork is almost at least half their job.
- [15:20]
A
You know, it's funny, this reminds me back when I was at, you know, GE and you know, it's where we met, right? I was put in charge of this cross functional initiative and included people from all over the world. But what I didn't have was I had accountability because the CEO said, I want you to report back to me how this program goes. But I didn't have authority. And one of the things that this research found was, boy, that Chief AI Officer needs to have real power to act. You know, just being a glorified office of Project Management, special projects doesn't cut it if you don't have the authority to actually make some decisions. Now you may need to really work hard on the collaborative side to get your peers on board. But end of the day, you need to be empowered to act and make decisions because this is where it Loops back to, you know, this point is you strip away any of these three pillars. The really disciplined focus on measurement, the ability to really build a team and have teamwork and collaboration, and third, the actual authority to make some tough decisions. You're going to have reduced ROI and thus you're not going to really have AI to roi.
- [16:40]
B
Right. Which is sad. It's really sad. So let's get to the last of these key points is let's say your company isn't ready to have a chief AI officer, but you need to put AI into your company. So our fifth point is you might not want to have the role yet, but you have budget constraints, you have organizational challenges, you don't have leadership bandwidth to actually do it. The responsibilities don't go away just because you don't have the title inside your company. So how are companies doing this when they're not ready to hire a chief AI officer, but they still need to get the benefit?
- [17:25]
A
Yeah, and this will also, whether it's a large enterprise that just haven't made that strategic commitment or investment in the chief AI officer or mid market and smaller companies, it's like I can't afford to have a dedicated person. So some of the options are you can have an AI steering committee, but that AI steering committee needs to be not only cross functional, but senior executives who not only meet regularly, but have specific objectives to implement AI from both a.
- [17:59]
B
What's your right? It's like a lot of companies have an investment committee before they spend significant dollars and you have to go to present to it to get the money. You need something similar and there's all senior people on that and they're usually pretty tough on the people that come visit them. So is this steering committee similar to that?
- [18:19]
A
The steering committee is, but you can have even a little bit more structure and you can have a center of excellence. And this center of AI excellence, you still maybe have a senior executive who leads that. Maybe they're not your chief AI officer, but they have been empowered by the CEO to say, hey, we want you to have some key project managers, almost like a program office. Have some real subject matter expertise on AI and you might even source some people like a top high profile, high potential person from marketing who goes into the center of excellence for 12 to 24 months, someone from supply chain, someone from finance, and they can help standardize the AI policies, the governance framework, the methodologies track and even help roll out. But it's really a center of excellence where you don't really have a chief AI Officer.
- [19:17]
B
Right. So that, that works as long as companies say, okay, well you've got a 45, 50 hour a week job, you're allowed to put 12 to 15 hours into this rather than having a side desk project attached to it. Right. You know, you gotta be serious about it.
- [19:33]
A
Might even go further. You might say, hey, this is so strategic. I'm gonna take that high potential person and make them the head of the center of excellence. And that is 100% their full time responsibility. Then they have people from each department or business unit where to your point, maybe it's a one day or two day a week job, but heck, it might be a full time job also depending on the size of the entity and in the level of strategic commitment to AI.
- [20:00]
B
Right. So most critically, someone needs to own AI roi accountability. So whether it's the cto, cio, business unit, lead ambiguity is bad. Like, so the success of these projects needs to be attached directly to the people who will be implementing and benefiting from them. Right. So. And that's not all you need. Governance frameworks for ethics, risk compliance. There's a lot of risk in some of these AI products because they're new and sometimes they're unpredictable unless they're really implemented in a fantastic way. My feeling is actually, of all the things in this newsletter, accountability has got to be right at the top of the list. Whether it's for the cause, accountability is certainly there for a chief AI officer, but it's also there for the people who benefit and implement. Right. They have to have shared accountability.
- [21:08]
A
Yeah. And that accountability also goes into, you know, how do I really leverage the collective experience, knowledge, best practices from across the company and bring that back up for the entire enterprise. And one data point that came out of the research which really kind of galvanized this for me and it was technology centric, but the average kind of large scale enterprise is using an average of 11 generative AI models. So. And by the way, I've seen it in the real world talking to CFOs at Fortune 1000 companies. Oh yeah, this, this organization is using chat, this one's using Claude. I got someone using Huggy Bear and they have a couple small language models and then there's that shadow model. It's like, yeah, it's not here. But hey, I, I found out one of my employees was using Deep Seek at home late at night because they really liked the capability of Deep seq. So that's one example. And as we see the rise of agentic AI, we're going from using a large language model to help us analyze data or make my emails a little bit better. But we're now going to have AI agents doing the work that used to be done by humans. The complexity of coordinating that across the company between departments, leveraging the existing systems of record and the data that you need to both use for the AI agent and then write back, that's going to be a huge challenge. If you don't have a, a challenge if you don't have a centralized organization and individual responsible for that, it's going to become a hot mess.
- [22:50]
B
Oh, no doubt. All right, so let's wrap it up. So the bottom line is this coming from our newsletter article. First, organizations with CIOs outperform companies without them in AI and the gap is actually widening. So 66% of CIOs expect that most organizations will have someone in their job within the next 24 months. And the window for competitive advantage through putting this function in place is relatively narrow because eventually everyone's going to have one. So if you can afford it and you want to be forward leaning in AI, now's the time to act.
- [23:39]
A
And the other data point I loved, we covered it early in Today's podcast. A 10% higher return on investment.
- [23:47]
B
Yeah, and I think that number is going to go up, Ray, because a lot of these C A I O people are new, they're getting into their jobs that the, the projects that the 12th project they do based on experience can be better than the third project they do.
- [24:02]
A
Yeah. And one thing that the report didn't cover, I mean, I hate referring back to this report because it's kind of been debunked on the rigor of the actual research methodology. But it's the MIT Nanda report that 95% of AI projects proof of concepts aren't really converting into production and delivering roi. Hey, if you don't have a person who's accountable to see how each one of those AI projects experiments, proof of concepts, call it once you will. If they don't have the success criteria already identified and someone who owns that, owns that report back to the CFO and CEO, you're going to have much less value. So I think this has been a great episode and I really appreciate all the hard work and research you did to help put this together. Peter.
- [24:52]
B
Well, and thank you so. And thanks for tuning in. To all of you. To AI, to roi. If you haven't done it already, subscribe to the newsletter. Is it AI, the number two roi.substack.com that's ai2roi.substack.com Leave us a review wherever you get your podcasts and we will see you next week.
- [25:12]
A
See you next week for the next big story.