Senior VP of Software & Chief Commercial Officer, IBM
The View From Davos with IBM's Rob Thomas on Turning AI Spend Into Real Enterprise Value
00:01:06 – IBM's Role in AI and Government 00:02:03 – Extracting Value from AI 00:02:44 – Challenges in AI Adoption 00:04:24 ...
About Robert Thomas
Rob Thomas, Senior Vice President of Software and Chief Commercial Officer at IBM, has been a frequent commentator on enterprise AI adoption and value creation. In early 2026, he stated that the most incorrect idea about AI is fear, arguing that productivity growth—driven by AI—is the only path to continued GDP growth. He noted that S&P 500 operating margins have risen from roughly 13% in 2019 to about 19% in recent quarters, attributing this in part to technology. Thomas has emphasized that 2025 would be a year of "value creation" with AI, moving beyond experimentation. He has also discussed IBM's own use of AI, stating the company generated $4.5 billion in productivity by automating manual tasks and has displaced some jobs while hiring into R&D roles. Thomas has been involved in several major IBM acquisitions, including the $11 billion purchase of Confluent, which he described as "the rails of this AI moment" in a December 2025 interview. He has also discussed the importance of open source in AI, arguing it prevents vendor lock-in and reduces barriers to adoption. Thomas has advised companies to focus on cost savings before revenue when starting with AI, and to embrace iteration over large, long-term projects. He has also spoken about the cultural and change management challenges of AI adoption, noting that a Fortune 50 CIO told him that while the technology works, getting people to use it remains difficult.
Transcript (32 segments)
H
Host0:07
The 65 is on the road with a view from Davos. Daniel, it's been a great show so far. The big political fireworks haven't hit yet, but I really did appreciate coming to Davos for a second year to get a much better view of what's going on at the intersection between AI, policy, technology, and a whole lot more.
D
Daniel0:30
Well, we're seeing the world change right before our eyes. This is an event where you bring together the business, economics, and policy leaders, and you're able to have those conversations in such a small amount of time. I don't think there's been a bigger topic, regardless of industry, than the impact of AI. I expect this week to not only talk about regulation, energy and AI, and the escalating tariff and trade policy issues, but also about jobs. What does it mean when you have software that writes itself and completely rewrites industry? There's a lot going on this week.
H
Host1:06
Yeah, totally. One of the big tech companies making major moves not only in enterprise AI but also in government AI based on its Watson X platform is IBM. It's great to welcome Rob Thomas back to the show. Rob, how are you?
R
Robert Thomas1:23
Wonderful, great to be with you both again. Another year in Davos, exciting.
H
Host1:26
Yeah, what has Davos brought you so far? I know it's early, but what do you want to achieve here?
R
Robert Thomas1:33
We're just getting started for the week, but I would say the feeling is different this year. From what I can tell, CEOs around the world have shown up aggressively and significantly. If I look at the meetings I have, heads of state as well. So I feel like Davos is really on the map this year for many of the reasons that Daniel just alluded to. It's the convergence of business, technology, and government and everything that's happening in the world.
H
Host2:03
Makes sense. This one's going to go really quickly. One of the things you've been focused on, running many parts of IBM, is extracting value from AI. We've had these years where there are five or ten companies that AI is their business and they're really making money, and then there are a lot of companies trying to figure out where AI drives productivity and efficiency to their P&L. What is, in your mind, working with so many enterprise customers, the gap between companies really getting to value and P&L growth because of AI and where they're at today?
R
Robert Thomas2:44
So maybe back up a little bit to the macro setting. S&P 500 in 2019 operating margins were 13%. Through 2025, for the most recent quarters reported, which is really calendar Q3, it's 19%. That's a pretty significant difference, and maybe another two points to come in 2026. So technology is having a big impact. To your question, I'm actually not sure that's even AI driven at this point. I think that's technology driven, which is people are bringing technology into their company, figuring out a way to automate processes, figuring out a way to become more productive and more efficient. So let's not act like there's not been a ton of progress. There has. You can see it's very deflationary. Absolutely. But for AI, that's a different question. And as I tell most clients I meet with, you are not going to get an ROI by having AI draft your emails and then when somebody receives it, having it be summarized by AI. And I think 95% today check the box on that basis. That's not going to be what this is about.
H
Host3:53
Yeah, it makes sense. I admit I might do summaries, but I don't have to write my emails. I have not crossed that bridge. I still think the human element matters, but I digress. When it's not delivering inside of an enterprise, are you seeing it primarily as an execution issue, a technology issue, or a people issue, or something different?
R
Robert Thomas4:24
First and foremost, it's culture more than imagination. Meaning, do we have the mentality to be aggressive, to try things, and to disrupt ourselves? Because done right, AI is going to disrupt how you operate as a business. It will change your processes. It will eliminate processes. So, do you have the courage to take that on? That's why I start with culture because it's going to be a massive change management exercise. A lot of people ask me about this and they say, 'Well, shouldn't our teams be doing this?' I think this has to be tops down in the short term to give people the confidence that we're going to try this and we know it may not be perfect. I hear a lot of people make the excuse of, 'Well, AI got something wrong.' I'm like, 'Look, the humans get a lot of stuff wrong.' I don't know about you guys, but my experience is humans get a lot of stuff wrong. So the fact that AI makes a mistake is not a valid excuse. I think it has to be tops down taking on the cultural and the change aspects and being willing to change completely how you operate in some cases.
D
Daniel5:33
You get a lot of that where people talk about the cars like, 'Oh, the Tesla had an accident.' It's like it had an accident. I'm like there was like 50 times more by humans in that given day. That kind of analogy or even worse are some of the people like, 'Oh, I used ChatGPT and it hallucinated something one time so you never used it again.' I mean, look at how fast and how exponential this is changing. They get things wrong and then a week later they ship new updates and it's better. So if you're a business or an employee of a business and you're being tasked with this, I want to drill into this culture one really quickly. How do you get, because that's a big part of your job, the IBM consulting side, getting people to change their mind? How are you getting them to change their mind and really drive ROI and drive adoption, drive behaviors that make this work in their business?
R
Robert Thomas6:24
I walk in with a point of view and I say actually AI is best used for boring tasks because you want to kind of catch people's attention. People are like, 'What does that mean?' I'm like, let's not go for we're going to put people on the moon or we're going to fundamentally create a new business. Let's go after the boring repetitive tasks in the company and let's say can AI make you more productive? Can it reduce the amount of repetitive tasks that you have to do? Can it increase your cycle times? Normally that starts with what I would call the back office of an enterprise. It's customer support. It's HR, supply chain, procurement, finance, operations. I think if people are willing to start there and the experience that we've had, you'll find an ROI actually quicker than you think, at least to build some confidence and get going. But if you start with we're going to use AI to double our revenue, I would say maybe there's a lot of different factors that go into that. So control what you can control which is the operations of the company.
H
Host7:26
Makes me laugh. You guys seen on X those people that'll post something like, 'Claude, grow me a billion dollar business, make sure that it makes 90% margins and make no mistakes.' That's the new Claude prompt that everybody makes. How quickly people are thinking that the technology is anything, 'I want 10 times my revenue.'
R
Robert Thomas7:44
And the ideas are good, like create a PDF-based business. That's great, but there needs to be demand for it. Yes. Which is why I say I would not start necessarily with the revenue piece. Start with the cost piece. And kind of back to what I shared at the top, if operating margins have improved like that without AI, that's actually a proof point that this is very doable because technology itself has always been able to do this. Now we can take a next step.
H
Host8:06
So Rob, I love boring is beautiful and backend, probably more front end in the future, but maybe not because we're dealing with customers. I've heard that use cases have been hard to define. I'm curious related to use case definition and metrics. What are some of the best practices that you're seeing with your clients that are delivering the goods?
R
Robert Thomas8:34
I think so in terms of KPIs, you've got to look at two things. One is can you reduce the cost of an existing process. That's pretty much all that matters. Secondly, and I would say almost equally important, is can you increase cycle times? Because if you can reduce the time that it takes to deliver something to one of your constituents, supplier, partner, customer, and you can take it down by 100x, that's value. That creates velocity in the business. So an efficiency play, then I would always start with one of those two because if you build enough confidence on we can use AI to increase productivity, then that gives you the option to go do anything else. So think about what are the things as you look at your company that take entirely too long to do in the year 2026. I think we all have examples of that.
H
Host9:28
You think there's a new role coming? I know there was a period a few years back in the early moment of ChatGPT where there was this chief AI officer role. It feels like it died down a little bit. Some places have it still, but do you think there's that role, that person that kind of used to be an almost COO type that goes now and looks at every process inside of a business and starts to figure out where AI can speed it up? Because I don't think I'm seeing that but it feels like an obvious opportunity. I'm doing that in my own business, but in big companies.
R
Robert Thomas9:56
You're the chief AI officer.
H
Host9:57
I'm the chief everything officer. That's the E by the way. The E is not executive. It's everything. But you work with a lot of these companies that have to look at so many antiquated processes, so many things where I call it people that just move data from one place to another all day long. This is like the perfect thing for AI to do.
R
Robert Thomas10:12
So, I'm probably mixed on whether or not there's a new role. It probably comes back to my cultural point. I think for some organizations that is necessary because that's how the company has always functioned. I think for others, the minute you do that, you take it off everybody else's plate. You say, 'You guys don't have to worry about it because we've created this new AI officer.' So, I think this is very company specific as opposed to there's a general correct answer.
H
Host10:38
Oh man, I was hoping you'd just say yes.
D
Daniel10:40
Yeah. And there's also another conversation going on and this intersects the cultural thing you talked about and that was, well wait a second, is the CIO, if the agents are being created by the CIO, do these agents report to the CIO? And if the agents are doing all this meaningful work, so we're getting into these, okay we have built these things that work now, who manages these and how do they spread across? So I think that'll probably play its way out in the next 3 to 5 years. So IBM, giant company, a lot of history, a lot of technology, you've made some changes with AI. What advice are you giving to other very large companies on where to start and has that evolved over the last two and a half years?
R
Robert Thomas11:35
One thing I would say is maybe the undertalked part of AI is what I would call applied AI. How are you using AI to improve something that's core to your operations? I'll give you an example. We've invested a lot of R&D on how do we use AI to make it run better, to be able to manage the landscape of a network, to automatically patch vulnerabilities. Most people think of AI as this chat interface kind of doing some of the basic tasks as opposed to this can change how we operate our technology infrastructure, how we become more resilient. If I think about one theme that I see, projecting forward for 2026 and probably beyond, I think it's the return of resilience. This is something that nobody talks about. Everybody knows they need it when they don't have it. But given everything that's happening in the world, I think we're going to see a big return to a focus on resiliency and I think AI can play a big role in that.
H
Host12:36
Makes sense. So, we're kind of winding down here. Appreciate all the time, Rob, but I guess what do you think people are getting wrong? While we're here in Davos, what's the big thing about AI that, I mean, there's the bubble talk, there's the it's going to replace everyone's jobs talk, there's the it's going to make the economy 20 times larger. There's a lot of big kind of hyperbolic comments and statements being made, but you're so close to it with so many. What do you think's not being correctly articulated out there in the media or in the market?
R
Robert Thomas13:09
I think the most incorrect thing is fear because there's an old equation in macroeconomics that says GDP growth comes from population growth plus debt growth plus productivity growth. You look around the world, we're not going to have population growth. Debt growth, I think, is probably uncertain given the way the world's changing. So the only way for the world to continue to grow will be productivity. AI can be the single biggest driver of productivity. So why are we scared of that? Technology has always had a positive impact in the long run. That doesn't mean there's not disruption in the short term. So I think anybody that's fearful of this, I think that's the biggest mistake I see at the moment.
H
Host13:53
Well, Rob, I want to thank you so much for spending some time with us here. It was great to talk. We're going to have to come back. We do this a couple times a year and it feels like this changes so quick. I laughingly said like a year ago I made some comments. I think you did too, but maybe not. I'm like I think the developer as we know it is over, and people thought I was crazy and a lot of the world's best developers said, 'Oh, no. It's never, AI will work with us, but it's never.' And now you got like Karpathy and Boris from Anthropic. They're like, 'Yeah, it's over. It's over.' I mean engineering, yes, workflows, yes, but coding, what I'm saying is like, so in six months, we could get together and like things that we said right now might be harder on it could be done.
R
Robert Thomas14:33
I'm still willing to debate that one. By the way, what we've seen as we've rolled out watsonx at IBM, we now have 20,000 users. It actually makes the best developers even better. Yeah. And it raised the bar a little bit, but I do think it really gives people superpowers. And I don't really think software development is going away.
H
Host14:50
Well, in fairness, by the way, these developers aren't saying they're just basically saying now they have superpowers, but they are basically thinking more as engineers and more as workflow experts and business people, not code writers, right?
R
Robert Thomas15:02
So, it evolves how you spend your time day to day, hour to hour, but that's also true if you look at the previous decade. Before there was code repositories or modern ones, there was a lot of manual effort in managing code and that kind of thing. So I think the role probably evolved.
H
Host15:18
Rob, so much fun. Always great to talk to you. We'll have you back again real soon and have a great Davos.
R
Robert Thomas15:21
Good to see you guys. Thank you.
H
Host15:23
Thank you everybody for spending a little time with us here with a view from Davos. This is the 65. We are on the road. Subscribe. Be part of our community. Check out all our great content here in Davos and everywhere else. We got to go. See you all later.