Delivering on the AI Promise: Economic Realities, ROI, and the Path to Customer Value

The AI revolution is here, but how do businesses make it work for them? 🤔

‍

At Six Five Media’s Summit: AI Unleashed, host Daniel Newman is joined by Rob Thomas, SVP Software & Chief Commercial Officer at IBM, to explore the practical strategies and cutting-edge technologies that are empowering enterprises to harness AI's transformative potential. The insights shared point to a significant shift in how businesses are leveraging AI to drive future operations and growth.

‍

Key takeaways include:

‍

🔹AI for Productivity: The conversation emphasizes the shift towards using AI to drive tangible productivity gains for businesses, streamlining operations, and optimizing workflows.

‍

🔹The Pragmatic Approach to AI: IBM's strategy focuses on delivering practical AI solutions that integrate with existing enterprise systems, enabling businesses to leverage AI's power without disrupting their core operations.

‍

🔹Agents as Workflow Orchestrators: The discussion highlights the role of AI agents in automating complex workflows and facilitating seamless human-machine collaboration within the enterprise.

‍

🔹Empowering Enterprises with AI: IBM is committed to providing enterprises with the tools and platforms they need to effectively deploy and scale AI, enabling them to enhance customer experiences and drive business growth.

‍

Learn more at IBM.

‍

Watch the full video at Six Five Media, and be sure to subscribe to our YouTube channel, so you never miss an episode.

‍

Or listen to the audio here:

Daniel Newman: Rob, how you doing? Great to be back with you.

‍

Rob Thomas: Daniel, great to be with you. Good to see you.

‍

Daniel Newman: We were sitting on a rooftop in Davos having a conversation, and one of the things that you brought up was your new book. I believe now you've made a lot of progress. It's out, yeah?

‍

Rob Thomas: It's out. I remember we were much colder in Davos than I am right now.

‍

Daniel Newman: Yeah, we were wearing winter coats, sitting outside. But it's sort of the vibe, right?

‍

Rob Thomas: Exactly. So yes, AI Value Creators is out and maybe to paint a picture of the book, there's a story we tell right at the beginning of the book and it's about the Statue of Liberty. Everybody knows the statue. If you look up close at her hair, the detail on her hair is incredible. Like these perfect braids. Everything is exactly precise. But here's the interesting point. The Statue of Liberty was built by Bartholdi in 1870. It was another 35 years before the first airplane. So why did he spend all this time, attention to detail, on hair that was never going to be seen? I think it's the instinct of an artist, a scientist, to do great work. And we use that to set the tone in the book, which is AI is not just about, "Let's try a thousand things and hope something works." It requires a little more precision. You need to think through, "How am I going to attack this problem? How do I make data ready for AI? How do I think about use cases?" We want people to be a little more precise. And so AI Value Creators is really a handbook for anybody that wants to get value out of AI. And we think we have some good stories in there, so a good lesson learned, so we hope people enjoy it.

‍

Daniel Newman: So I have to ask. That particular story, that anecdote, what, unsurfaced that for you as you were probably looking for that kind of story, what made that one resonate so much?

‍

Rob Thomas: I read a lot and I always keep notes of different stories, anecdotes, analogies. So I literally was just going through a book of... It's a number of pages, the old notebooks, and it just came out and I was like, "That's it. That hits the mark for what I'm trying to convey," because at the start of the book, you have to give people a reason to even want to turn to the third page.

‍

Daniel Newman: Oh, yeah. No, it's hard.

‍

Rob Thomas: As you know. And so I was like, "Maybe this incites people to think this is a little bit different," 'cause I think most people expect... They hear AI. There's two camps. One is, "Everything's amazing, everything works." And the other camp is, "You must try 10,000 things." And I don't really believe in absolutes. I think the right answer is, "Let's be thoughtful, let's be intentional." And that's what I wanted to convey.

‍

Daniel Newman: Yeah, I'm incredibly optimistic, but I think there is a difference, Rob, between being incredibly optimistic about the potential in society. You know like the milk, and Paul Tudor Jones went on and talked about... He's a big investor for everyone that doesn't know it, and he talked about the existential risks of AI. And I think there are the doomsayers out there. It's the same people that are permabears in the market that believe everything's going to fail. But I think deep down, most of us have come to the conclusion early and often that this is going to accelerate society, that it's an augmentation, it's a value-add, it's going to make businesses more efficient, it's going to make individuals more knowledgeable. It's going to remove some of the chaotic or wasteful time that we spend trying to find things, make it more accessible to us. I mean, there's a lot of reasons to be overly optimistic despite the fact that I think it sounds like you're coming at it through your book with a bit of realism that needs to be balanced in terms of any company or individual looking at AI.

‍

Rob Thomas: I think a lot of people love to talk about the edges or the extremes. To some extent it creates a level of attention. But I think there's two big ones out there right now that I probably completely disagree with. One is the AI is bad, this is a doomsday. I'm not a believer in that. Two is, there will never be software developers in a year. All software developers are going away. I think both of those are completely incorrect. I think there's a nuance in all of this. Will the role of software developers change and evolve? Absolutely. But I am willing to bet we are going to have more software developers in 10 years than we do right now. What they do day to day may change or evolve. And I believe AI, like all technology has always moved the human condition and society and humanity forward. It always has. And yes, there can be bad uses, there always has been, but I don't always believe in the extremes.

‍

Daniel Newman: And by the way, for decades of machine learning, there's been uses. We talked a little offline. Machine learning is a version of what we really consider in the bucket of AI. People have been using algorithms and patterns for four or five decades now to try to modernize, compute, and by the way, for both black hat and white hat purposes. So I think there's a lot of just that minimalist maximalist view. Even that whole, "All jobs will be replaced and everybody's going to go on UBI."

‍

Rob Thomas: Right.

‍

Daniel Newman: It's like, well, if we don't evolve, that would be the truth. I still remember that I'm looking out at the streets here of Boston and there's sidewalks and lamps out there. And by the way, they used to have to go every night and light the gas lights.

‍

Rob Thomas: Yeah, I wasn't around then. Maybe you were, Daniel. That was before me.

‍

Daniel Newman: I'm bald. I'm not that old. But there was literally gas lights. When they came up with electricity, everybody thought, "Well, that person will never work again." That person used to run around, get on a ladder and like, "Yo, you know, we make progress." And there's so much importance that we talk about making progress. And one of the things I think your company, by the way, beyond just the book in value creation, has made so much progress is sort of thinking about open, thinking about hybrid and really staying the course. I had the chance to sit down with Arvind. You and I have sat down a few times this year and you've not deviated at all, really since Arvind took the helm of this company, on hybrid cloud and AI. And by the way, you were on AI before it was really popular. It was a thing, but it wasn't so popular. I mean, talk about why these two things are so inextricably linked together, they have to be brought together and you have to bring them forward together.

‍

Rob Thomas: I mean, all the credit goes to Arvind. I think he had the insight of... IBM is a big company, it's a very important company in the world. You can't change strategy every year. So he was looking for what are the trends, the themes that can drive decades of investment. Hyper Cloud and AI was the answer. I want to come back to your point on AI for a minute and then we'll get to Hyper Cloud.

‍

Daniel Newman: Yeah. Absolutely.

‍

Rob Thomas: John McCarthy, famous computer scientist. It was 1959 that he said, "Once it works-"

‍

Daniel Newman: I wasn't around.

‍

Rob Thomas: He said, "Once it works, it's no longer called AI." And think deeply about what he was saying. He was saying, "The goalposts are always going to move. The minute that we accomplish something, we're going to say, "That's no longer AI," and the minute we can do something, "That's no longer AI." And I think when I hear these comments of, "We're not there yet," or, "AGI is coming," I think we forget this basic. Anytime we do something, it's no longer AI. When you watch the movie, Her... Have you seen that movie?

‍

Daniel Newman: Oh, yeah.

‍

Rob Thomas: I think that was 2013. If you were watching Her in 2013, you're like, "If this is actually possible that I can chat with something and it's responding back and it's AI, at that point, we'll be an AGI." And we can do that today and nobody thinks we're in AGI, so the goalposts are always going to move. What is our role as a business? Our role as a business is to help clients get value out of AI, and to some extent, which technique they use. It doesn't really matter. It's what is the lowest cost? What is the highest performance? How do you get to the outcome? That's machine learning. Great. It could be a Jupyter Notebook that's making predictions. It could be generative AI, and in many cases it will be. It's definitely going to be agents as those start to evolve and we go from there. So I think it's about, I'd say, the pragmatic approach to AI. Hybrid cloud, I would say we're just getting started. Believe it or not, we're still in the early innings. I think now every company realizes their strategy is hybrid. They're never going to all be on one public cloud. That doesn't work, especially not when you think about sovereignty outside the US. So we've hit the tipping point, but we're still very early.

‍

Daniel Newman: Yeah, it feels so opportunistic for IBM. Every time I hear about why would a company pick to partner with something like AI or especially agentic orchestration, and we'll come back to it, but I can't really wrap my head around why you'd want to rate limit yourself either through an application layer where you're going to centralize all your agentic work through a single application or start there, 'cause I actually believe there's really meaningful change coming to the entire software industrial complex. I also want to understand the comment you just made about cloud. There's many reasons you function in many clouds and some have been more multi-friendly than others, but I still think it's like they're all sort designed to keep you on that cloud no matter what. And by the way, that's good business. I am not criticizing that. But you kind of just said it, and by the way, Arvind said something a little different. So Arvind on the stage said, "We're sort of past the POC." He said that today in his keynote. He came up with a, "We're past it, we're going commercial, we're going big." But we really haven't necessarily seen that scale yet. What, in your mind, is preventing some of the clients... You've worked very closely with the clients. What's preventing clients from getting their AI POCs really up to scale?

‍

Rob Thomas: It is the classic problem of data, I would say. It always comes down to either data or skills that would slow you down. Data unlocks AI, and so once you're able to identify use cases, there tends to be some level of data preparation that's required to make sure you're going to get the accuracy, the performance that you want, so that could slow you down. Second would be skills. Do you have the skills to do this? I think that's why IBM is quite well positioned. We have a consulting business that can provide all the skills that you need to work with any AI model, any AI stack. And then we have the technology with open source now building applications on top of our open source models like assistants and agents. So that's what the world needs right now.

‍

Daniel Newman: So let's flip to the macro a little bit. It's been a really wild year. We did this massive survey coming into the year, Rob. We talked a little bit about a Davos with 211 CEOs, billion dollar plus companies, and AI was by far and away the biggest board priority. Now, April, we had Liberation Day. We're not going to debate politics here. But now the number one for many companies has become supply chains again. I mean, it's a chain. But largely what I'm hearing is that AI is tariff proof, especially, and you bifurcate consumer and enterprise. On the enterprise side, on the CapEx build outside. But you're talking a little bit more to the client's deploying. So obviously we saw numbers. Meta's going to keep spending and Microsoft's going to keep spending. They're going to build the data centers of the future. What about the enterprises that you're talking to? Has there been any change, any halts, any slowdowns, any impact, or are they seeing this deflationary and seeing it as a go, go, go, 'cause they need to get there?

‍

Rob Thomas: Maybe two times scale. So go back to January and Davos that you mentioned.

‍

Daniel Newman: Yeah.

‍

Rob Thomas: There was a lot of optimism at that moment, from every CEO that I spoke to. And that was really the view that we think we're entering an environment with less regulation. That's actually true. We shouldn't forget that. Less regulation is generally good for business. So I think that optimism is still there. Now, yes, I think today there's a little more uncertainty and people are trying to decide what to do. I think that's actually a catalyst for AI. So the biggest change in discussions in the last 45 days has been, "We're going to keep investing in technology. We have to do AI, but we're really only interested in use cases that drive productivity." I'm not sure that's a bad thing for businesses. It's kind of focused on the basics. How do you get more productive with what you have? And I think that will be the catalyst for this year and perhaps even into next year. It's going to be all about productivity. And again, I think for AI to work, the more pragmatic and value accretive, the more likely companies will stick with it. So I think this is actually a good thing.

‍

Daniel Newman: Yeah, I think at any time, I know during '22 when we had QT and the market really fell, I saw companies get sharper. That was actually what I called the AI boom. I remember I went on Squawk Box and I actually said... NVIDIA, in July of '22, it was down like 70%. I'm like, "AI." I'm like, "Watch NVIDIA." I'm like, "Because the deflationary nature of it..." And that's what you're saying, is companies get smart. When there's excess, when the growth is coming without a lot of effort, you over hire, you over invest, you do more events than you need to do. You just spend money and really good prudent businesses get back to basics when the macro gets a little more complicated. But now we've got a new potential variable. We've got agents. So agents now basically can put a company in the driver's seat to say, "Look, we're going to augment, assist, displace, replace," I'll let you fill in the blank there, "a number of different roles, and we're going to get a ton of scale in our business. We're going to get tasks done more quickly, more efficiently." How does agents change the game and how does IBM think about this because you seem to be in a really good position with Orchestrate and what's coming next to solve problems for your enterprise clients?

‍

Rob Thomas: So let's talk about how this evolved. We started Orchestrate back in 2021, and I think-

‍

Daniel Newman: I was around then.

‍

Rob Thomas: ... this is probably the biggest thing that we've learned in how do you innovate as a big company? You have to be willing to iterate. At that point, we were really just focused on digital labor, how do we automate tasks? So we went forward a few steps back, a few steps. We're now four years later. Two years ago, we brought out Orchestrate, we made it generally available, and we actually got a lot of customer traction with automation. But then we retrenched a little bit and we said, "There's something happening with agents." And I will tell you the value in Orchestrate. It's not necessarily the agents themselves, it's literally what I'd call the middleware of agents. How do you get agents to work together? How do you deploy with multiple models where you can use Granite for one query, Meta or Llama for another query, Mistral for another query. So that whole orchestration piece, hence the name, is where the value is, but we announced an agent marketplace. We've recruited nearly 30 ISVs. We have some of the biggest in the world, Salesforce, Adobe, Workday, and we got startups like 11x, Simplistic AI, who are building on Orchestrate, and we help create market awareness for them and demand generation. I think we've hit a sweet spot here because clients are going to come to IBM for every agent. So we want to say, "We can give you the middleware and then you can use IBM agents, or you can use an agent for one of our partners and all of that will work together."

‍

Daniel Newman: Yeah, you're not really trying to displace the partners. You're not telling the application layer not to build what they're building. You're saying, "We're going to make it work better and we're going to make it work across your stack." Because that's my biggest problem, is it can't just exist at the SaaS layer. It can't just exist at the infrastructure layer. You need these things to really coexist, work together, and of course have these handoffs and exchanges and solve a lot of problems that historic automation and IPA and RPA just didn't solve, and I think you've shown some really good results over the years with digital labor, with automation. But I do think the machine to machine handoff, the human in and out of the loop selectively, I mean, it's exponential, right? I mean, the difference now is you've done something where you did one repetitive task and then maybe added one more, and now it's like you can go really fast.

‍

Rob Thomas: Now, I do think we're headed probably towards some level of disillusionment on agents. And here's the reason. I think there's a view of, "Hey, agents are amazing. They will displace all SaaS, all software." I'm actually incredibly skeptical of that. If you think about an agent running in a company, I think we are a long way away from, "I'm going to tell you an outcome, and the agent will decide all 50 tasks and perform them autonomously." I'm very skeptical of that. That could work in some B2C use cases. I don't think enterprises. They're all too different to say that's going to happen overnight. What, I think, can be done is you can start with agents now, you can execute some tasks, there's likely going to be a human in the loop to get started. So again, I would recommend, let's be pragmatic here. Let's not think that this is going to displace everything that you're doing.

‍

Daniel Newman: Well, I think the only question I would ask is how fast it goes from where I agree we are right now, and I think you're right. We cannot just hand over the keys. I've seen some impressive demos, but again, it tends to be inside of a sandbox in a controlled environment. You can't necessarily let this thing run free and start making major decisions on your maintenance of your airplanes or major operations in hospitals. But over time, I think the goal is that these things can become incredibly intelligent and be able to do a lot of things and give great efficiency to your businesses. And so you started own the path here, and maybe we'll finish here, Rob, is all the wisdom you've gained from your book, the wisdom that you've collected working with so many customers leading now software and the commercials of this company, what's the most prudent advice that you're giving to these companies now? Because by the way, we're inundated in this stuff, meaning we wake up and all this stuff is normal to us, but if you're running an industrial company, manufacturing, if you're in transportation, logistics, oil and energy, it's there. But this speed has to be absolutely overwhelming.

‍

Rob Thomas: It is. Look, if I sit down with a CEO, I go through the same question in my mind. One is, do you believe technology is key to your competitive advantage? Obviously, everybody says yes to that, but you can tell from the body language, do they truly believe that? So I think one is, that is a fundamental belief you have to have if you're leading a company. Two is can you iterate? Are you willing to put things in production, iterate off of that? That is a big culture change for most companies, including ourselves, which I think we've gone through in the last few years. And then three is can you then start to hone in on value creation and how does this augment my current workforce? Bigger companies do that. This is going to be a home run. It's going to make you a more productive company, better working with customers, improving your supply chain, but you've got to start from those areas.

‍

Daniel Newman: Absolutely. Well, we're two believers here. The pace that'll be determined in the future, we'll see next year when we sit back down, probably together and have the conversation just like this, just how far we've come. Rob Thomas, IBM, thank you so much for joining me.

‍

Rob Thomas: Daniel, good to be with you.

‍

Disclaimer: The Six Five Summit is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

‍