The AI world according to Boomi CEO Steve Lucas

Boomi CEO Steve Lucas riffed on the company’s future bets, where he sees enterprise AI going, why he’s a headache as a vibe coder and why enterprises will always follow the economics over the fancy models.

Lucas’ keynote at Boomi World 2026 was entertaining, but also shed light on the state of enterprise AI as well as the company’s strategy. Here’s a look at the takeaways from Lucas.

• Boomi to buy Lunar.dev, eyes intelligent prompt and model routing
• Boomi builds out data-to-agent platform, says welcome to headless party

Boomi's approach and how it stands out. Lucas noted how Boomi has evolved its stack from iPass to data movement and agent orchestration.

There's also a lot to be said for neutrality. He said:

"It's about the journey, not the destination. At Boomi we're not telling you where to put your information, but we make the journey profoundly better. Boomi is never going to mandate a lake. Boomi is never going to mandate a model. But we are going to make the journey the first class, best journey you could possibly go on with your enterprise. Our mission is to deliver the Boomi Enterprise Platform across the connected, automated and digital enterprise, leveraging all of your data, activating it, driving it to a place where you can monetize it."

Data activation. "Only 7% of your organization's data is AI ready today… and these projects, they're going to be abandoned unless you have your data activated and ready," said Lucas.

Lucas added that data is simply sand. "There's tons of it, not all of it's useful, but we can do amazing things with it. We can do these things with data, but it's only if you put it to work. It's only if you activate that data. Now activating your data. It's not just inside of the traditional data sources that you think about," said Lucas, who noted that data is trapped in the physical environment, devices, operations technology, Internet of things and systems.

You need referees and rules for AI, which has deleted the rules. “The reason we invented referees is so that we play by the rules. And the referees in our modern era are things called ERP, CRM, silly little things like accounting rules," said Lucas.

"We're rethinking everything in this world of acceleration. My thesis on that is because it's just too easy to start. It's like, well, I got an idea. Let's skip the whole planning thing and requirements analysis and send it and in this send it culture that we're building with AI. We're missing the mark on ROI," said Lucas. "We should find ways to harness AI, but AI alone is not the answer, especially in the enterprise. We know that because we have silly little things like people, trust, rules, and speed of adoption. AI will meet reality, and that is in the enterprise."

Agentic capabilities must merge with deterministic referee systems.

Software is proliferating and that's an issue. Lucas said:

"More of you are now vibe coders. Just like me you've convinced yourself that you're a wizard. Every three days, I start doing something with Claude Code, and then I send it to my CIO, and I'm like, I have solved all of your problems. And in care, he's like, not at all. In fact, you just created more. Because this doesn't scale, it's not secure, it's not compliant. Are you going to manage it Steve? He doesn't say that part to me because he still works for me."

Focus on the planning and governance as well as architecture or you'll be doing mop-up later. "I promise you that two years from now, AI will make 50% of those decisions. You'll never even see them," said Lucas.

In other words, governance matters. Lucas said:

"It's not just build, it's also govern. Why jump to govern? Because no matter who is building the agents, somebody's got to watch these things. Someone has to know what data source, what API. Are they duplicates or triplicates? Are they reducing or helping with the amount of work? Are they burning too many tokens that you need to have kill switches? That's partly what Boomi does with governance. What we do with governance is we inventory agents for their skills. This matters because you have to understand what these agents are actually good at."

Enterprises will apply economic reality to AI. Lucas said enterprises will start routing work to AI agents and models that are the most economical. Follow the returns.

He said:

"If you go to Hugging Face and you look at models, there are millions of open weight models and domain specific models. We figured this out. Now the challenge is, well, where do I run the open weight models? Do I run those in my own cloud, on premise? Do I do in the sovereign data center? And the same thing for domain specific models? The point is, all of these things exist together."

In the end, open weight models win, argued Lucas.

"They cost money to run from a compute standpoint, but they are profoundly less expensive. Do we continue to pay every year 10 times what we did last year for a proprietary frontier model? Or do we start to do what enterprises always do, which is pursue the lowest economic option with the highest rate of return economics? ROI always wins. It will profoundly supersede AI at all times."

There's an opportunity to simplify. Lucas said it's still very early with AI and there's a forever opportunity to "fight complexity with simplicity, and at the same time, prepare your enterprise to be the programmable, headless and AI ready."

Enterprise complexity, however, has only become worse. Nevertheless, focus on simplification. Lucas said:

"It's so early with AI and now I have these amazing results, there's a question: Shouldn't things be getting simpler? Shouldn't they? I mean, we have AI. It's a wizard. Let it go, do things. But are they? Do you feel more or less calm, The Forever problem is things aren't getting simpler, they're getting more complicated. I believe there is a forever opportunity, not just a forever problem. That forever opportunity is to fight complexity with simplicity."