Share this
As OneStream’s AI expert, Dylan Meroz talks to finance leaders about AI every week. He has concrete answers, grounded in real client conversations. At Future Foundry 2026, he’s a panelist and breakout session speaker. We spoke with him ahead of the event about making AI tangible, what that means for your team, and what comes next.
1. What’s your vision on AI in finance?
AI in finance isn’t optional anymore. You adopt it, or you get left behind by competitors who already have. The question I keep coming back to is: how does it impact the lives, and the future lives, of finance professionals? Picture doing all your financial work the way you already work with ChatGPT or Claude. You’d use that same direct, natural interaction, except it runs on your own financial data, secured and governed underneath. Those are the conversations worth having.
2. Are there any other areas where AI can have a significant role, for example, in ESG reporting?
AI can add real value to ESG. From our perspective, it has a role as an intelligence layer. Performance Management platforms like OneStream provide the governed data and the underlying calculations, and AI makes that information easier to interrogate, analyze, and explain. Say a user asks why emissions exceed a target, and which entities are driving that: AI builds that management summary fast, but it doesn’t touch the underlying controls or the human sign-off ESG reporting still needs. I don’t see it that differently from AI in finance, or other areas like operations.
3. With AI moving into so many parts of organizations, how do you figure out where to start?
What we recommend, and this isn’t specific to finance, is to start with one clear pain point, not a company-wide AI rollout. For every person excited to fix everything with AI, there are three who aren’t exactly eager to change how they’ve worked for the past fifteen years, so start small. Every company reports, and every company knows which parts of that reporting are painfully manual. Find the report that takes the most work today, and fix that first. That’s also where we see the most successful projects: a small team, a clear goal, quick results, and other teams asking for the same once they see what it delivers.
4. Where do you draw the line between AI and human judgment?
The line is simple: AI can analyze, draft, and spot patterns. It doesn’t decide. We always tell clients starting this journey to write a policy on responsible AI use: where they use it, and where they don’t. As a person, that means using AI as an assistant, faster analysis, pattern recognition, a first draft, while you stay responsible for the judgment and the final call. As an organization, it means reliable data, clear access rights, and transparency about outcomes and oversight. Our agentic solutions include a control tower, so administrators can see exactly what users are doing with our AI. Finance professionals need to trace where numbers come from, or they won’t use them.
5. What are the biggest misconceptions you run into?
Plenty, and I hear most of them daily. “We don’t have the budget or the people” is a classic, though these projects don’t have to be invasive: they can run alongside other work. “Why go with specialized AI when we already have ChatGPT” is another. People think they can just connect their data to a general model and it’ll work. In practice, however, it’s not that simple. Then there’s security: “What data does AI use, and is sensitive financial information protected?” With OneStream, that one’s simple; the security rules already in place for our platform carry straight through. There are more, but I’ll save the rest for the panel discussion and the breakout session during Future Foundry.
6. What’s the next big step in AI?
Today’s AI answers a question when you ask it. Agentic AI goes a step further: agents that carry out tasks and connect to other systems on their own, within the boundaries you set. The next step beyond that is what we call the agentic mesh, a web of agents that all talk to each other. Every system a company uses eventually launches its own agent, whether that’s CRM, expenses, or planning, and those agents connect to each other, not just to OneStream’s. I believe that’s where this is headed, while humans remain always in control. I don’t buy the idea that we’ll be without jobs or money in fifteen years. I’d compare AI to the internet or the mobile phone: hugely disruptive, and yet day to day, people carry on. What will change is that manual, repetitive work nobody wants to do anyway ends up being done by AI.
7. At the Future Foundry, you’re both on the panel and hosting a breakout session. What can people expect from each?
The panel is plenary: a lively discussion with a mix of perspectives, practitioners like me, users, philosophers, and scientists. The breakout session is where it gets hands-on. You get a short demo where we put agentic AI to work, a chance to touch the technology, and at a look at how our agent performs. Bring your objections too, like the fear that your team gets too dependent on AI. I’d love to answer these live.
Want to hear more from OneStream’s AI expert Dylan Meroz?
Join us at Future Foundry 2026, Swap Support’s annual event for professionals in finance, BI, and ESG.
Share this
Join the Future Foundry
