Driving Operational Impact in BFSI with Agentic AI - with Yoav Naveh of Reindeer AI

**Daniel Faggella** (0:12)
Welcome everyone to the Emerge AI in Business Podcast. Today's guest is Yoav Naveh, co-founder and co-CEO at Reindeer AI. Reindeer AI builds agentic AI systems that automate high-volume labor-intensive digital processes for enterprises, including banks and financial institutions. Today's conversation moves past theoretical use cases and into what's really working. Agentic AI running live inside document-heavy regulated processes like AML validation and account closures with human oversight built in from day one. Today's episode is sponsored by Reindeer AI. In this episode, we cover how banks are digitizing document-heavy compliance workflows like AML validation and using public records and alternative data to strengthen both risk decisions and customer attention. To go deeper on this topic and learn how financial institutions are digitizing paper-based records to unlock usable data for AI and using alternative data like public web and social signals to enhance risk assessment, download our free PDF report AI in Financial Services Executive Cheat Sheet at emerge.com/fcs1.
That's emerj.com/fcs1 to download your copy. Now the conversation with Yoav.
Yoav, it's great to have you back on the show.

**Yoav Naveh** (1:47)
Great to be here again.

**Daniel Faggella** (1:48)
Yeah, you left us with a lot of thinking the last time, especially with your closing statement. And I'm excited that our episode today is going to be more about what does it look like when it's working in terms of agentic AI and financial services. And I'm going to hit the ground running with question one already, because the conversation about agentic AI and financial services has been going on long enough that we should be able to move past that, what could it do and more into what is it doing. And the use cases that are genuinely working are the ones that we should be focusing on. The ones where teams are expanding rather than quietly just shelving what they have. And it feels like there's a lot that we need to unpack there, whether it's the shape of the work, the kind of data involved, the compliance context. So which workflows are delivering and what makes the ones worth paying attention to different from the ones that we shouldn't be paying attention to.

**Yoav Naveh** (2:42)
I think there is the place we all want to get to. I think that sometimes that conversation about use cases could feel tactical for a CFO, for example. Okay, I'm going to get some better efficiency on my account payable process, or I'm going to get my AML alerts being handled with higher accuracy. I think that the end game where we all want to get to is the ability to have data available, more readily available, be able to make smarter and better decisions on data that is actually accurate and as close as possible to real time in life and available. I think the path to get there goes through a lot of different use cases and workflows that you have to implement, and all those pieces of getting everything in your operations streamlined. One of the biggest challenges when we speak with enterprises is, how do we get, we all want to talk about that big vision, but how do we actually choose where to start? Some of them are thinking, let's take a use case we have and reimagine it. AI can do, AI shouldn't replace the manual process we have today. Let's already start reimagining. And I think sometimes you have forward thinking leaders that are able to kind of have this amazing vision of how they want use cases running differently.
I actually think that sometimes it's a scary thought that delays the process. And I think that a good first step to taking use cases that you have, pretty good grasp of what you like and what you don't like about them. And at the first stage, shifting them to AI, I think about it as even as digitizing them through AI. And then once you do that, it's so much easier to innovate on them and improve them. And I mentioned AML before, I think that's a really good example. We had a process where there is an alert that was coming in, and customer is required to upload various different documents to show that the deposit they made is a valid one, that the source of fund is valid.
And you can think, oh, I want to do this process entirely different. It's actually started by just doing the same process people were doing, by being able to read the documents, extract them, validate them against the policy.
Once we've done that, it was so easy to say, hey, you know what? How about we start doing public record search? And maybe we don't even have to require a lot of documents from someone, because we're able to find enough information online about their income, about what they do for a living. Doing that with people, having them kind of play around with all these changes is very difficult. First of all, it takes them a lot of time. It may only be relevant for 2% of the cases, but doing that with agents is a much easier way. But you have to start with the baseline. So I think often our approach is take a use case, do the initial version very close to what people are doing today, and then start reimagining after that.