Data Faces Podcast — On Location · Gartner D&A Summit 2026
Caleb Watkins, Solutions Engineer at Solidatus, on bring-your-own-LLM access to lineage work, trust, governance, and the tasks an AI assistant can perform.
Listen: YouTube · Spotify · Apple Podcasts · Amazon Music
About Caleb Watkins
Caleb Watkins is a Solutions Engineer at Solidatus, where he helps organizations understand how data lineage, metadata, security, and governance support practical AI adoption.
In this interview
- How Solidatus’s lineage assistant works with a customer’s preferred LLM
- Why access, security, trust, and governance carry through to AI-assisted lineage work
- What users can do through the assistant without being Solidatus experts
→ Read the companion article: What if your best AI governance asset already exists?
→ Browse all on-location interviews: Data Faces Podcast — On Location
Full transcript
David Sweenor 0:00 Hello. Welcome to the Gartner, Orlando 2026 data and analytics conference. I’m here with Caleb Watkins. He is a solutions engineer for solidatus. Caleb, how’s the show going?
Caleb Watkins 0:10 It’s going great. So far, having lots of good conversations, hearing more about the industry and what everyone’s doing with AI, including the new solidatus AI agent, yeah.
David Sweenor 0:22 I think we have a scoop. Tell us about the lineage assistant. What does it do? Yeah.
Caleb Watkins 0:26 So our lineage assistant, so first of all, it’s a bring your own LLM model, and then we provided the context of how to work within solidatus. And with our AI agent, you can perform any tasks that you normally could do in solid data is directly through the AI without having to be an expert in solid data to use it.
David Sweenor 0:44 Okay, you know, I bet one of the big questions here is it safe? People are worried about trust and safety.
Caleb Watkins 0:49 Yes, yes, absolutely. So first of all, solidatus is built with security and trust and governance principle as far as access of data and metadata in line, and that carries over to the new AI agent
David Sweenor 1:01 as well. So what sorts of things can it do? It’s got, like, couple of use cases, yeah,
Caleb Watkins 1:06 so just couple of use cases off my head. So one is code scanning. So that was historically something that was fairly difficult to do with the traditional lineage tools that exist on the market all you know. So with our new AI agent. We can provide it, for example, Python queries. And we can say, hey, scan this code, provide it some prompt engineering, and it will go through and it will map out the data flows for that Python code. And then we can also tell it to link it up to the rest of our data flows, so we get that true end to end picture with the code
David Sweenor 1:42 in there as well, and to do that manually. How long would that take before the assistant, before manually?
Caleb Watkins 1:47 I mean, probably take you several days to do. And now with the assistant, you can do it in about 510, minutes, right? So it’s really multiple times faster than than doing it with a human by hand.
David Sweenor 2:01 Okay, another thing I heard that came up is the regulatory landscape. Can this help understand, you know, the state of your data in relation to different regulations like BCBS 239, AI act and things like that?
Caleb Watkins 2:12 Yes, absolutely. So that’s another use case that we’ve been building out. So because you have all of your data centralized in one place, all of your metadata. And then you can also provide the regulation, the regulations in a reference model in solidatus, as well as all the other reference data that exists within your organization. We can provide solidatus Our BCBS, 239 regs. We can train solidatus up on those regulations, and then we can ask the assistant to assess your models for these for compliance with these different regulations, to make sure that you’re heating, you’re meeting all of your objectives with these regulations.
David Sweenor 2:56 And what’s been the feedback on the assistant today? I know you’ve talked to a lot of people. Was hopping earlier, yeah, yeah.
Caleb Watkins 3:01 So it’s something that we’ve been in private preview with, with some of our customers. We’ve obviously shown it to several people today. The reception has been really good. People really like that. They can ask simple questions to the agent, and it can then go through and answer those questions, and not just answer those questions, but the agent can actually perform the work within solidatus itself, to do the mappings, to do the assessments, to turn on display rules within the solidatus platform.
David Sweenor 3:29 Okay, great, and people want to find you at the show. What booth Are we at?
Caleb Watkins 3:32 We are at the solidatus booth in nine to nine over by theater three. All right.
David Sweenor 3:39 Well, thank you. Caleb, sounds amazing. Awesome. Thanks. David. Cheers.