The knowledge graph is fast becoming a foundational layer for enterprise AI, as organizations race to turn scattered data into answers that leaders can trust. As IT stacks that went cloud-native now go AI-native, a new class of layered data architecture is taking shape — one built to give models the context they need to reason, not just retrieve.

That shift is playing out across a fragmented data landscape of lakehouses, operational databases and customer profile stores, where the hard problem is accurately describing an enterprise’s information so AI can make sense of it. Graph technologies are increasingly serving as the connective tissue for AI agents and GraphRAG architectures, according to Tristan Baker (pictured), senior director and head of data architecture at Salesforce Inc.

“It’s becoming that critical piece that helps tie the end agentic experience that you want to deliver,” Baker said. “It’s almost like the glue that stitches what the customer or the person is asking to the data in the context that is needed in order to answer that question.”

Baker spoke with theCUBE’s John Furrier at the Neo4j GraphTalk event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how knowledge graphs, ontologies and layered data architecture are reshaping the pursuit of an enterprise system of intelligence. (* Disclosure below.)

Knowledge graphs and layered data architecture

The goal every leader now shares is a conversational interface that returns trustworthy answers in seconds, Baker noted. But delivering that at conversational speed requires far more than a single database, since different questions demand different retrieval structures.

“Don’t just give me the number,” Baker said. “Tell me how you figured it out and give me all the things it’s connected to.”

Baker frames the solution as a stack. At the bottom sit familiar lakehouses, operational and time-series databases, each optimized for a different kind of query. Above them, a metadata layer tracks where the truth about a customer lives across dozens of copies and maps business terminology to underlying columns. The graph, he explained, manages the relationships and context that connect those pieces together.

“Unless somebody’s withholding information from me, I don’t know anybody that’s actually totally solved this one yet,” Baker said. “So like most things, many systems are garbage in, garbage out — you can have the best technology in the world, but if you’re not careful about the content you’re exposing it to, then you’re going to end up with nothing very useful.”

The most under-discussed challenge, Baker added, is governance. As metadata moves up to a higher semantic layer, access control can no longer live only inside individual databases. Master data management and governance, he said, go hand in hand.

“In addition to needing a semantic description of the data, you start to also maybe need a semantic description of your access policy,” Baker said. “Your legal team … they’re not going to say, ‘Here’s how data should be protected in Postgres.’ They’re going to say, ‘People like this should not be able to access data like that.'”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the Neo4j GraphTalk event:

(* Disclosure: TheCUBE is a paid media partner for the Neo4j GraphTalk event. Neither Neo4j, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE