Raymond James Uses Neo4j Graph to Evaluate AI Agent Results

A Neo4j knowledge graph is helping Raymond James Financial combine information from separate systems and check AI-agent results against trusted internal sources before launch.

Naveen Mahankali, the firm’s vice president of technology, described the checks as comparisons with the company’s “golden sources.” “We run evals compared to the golden sources,” Mahankali said.

Jeff Griffith, the firm’s vice president of AI engineering and data science, said graphs can help with requests that cross business areas or go beyond information already surfaced in existing applications. He described their use for cases “when you want to cross those domains or do some unique thing.”

Raymond James is also working on dynamic data queries for financial advisors. The planned use would let advisors request product analysis or compare account balances without relying on fixed dashboards.

Knowledge graphs have also featured in enterprise AI infrastructure work, including McKinsey’s EcliptOS project.