From Rules to Reasoning: How Context Graphs Are Powering Smarter AI

Finance teams are racing to add AI agents to their workflows while trying to determine what controls and checks are critical to have in place. But Foundation Capital argues that agents need more than clear rules. They need “decision traces”: a record of what context the agent used, which policy it applied, what exception it granted, who approved it, and what precedent shaped the outcome. Over time, these decision traces connect into something bigger: a context graph that can break down the AI’s reasoning for a certain decision.

Hear from Jaya Gupta, partner at Foundation Capital and co-author of “AI’s trillion-dollar opportunity: Context graphs” and Ramp’s Gayatri Sabharwal, who writes on AI trends, on context graphs and what comes next for AI in finance.

Learning Objectives:
• What decision traces are and why they capture what rules never could - the context, precedent, and approvals behind every automated decision.
• How context graphs connect approvals, policy exceptions, and precedent into a record your AI can actually learn from.
• How systems of agents that capture decision-time context pull ahead of legacy architectures over time and why the gap widens with every decision.
• How to focus on the use cases that matter in day-to-day finance work, and where context graphs play a role as workflows get more rules-driven.