A Foundation Capital paper introduces 'context graphs' and 'decision traces' as a new paradigm for enterprise AI. Context graphs are knowledge graphs that capture the full reasoning, causal relationships, and decision history behind business decisions, complementing existing systems like ERPs and CRMs rather than replacing them. The author argues that decision traces alone are insufficient — AI also needs semantic memory (facts and schemas) and procedural memory (skills and operating principles) to avoid hallucination. Context graphs act as a 'graph of graphs,' guiding AI agents across enterprise data sources. GraphRAG is identified as a key component for structured knowledge retrieval. The overall advice to enterprise teams is to experiment with these ideas but avoid locking in any single architecture given how rapidly the landscape is evolving.
