Opening Keynote: Exploring Context Graphs: From Data to Decisions
summary
Panel claims a decision trace—a recorded chain of agent tool calls and human interactions—lets systems capture the “why” behind outcomes and form a reusable context graph. Demos show a financial‑services full‑stack app that ingests Google Workspace, Linear and code history, uses 22+ ontologies via Create Context Graph, and materializes structural embeddings for retrieval. Production challenges are concrete: instrumenting decisions, governance and access control across systems, and applying trust scoring to decide when agents proceed or escalate.
AI-written summary · about
chapters (7)
- 0:00Introduction and context
- 2:04Defining decision traces
- 5:18Infrastructure and trust
- 8:19Full stack context graphs
- 12:15Challenges in implementation
- 16:00Ontology and discovery
- 23:52Future outlook and wrap-up
related talks
56mConnecting the Dots with Context Graphs
17mContext Graphs for Explainable, Decision-Aware AI Agents
20mWhy your agents need decision traces, not just documents
23mConnecting the Dots With Context Graphs
38mSearch is Back: Solving the "Context Crisis" for AI Agents
18mConnecting the Dots with Context Graphs
29mHow to Structure Context to Make Your Agents Smarter
15mUnderstanding Context Engineering for Agentic Systems