Human in the Loop: Governing API Evolution in the Age of Agents
summary
Claims a proposal-based governance flow that keeps humans in the loop while agents move fast: agents propose schema changes, create branches, and the system routes each proposal to the right ownership at type/field level, runs deterministic federation checks (the demo uses WonderGraph Hub and a fision engine), shows blast radius via impact analysis against real traffic, and requires explicit approval or rejection; works with hundreds of subgraphs (customers report >250), supports override directives but blocks unsafe federation, and ships only after gated reviews.
AI-written summary · about
chapters (9)
- 0:00Human in the loop API governance
- 2:35Agents in API evolution
- 4:22Federation complexity issues
- 8:35Communication overload in APIs
- 10:42Ideal schema evolution workflow
- 12:05Current schema review flaws
- 14:10Layered API governance solution
- 18:25Live demo of governance
- 25:02Q&A
related talks
25mThe Invisible Fortress: Embedding Zero-Trust Governance in the Supergraph
33mHumans on the loop, not in it: Taking agentic engineering end to end
21mThe Future of API Management: Autonomous Agents in Action- 33mHumans on the loop, not in it: Taking agentic engineering end to end
33mGoverning the Ungovernable: Security and Compliance for AI Agents in Open Source Projec...
21mPlatforms for Humans and Machines: Engineering for the Age of Agents
45mFishbowl Discussion on APIs and AI
23mDesigning APIs Both Humans and AI Love