The role of APIs in Agentic Transformations
summary
Shows a concrete architecture for turning APIs into intent-driven, discoverable capabilities using a semantic layer over existing systems, demonstrated with a grocery order fulfillment example that decomposes one customer intent into inventory, pricing, fraud, payment and delivery tasks. Proposes an extensible orchestrator built on ReAct patterns that uses a task/knowledge graph, skills registry, and specialized sub-agents, plus Open Policy Agent for rules, plus auth, rate limits, input/output validation and observability to prevent runaway or unsafe agent behavior.
AI-written summary · about
chapters (11)
- 0:00APIs in agentic transformations
- 1:35Retail order fulfillment example
- 2:58Redesigning APIs for self-service
- 5:50Customer intent to task execution
- 7:55Simplified agentic orchestration
- 9:59Semantic layer over enterprise systems
- 11:04Security for autonomous agents
- 13:12Auditability in agentic systems
- 13:55Design principles for agentic APIs
- 15:39Industry use cases
- 16:20Q&A
related talks
35mUnleash Your Inner API Architect: Build Agents That Deliver
21mThe Future of API Management: Autonomous Agents in Action
29mAgent-Ready Orchestration: Engineering the Execution Layer.
26mDo you speak Agent? The convergence of APIs, MCP, A2A, and Agent Languages.
22mSurviving an interconnected world with API governance
22mWhen AI Agents Start Acting, APIs Become Execution Contracts
07mThe Transition from Strict APIs to Flexible, Context-Aware Interfaces
45mFishbowl Discussion on APIs and AI