Panel: DX in Enterprise APIs with our new consumers - Agents
summary
Panel lays out concrete moves: an agent that analyzes DOS spikes and ASNs as a working Vercel firewall observability agent, a data‑science agent built from an initial single‑prompt then expanded into 20 tools across planning/build/execution phases, and an MCP pattern that bundles linear steps into one deployment‑diagnostic tool which improves reliability vs 80%‑right/20%‑fail chains. Notes security gaps: MCP servers lack versioning, clients cap useful tools (~28 tools), and guardrails, persistent queries and gateways are necessary.
AI-written summary · about
chapters (10)
- 0:00Panel on enterprise APIs and agents
- 1:04Next-gen AI agents
- 6:01APIs for AI-driven workflows
- 8:52MCP servers for agentic experiences
- 12:26Security in agent automation
- 15:58Flexibility versus control in agents
- 21:55MCP’s role in enterprise systems
- 25:52Agentic development tools
- 33:57Q&A on MCP gateways
- 44:50Closing remarks
