AI and the state of APIs in 2026
summary
Prediction: by 2029 about 30% of AI usage runs through agents, not humans, and APIs remain the plumbing that agents call. Practical moves include adding an MCP bridge over existing gateways, providing machine‑readable specs and sandbox keys for easy agent access, and building an AI gateway that mediates LLM, tool, and agent calls. Expect screen scraping to cost up to 45x more tokens than structured APIs, so governance, discovery, and token‑aware throttling become urgent.
AI-written summary · about
chapters (17)
- 0:00AI and APIs intro
- 1:09Agentic AI vs APIs myths
- 2:55APIs still foundational
- 3:38AI-built APIs and tools
- 6:05AI consuming APIs
- 7:30Agent-driven headless trends
- 9:36APIs questioned future
- 10:59API strategy for agents
- 12:27MCP and A2A protocols
- 15:14APIs as AI toolkits
- 17:00MCP governance challenges
- 18:22AI gateways rise
- 20:08MCP and AI gateways vendors
- 21:34Enterprise agent governance
- 22:37AI ecosystem discovery issues
- 23:41Three future strategies
- 24:43Q&A
related talks
23mAI and APIs in 2026
29mAPI Management Recent Trends
22mLeading the future: AI and APIs, a powerful duo
26mIf your APIs aren’t agent‑ready, your AI agent is just chat: [...]
27mAre APIs becoming obsolete? Think MCP, ACP, A2A, AGNTCY etc.
22mBusiness Wants In—Here’s How to Drive API Wins (and Keep Your Sanity)
13mAre Your APIs Ready for AI Agents?
45mFrom Tools to Partners: The Rise of Agentic AI