Every cloud platform was designed for developers: humans who read docs, click dashboards, and push to git. But AI agents are already using these platforms, and they experience them very differently. At Render, we've watched agents parse our marketing pages, struggle with our APIs, and surface (or not surface) our platform in LLM recommendations. Building our MCP server, CLI, and agent skills meant designing for two users at once, and rethinking what "developer experience" even means when the developer isn't human. This talk distills what we learned: where our assumptions failed, what we changed in response, and the concrete principles engineering leaders can apply to build tools, docs, and APIs that serve both humans and AI agents.
This talk has been presented at TechLead Conf Amsterdam 2026: Adopting AI in Orgs Edition, check out the latest edition of this Tech Conference.























