AI agents are now the first to read your docs and touch your platform—not developers. DevRel must adapt by writing for machines: structured, unambiguous, and pattern-rich. If the agent doesn’t get it, the developer never will.
We’ve long defined DevRel as the function that helps developers succeed on a platform through docs, demos, content, and community.
The goal has always been the same: reduce friction, build trust, and spark adoption.
But something fundamental is changing right now, which is why I'm writing this post.
Developers aren’t the ones reading your docs line by line anymore. They’re asking Cursor. They’re pasting errors into Claude. They’re prompting Copilot to build entire flows.
The developer is still there, but the first touchpoint with your product is often mediated by an AI agent.
That changes everything. It means you’re not just advocating to developers anymore. You’re advocating to the AI agents they rely on. If your docs are ambiguous, the agent will hallucinate. If your error messages are vague, the agent won’t fix them. If your SDKs don’t follow conventions, the agent won’t use them right. So what does this mean for DevRel? It means you need to think like a prompt engineer. Write docs for LLMs, not just humans. Use clear, consistent patterns the model can recognize. Tag content, surface context, clarify edge cases. Make your platform machine-readable in every sense. The best DevRel teams will treat AI agents as a new developer persona. The agents are the ones writing code now. If they don’t understand your platform, the developer won’t either.
DevRel’s next frontier isn’t just developer-facing. It’s agent-facing. And if you’re not thinking about that yet, you could already be falling behind.
