Sourcegraph Agent Advocate Stephanie Jarmak argues that developer advocacy is evolving into agent advocacyAI agent advocacy is the practice of making developer products understandable, discoverable, and usable by software agents while preserving trusted relationships with human developers.. Coding agentsAn AI coding agent is a tool-using AI system that can inspect, modify, and validate software within a repository. are becoming users of developer tools in their own right, and their behavior can influence which products developers discover, adopt and trust.
Jarmak explains that agent experienceAI agent experience describes how effectively a software agent can discover, understand, and use a product's interfaces, documentation, tools, and error feedback. is unusually measurable. Product teams can inspect traces, tool errors, token use, speed and task completionAgent evaluation tests whether an AI agent completes tasks correctly, consistently, and within its required boundaries. to identify friction, then improve documentation, APIs, MCP integrations and code-navigation tools for machine users as well as people.
Agents can also act as imperfect product recommenders because their training data and search results may be stale. Jarmak recommends current, structured and machine-readable product informationMachine-readable product information is current product data expressed in a structured format that software can reliably parse, compare, validate, and use without inferring meaning from presentation alone., while preserving human communities, credibility and feedback as central parts of developer advocacy.
Watch on YouTube
