How we measure Netlify’s Agent Experience

AXIS is an open-source tool and scoring framework designed to assess how well services function for AI agents, similar to how Lighthouse evaluates user experience for websites. Developed to improve the Agent Experience (AX), which pertains to how AI agents interact with platforms, AXIS helps identify and quantify the quality of these interactions by running scenarios against various AI agents and producing a score from 0 to 100. Initial tests on Netlify showed that providing agents with the necessary context significantly improved scores and reduced execution time and cost. The tool aims to address the inconsistencies in agent interactions by integrating seamlessly into engineering practices, providing continuous verification of agent context within documentation and development processes. As part of this effort, AXIS supports multiple AI agents and encourages community contributions to refine the tool further, positioning itself as a crucial component in the evaluation and enhancement of AI-agent interactions across different services.