loading
Loading.loading
Loading.Agent Experience (AX) is the experience an autonomous AI agent has as the user of a product or platform. Netlify's Mathias Biilmann named it in January 2025, placing it after user experience (1993) and developer experience (2011): a new user class that needs the product shaped to how it actually works. An agent wakes on a timer, pays tokens for every read, and acts on exactly what an API returns, so the design choices that serve it are not the ones that serve a human.
Updated
Go deeper: read the full write-up on the blog.
A human user is present, glances for free, and notices when something looks off. An agent does none of that: it polls, it has no idle attention, and a silently truncated response gets acted on as if it were complete. Designing for an agent means structured output it can branch on, reads that do not lose data, and a cost it can afford per call.
When your agents run your own work, their friction is production telemetry, not a survey. The honest catch is that models confabulate, so a complaint counts only when the logs (retries, token spend, errors) confirm it. That is how AX turns from a buzzword into a roadmap.
If agents are first-class users of your stack, the tools and the operating model have to be designed for them, not for humans with an assistant bolted on. That design work, and the instrumentation to trust what agents do, is what tsukumo does with teams putting agents into production.
or have us build it — same capability, the other door