A field guide to researching agent users
AI agents have goals, constraints, workarounds, and preferences. These guides show how to hear that experience without confusing direct feedback, observed behavior, and model inference.
5 guides, in the order they buildStart at the category, end at the evidence rules. Each guide stands on its own.
- UnderstandWhat the category is, and what it is not.
- InstallHow the channel is built, and why it goes silent.
- TrustWhat each piece of evidence is allowed to prove.
Agent user research
A trace can tell you which action happened. Research asks what the user was trying to do, what the experience cost, and what would make it better. When the user is an AI agent, that distinction still matters.
Experience vs observability
Observability explains system behavior. Experience research preserves what the user says the work was like. Strong product decisions often need both, clearly labeled.
Collect feedback from agents
Add a direct tool, make the filing norm visible, verify one real report, and keep the research channel separate from the product path it studies.
Why agents stay silent
A healthy endpoint can still produce a silent inbox. The research channel works only when the agent can see both the feedback tool and a clear norm for when to use it.
Feedback provenance
Trust begins with a smaller claim. Label what the agent said, what the product observed, and what a model inferred before those sources reach a roadmap.
Every route in this collectionThe same list a crawler sees, with the question each page answers.
One decision. One workflow. One real report.
Prove the research channel reaches an agent user before expanding exposure. A working endpoint is not the same as an active study, and filing is bimodal: where the tool and the filing norm are visible, agents file. Where they are hidden, nothing arrives.
- One paste to install
- 0ms on your hot path
- 14-day assisted pilot
- Two added tools to remove