The Sightspool Journal

Thinking about the UX function after AI.

Practical essays on connected evidence, specialist agents, product judgement and the operating systems that help teams learn from what they ship.

Product evidence24 July 20267 min read

Analytics is a sensor. It is not the UX function.

PostHog can tell you what its instrumentation captured. The harder work is deciding whether the right thing was captured, what it means and what the team should do next.

Read article
Operating model24 July 20268 min read

What an AI-native UX function actually does

The UX function is not a list of roles or deliverables. It is the capacity to notice uncertainty, choose the right method, change an action and learn from the result.

Read article
Agent governance24 July 20266 min read

No proof, no post: building UX agents a team can challenge

A persuasive AI answer is not the same thing as a defensible UX read. Trust comes from structural proof, visible uncertainty and the ability to disagree.

Read article
Human assurance24 July 20267 min read

Human escalation is a product feature

The strongest AI product is not the one that hides every human. It is the one that knows which calls deserve human judgement and can explain why.

Read article