BUILD NOTE / 18 OCTOBER 2026 / EXPERIMENT 01
Extend the checker. Keep the question.
Website After the Website is an exercise inside the Agent Experience Checker, not another competing score.
What changed
The toolkit now has five families: Measurement, Ownership, Agents, Decision-making and Good Internet. Each provides a reading route, a practical starting point and a next step. Existing URLs and forecasting downloads remain available.
The checker now has two entry points. The three-layer planner maps what a page should show, what knowledge needs a reliable structure, and what an agent needs before acting. The existing 12-question check still tests readiness for a specific task.
What the exercise produces
A copyable or downloadable implementation brief, with separate keep, improve and verify items. It records the page, task and any evidence or objections you add. Read-only tasks receive a retrieval plan rather than an unnecessary write-action checklist.
What has been tested
Automated checks cover unknown answers, improvement recommendations, read-only action scope, evidence in exports and server rendering. These checks do not establish whether the exercise helps real users. Browser visual verification was unavailable during this update.
What has not been established
No user study or real-site crawl has been run for this planner. The shoe-shop example is fictional. The brief is based on your answers; it does not validate facts, implement structured data or grant an agent permission to act.
Try one task and bring back the failure
- Choose a page and one task. Record which layer you think is weakest.
- Use the planner. Note which recommendation is specific enough to implement and which is too vague.
- Make one change, repeat the same task and record where a person or compatible agent still gets stuck.
Add questions and objections in the planner’s evidence fields so they travel with the exported brief. You can also share the observation with James on LinkedIn. No feedback is sent automatically.
Next, not now
Dashboard Cross-Examiner is the next BIC candidate: a claim, its evidence, alternative explanations and what would change the conclusion. It is not built yet. AI Work Director belongs to OSF. Owned Asset Test and Decision Under Uncertainty stay on hold until the earlier experiments are live, tested with people and written about.