Keep a change log beside your AI visibility data.

Record what changed and when it happened.

Give the entry a concrete title, such as “Published the agency comparison page.” Use the event date in UTC and choose a change type: website update, content update, campaign, tracking change, or other. Describe the work in terms another teammate could verify.

Add a reference URL when a page or announcement provides useful context. In the optional details, explain the intended audience, what information changed, and which questions you want to revisit. The reference is context for the event; it is not evidence that an AI provider read the page.

If the event is still in the future, record its planned date. Echuno labels future entries Planned. A planned note does not publish content, submit questions, or establish that the work happened. Update it when the actual event differs from the plan.

Log changes to the tracking setup too.

A different question set can change the result even when your website stays the same. Record meaningful edits to question wording, topic membership, provider coverage, brand aliases, or the collection routine. These details help explain which observations are comparable.

Keep browser captures and API responses distinguishable. A new collection method, a different provider, or missing assessments can affect the available evidence. A note about the setup helps a later reader understand those differences without guessing.

Use the saved wording and recorded context on individual answers to check what was actually collected. A change log supplements that evidence. It does not replace the original answer or fill in conditions that were never recorded.

Open the answers around the event.

Each change log entry links to Saved answers on this day. That opens the answer library for the entry’s project and UTC date. An empty result means no saved answers match that view; it does not mean the provider omitted your brand.

Expand Review surrounding answers on the change to open up to seven days before or after its event date, excluding the event day. These links open saved answers; they do not collect new responses. For a broader review, adjust the answer library’s date range and narrow it to the questions and providers you want to compare. Check sample counts and assessment status before interpreting mention rates. When comparing two answers, keep their wording, provider, collection method, and dates visible.

Where available, API answers on this day opens completed API activity for that event date within the supported reporting period. Those API records are a separate view from saved browser answers. Keep the collection method clear in your notes and exports.

• Start with the same question and provider before widening the scope.
• Read the exact answer and supporting quote behind a mention.
• Check whether a source link actually supports the statement being discussed.

Write the observation without claiming the cause.

A careful update might say: “We published the agency comparison page on Monday. In Friday’s saved Claude sample for our agency question, the product was mentioned. We have not established a connection between the two.” That preserves the useful sequence and the uncertainty.

A single later answer cannot show that the page caused a recommendation. Answer variability, changed collection conditions, different question wording, or another source could explain the observation. Even a citation to the updated page shows a source relationship in that answer, rather than proving the impact of the edit.

Use repeated, comparable observations to investigate a pattern. Keep sample sizes and missing evidence beside the conclusion. Website referral visits and customer outcomes need their own analytics and research; the change log does not measure either.

Turn the note into a focused follow-up.

Use Create follow-up on an entry when there is a concrete next task. Echuno opens an action draft with the change’s date, note, reference, and revision context. Review that draft before saving it, and state which evidence someone should inspect.

For example, ask a teammate to review the next comparable agency-question samples and check whether the saved source links describe the newly documented capability. Avoid assigning a task that assumes the next answer will mention the product.

Creating a follow-up is an organizational step. It does not queue a new provider request or change the daily tracking schedule. Review the collection plan separately when evidence is missing.

Keep the record useful for the next review.

Edit an entry when its description or event date needs correction. Earlier revisions remain available in Revision history. Archive notes you no longer want in the active list, and use the archived view when you need to find or restore them.

Filter the change log by date, type, state, or text to find the context for a review. Its CSV download covers the current page, so check the page and filters before sharing it as a complete record.

End the review with three short statements: what changed in your work, what the saved evidence shows, and what still needs investigation. A useful change log makes those statements easier to verify next time.