Action centre
The audit does not stop at a score. It writes a numbered list of website changes, each tied to the buyer questions you lost and the competitor pages that won them.
What a row looks like
In the dashboard the list is titled Website Improvements. Each row has a title, a one-line reason, and a collapsed proof block: the lost questions, who was named instead, and the source domains. You can mark a row started, done, or dismissed.
- 01
Publish the missing comparison the answers already cite
Lost on several buyer questions to a competitor whose comparison page was in the citations.
- 02
Put pricing and limits on a crawlable page
Models skipped the brand on purchase-intent questions where rivals stated price in plain HTML.
- 03
Name the category the way buyers ask it
Answers used a category phrase that does not appear on the homepage.
These three rows are the shape of the list, not a customer result.
How priority is set
The writer ranks fixes by how many lost answers a change could move, using the stored answers, citation gaps, and competitor pages it actually read. The number you see is that sort order — 01 first — not a separate published scoring formula with weights.
The free audit includes the first prioritized fix. Plus and Pro see the full list. Recommendations are directional. They are not a ranking guarantee.
The coding prompt
Plus and Pro can copy one prompt that restates the open fixes for Cursor, Claude Code, or a similar tool pointed at the website repo. It includes the company facts from the audit, the lost buyer questions, and the competitor URLs to study — not copy. That is the same evidence already on the improvements page, worded so an agent can implement it.
Impact after you ship
On Pro, marking a fix done stores the visibility score at that moment. Later audits show the difference since then. The product copy says “since completing,” never “because of.” AI answers move for many reasons.