Citability audit

The citability audit is the centerpiece of the diagnosis: for one prompt where a competitor is cited and you are not, it fetches both pages, splits each into sections of about 120 words, picks the passage on each side that best answers the prompt, and returns the likely reason the competitor wins plus what to change. The result is a hypothesis to test: the next runs show whether your change moves the citation.

Where to find it

On the Gap diagnosis tab, pick a prompt in the gap list: its audit opens next to the list, or as a sheet from the bottom on a phone. Until you pick one, the first listed prompt with an audit is shown. A prompt that has no audit yet offers Audit this prompt.

How the competitor is chosen

The audit compares against the most-cited domain that is not yours across the AI answers for this prompt in the selected period, and audits that domain's most-cited page from the same answers. For prompts where you were never cited, it judges parity differently: whether your content holds up while the gap lies elsewhere.

What the audit examines

1. Passage diagnosis: your best passage versus the competitor's winning passage, scored on how well each answers the prompt
2. Concepts your page never addresses, mined from the actual AI answers for the prompt
3. Answer shape: does the page answer immediately, use question-and-answer structure, show freshness, allow AI crawlers, and respond fast
4. A suggested direct-answer sentence, with the placement, ready to copy into your page

Reading the results

The audit opens with the prompt, its result class, the competitor page it analyzed and the audit date, then the signal comparison, the passages with the suggested sentence and a short summary of why the competitor wins. The numbered fixes and the earned-media pages are folded away below the summary; open them when you are ready to act.

The written parts of the audit (summary, fixes, suggested sentence and its reasoning) are in the topic's language, so a German topic gets a German audit whatever language your account is set to, and everyone on the team sees the same text. Labels and signal details follow your interface language.

The score bars compare page content only. Which sources AI assistants actually cite is also driven by off-page presence: mentions, reviews and communities on third-party sites. When your passage matches the competitor's but you still lose the citation, the audit says so and points you to the earned-media view: the pages the AI relies on and whether they feature you.

Fallbacks and limits

If your page blocks automated access, the audit says so and analyzes an archived copy when one exists. Pages that load nothing but navigation, or block bots without an archive, produce a summary-only audit. When the winning source is a video (YouTube, Vimeo), no audit is offered: a video page has no readable text to compare against, so the audit shows a note instead of the Audit this prompt button. When the inputs change after an audit, for example a new winning page or a newly pinned target page, the audit offers a re-run. Audits use a monthly quota: 25 on Pro, 100 on Agency.