> ## Documentation Index
> Fetch the complete documentation index at: https://manual.botric.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# GEO Audit

> A readiness check on how citable your site is to AI answer engines.

Where [Site Audit](/search/site-audit) asks *is this site technically sound?*, GEO Audit asks *is this site something an AI engine would confidently cite?* Different question, different findings.

## Running it

GEO Audit runs automatically the first time you open the page for a site that's never been audited. After that, **Re-run** starts a fresh scan. A running scan shows live progress; the finished report can be downloaded.

If a workspace has no site yet, the page prompts you to add one first.

## What it scores

The report opens with an overall citability score and a band — **Excellent**, **Strong**, **Fair** or **Needs work** — then breaks the score into sections.

<AccordionGroup>
  <Accordion title="Technical readiness">
    The mechanics an answer engine depends on: clean markup, structured data, and content it can parse without executing your application.
  </Accordion>

  <Accordion title="AI crawler access">
    Whether the crawlers behind the major AI platforms are permitted to fetch your pages, and whether anything in your configuration blocks them.
  </Accordion>

  <Accordion title="AI platform readiness">
    How well your content is shaped for the way each platform assembles answers.
  </Accordion>

  <Accordion title="Trust stack">
    The corroborating pages that make a brand safe for a model to cite: an **About page**, **contact information**, a **LinkedIn** presence and similar signals. Models prefer sources they can verify.
  </Accordion>

  <Accordion title="Knowledge-graph presence">
    Whether your brand exists as an entity where models look it up — **Wikipedia**, **Wikidata**, **Crunchbase**. Missing entries are called out with a **Claim listing** action.
  </Accordion>

  <Accordion title="Query intent coverage">
    Which intent types your content covers, shown on a radar chart. Gaps here map directly to prompts you're likely to lose.
  </Accordion>

  <Accordion title="Freshness risks">
    **Content decay** — pages stale enough that a model may prefer a more recently updated source.
  </Accordion>

  <Accordion title="Risk flags">
    Things that actively work against you:

    * **Hallucination bait** — vague or contradictory claims a model may reword into something wrong.
    * **Keyword stuffing** — patterns that read as manipulation.
    * **Prompt-injection surface** — page content that could be interpreted as instructions by a model reading it.
  </Accordion>

  <Accordion title="Top citability gains">
    The ranked list of changes that would most improve your citability. Start here.
  </Accordion>
</AccordionGroup>

## Using the results

<Steps>
  <Step title="Clear the blockers">
    AI crawler access and prompt-injection surface first. Nothing else matters if engines can't or won't read you.
  </Step>

  <Step title="Build the trust stack">
    An About page, real contact details and a knowledge-graph entry are cheap wins with outsized effect.
  </Step>

  <Step title="Close intent coverage">
    Use the radar chart to find the intents you don't answer, then send those to the [Blog Agent](/content/blog-agent).
  </Step>

  <Step title="Re-run and compare">
    Re-run after publishing and watch the band move.
  </Step>
</Steps>

## If a scan fails

A network or scan error stops the audit and the page offers **Retry**. A previously completed report stays available in the meantime.

<CardGroup cols={2}>
  <Card title="AI Visibility (GEO)" icon="sparkles" href="/discoverability/ai-visibility/overview">
    Readiness is the input; visibility is the outcome.
  </Card>

  <Card title="Blog Agent" icon="pen-nib" href="/content/blog-agent">
    Fill the intent coverage gaps this audit found.
  </Card>
</CardGroup>