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

# Use cases

> Sixteen questions to ask once the MCP server is connected.

The server ships no prebuilt prompts, so this page is the substitute. Every question below is answerable with the tools your assistant already has, and each one names the tools that do the work.

<Note>
  Start any session with "list my brands in FixAEO". The slug that comes back is what the brand-specific tools need, and your assistant will reuse it for the rest of the conversation.
</Note>

## Visibility

**"Pull my visibility for the last 30 days and tell me which days moved most."**

`get_visibility` returns one row per day. Ask for the percentage as `brand_mentions / total_responses`, which is how the dashboard computes it — the raw `citation_rate` field is a 0..1 fraction and not the number you see in the app.

**"Which engine mentions us least, and is it getting worse?"**

`get_model_channels` gives a 0..100 score per engine over 30 days, with a delta that compares the recent half of that window against the older half. It's the fastest read on whether one engine has gone quiet on you.

**"Show me every negative mention from the last month, with the prompt that produced it."**

`list_mentions` with `sentiment: "neg"`. Each row carries the prompt, an answer snippet, the engine, and the source URL the engine cited.

**"Across all my brands, which one is falling behind?"**

`get_portfolio` returns top and bottom performers plus a per-brand table with deltas. Useful the moment you track more than one brand.

**"Which of my tracked prompts have never been scanned?"**

`list_prompts` marks each prompt with `last_snapshot_date`. A prompt with no date has never produced a snapshot, which usually means it was added since the last scan.

## Competitors

**"Who owns the biggest share of AI mentions in my category, and how far behind am I?"**

`get_industry_metrics` returns a ranked list with each brand's share as a percentage and a flag for your own row.

**"Which competitor gained the most ground over the last 90 days?"**

Same tool with `range: "90d"`. Each row's `delta` is in percentage points against the oldest snapshot in the window.

**"Build me a weekly table of my mentions versus my pinned competitors."**

`get_industry_metrics` again — the `time_series` block has per-day counts for every competitor, and `tracked_competitors` tells your assistant which ones you pinned deliberately.

## Sources

**"What are the top 20 domains AI cites in my space, and am I anywhere on that list?"**

`list_citations` ranks the domains and returns a `rank` block with your own highest-placed domain. Position `0` means you're not cited at all.

**"Which sites cite my competitors but never me? Rank them by opportunity."**

`get_industry_gap`. Each row carries a gap score, which competitors were cited there, and which engines did the citing. Your own domains are already excluded.

**"Do Perplexity and ChatGPT cite different sources for us?"**

One unfiltered call is enough — every domain row lists the engines that cited it. If you'd rather see each engine on its own, call `list_citations` twice with `engines: "perplexity"` and `engines: "chatgpt"` and diff the two lists.

## Agent traffic

**"Which AI crawlers hit the site this month, and are any of them getting errors?"**

`get_agent_traffic` returns a per-bot breakdown, a failure rate, and a per-status-code table. A rising failure rate usually means the crawler is hitting redirects or blocked paths.

**"What single page do AI crawlers read most, and which section of the site do they prefer?"**

Same tool. `top_page` is the most-crawled URL, and the per-folder breakdown shows whether they're living in your docs, your blog, or your product pages.

## Search gap

**"Which Google queries do we rank for but get ignored by AI? Give me the top 10 by opportunity."**

`get_seo_aeo_gap` classifies each of your top organic queries. Filter to `ai_state: "tracked_not_cited"` and take the first rows — they're already sorted so the biggest opportunity is first.

**"Which of my best organic queries aren't tracked as prompts yet?"**

Same response, `ai_state` of `untracked`. Those are candidates to add on [Setting up your prompts](/setting-up-prompts).

**"Compare our Google clicks with our AI visibility over the same 30 days."**

`get_search_performance` for the search side, `get_visibility` for the AI side. Both cover 30 days by default, so the windows line up without any extra work.

<Tip>
  The strongest habit is a weekly loop. Ask for `get_visibility`, `get_industry_metrics`, and `get_industry_gap` in one go, then have your assistant write the summary. It has the raw rows, so it catches the movement a chart flattens out.
</Tip>

## Related pages

* [Tools reference](/mcp/tools) — every parameter, and what each field means
* [Setup](/mcp/setup) — config blocks and the common connection errors
* [Improve](/improve) — what to do with a gap once you've found one