Read citations coverage
Which sources do AI engines cite when they answer about me?
When an assistant answers a question about your category, it reads sources first. Citations shows which ones — split into the directories that cite you, everything else the engines pulled in, and the well-known directories that never appeared. The third number is the only one you can act on directly.
The three tiles
On a Miami real-estate agency the split reads 4 · 87 · 7:
Directories citing you
- Count here
- 4
- What it is
- recognised directories that appeared as a source, with you in them
- What to do
- nothing — these are working
Other sources detected
- Count here
- 87
- What it is
- everything else the engines cited — blogs, local press, agent roundups, rival firms' own sites
- What to do
- read them for what earns a mention in your category
Key directories missing
- Count here
- 7
- What it is
- well-known directories that never appeared as a source at all
- What to do
- get listed, starting with the ones your customers actually use
Those three add up to Sources cited in the KPI row — 98 here. That total counts distinct sources across every answer, so it grows with how much the engines read, not with how well you did.
The missing list is the actionable part
Under Missing — get listed here the product names the directories outright rather than making you infer them. On this location: Google Business, Tripadvisor, Bing Places, Apple Maps, BBB, Foursquare and MapQuest, with a Fix in Listings → button that takes you to the tool that fixes it.
Google Business appearing in that list on a location scoring 81 is worth pausing on. It does not mean the profile is broken. It means that across these particular answers, no engine cited Google as a source — they cited property blogs and agent directories instead. Different categories get read differently, and a directory that matters enormously for a restaurant may never surface for a real-estate firm.
Read the per-engine columns for where to publish
Below the tiles, each engine lists the sources it actually cited. They overlap far less than you would expect:
| Engine | Sources it leaned on for this location |
|---|---|
| ChatGPT | Almost entirely individual agency and agent websites — bluepropertiesre.com, kamanyrealty.com, mikeplaton.com |
| Gemini | Review and comparison sites — goodfirms.co, ratemyagent.com, expertise.com, birdeye.com |
| Perplexity | Large portals and trade press — realtor.com, zillow.com, therealdeal.com |
That is a content strategy in three lines. Being cited by ChatGPT in this category means having a substantial site of your own; being cited by Gemini means being present and rated on comparison sites; being cited by Perplexity means being in the big portals.
A high citation count is not visibility
The clearest illustration is a location with zero AI visibility. A Fairbanks bed & breakfast scores 0/100, is named in no answer at all, and still shows 36 sources cited.
Nothing is wrong with that number. The assistants read 36 sources while answering questions about bed & breakfasts in Fairbanks — none of them about this one. Sources cited measures what the engine read, never who vouched for you.
Questions people ask
Getting listed didn't change my citations. Why?
Reports are monthly snapshots, and being listed is not the same as being cited — the engine has to choose your listing as a source when it answers. Expect the missing count to move first, and citations to follow only if that directory is one the engines read for your category.
Why is a competitor's own website in my citations?
Because the engine cited it while answering your question. Rival sites appearing as sources is normal and is a useful signal: it tells you what the assistant considered authoritative enough to quote.
Where to go next
- What should I fix first? — where the missing-directories count becomes a ranked action
- How do I update a location's listings? — the tool the Fix in Listings button opens