Why it matters
Web-grounded assistants search first and write the answer from what they find. Perplexity documents PerplexityBot as the crawler behind its results, and OpenAI documents OAI-SearchBot as the crawler behind ChatGPT’s search features. Which brands an answer names depends on which pages it read.
If the comparison article the assistant retrieves lists three competitors and not you, the answer does too. The shortest route to being named is often a page you do not own.
The workflow
Collect the citations behind your category’s answers
For each tracked question, note every domain the answer cites. Perplexity, Google AI Overviews and AI Mode, and ChatGPT’s search results all show their sources. Do this per market: the sources in Colombia are not the sources in Brazil.
Classify the sources and count them
Sort the domains into review platforms, directories and marketplaces, press, forums, public bodies and competitor sites. Count how often each is cited across your questions; that is its citation share.
Check whether each source names you
Open the most-cited pages and read them. Which brands do they name? A “best payroll software in Mexico” article that lists five competitors and not you is a source gap. Note which brands are named alongside each other; that is who the assistant thinks belongs in your category.
Sort the gaps by severity
The domains cited often with competitors and rarely with you come first. Then weigh how realistic each one is: a directory or marketplace profile can be completed this week, a review platform needs customers, a publication needs a reason to write about you.
Work the list, one source at a time
Create or complete profiles where comparisons happen. Ask customers for reviews on the platforms the assistant reads, not only on your site. Pitch the publication with a correction or an inclusion backed by plain facts. Where your own page was retrieved but not cited, make it answer the question directly, short answer first. None of this guarantees inclusion; it changes what the assistant can find.
Let VozRadar find and sort the gaps
Sources builds this list from your tracked answers, checks what the cited pages actually say, and turns each gap into an opportunity with a suggested next step.
Run it in VozRadar
Sources
- Per cited domain: its type, the answers citing it, citation share, whether you were named when it was cited, and the brands named alongside
- Source gaps tab: third-party domains cited with competitors but rarely with you, each with severity, evidence and a suggested next step
- “Check cited pages” fetches the most-cited third-party pages (robots.txt respected) to see which brands they actually name
- Your pages tab: which of your pages are cited and which are retrieved without citation; gaps become opportunities with an outreach recommendation
In the app: Project → Sources. Source-gap opportunities appear under Fixes for the primary market.
See how AI sees your brand.
Start with one market. Expand when you're ready.