The sources AI assistants cite, and how to earn a place in them
By Ben Frampton · · 5 min read
Short answer: for "which product should I buy?" questions, AI assistants mostly cite pages that aren't owned by any one brand: review and comparison articles, community threads (Reddit and forums), review platforms, directories, trade press and sometimes YouTube. Find out which of those sources the assistants cite for your buyer questions, then earn an honest mention in the ones where your competitors appear and you don't.
Why do third-party sources matter so much?
A brand's own site is, understandably, biased. When an assistant answers "what's the best X for Y?", it tends to prefer pages that compare several options or reflect many people's experience. Those pages also tend to rank well for recommendation and comparison searches, which is what assistants retrieve.
The result: your own page may be cited when someone asks about you specifically, while competitors get recommended for the general question because they appear in the roundups and threads the assistant reads.
What kinds of sources get cited?
It varies by assistant, category and country, but the usual types are:
| Source type | Examples | Why assistants use them |
|---|---|---|
| Comparison and "best of" articles | Publisher roundups, independent review blogs | Directly answer "best X for Y" |
| Community discussions | Reddit, specialist forums, Q&A sites | Real users' experience and opinions |
| Review platforms | G2, Capterra, Trustpilot, TripAdvisor (by category) | Many reviews, structured comparisons |
| Directories and marketplaces | Industry directories, app marketplaces | Category lists with descriptions |
| Trade and news press | Industry publications, local news | Authority and recency |
| Video | YouTube reviews and tutorials | Hands-on comparisons |
| Reference sites | Wikipedia, Wikidata | Background facts about companies |
Assistants differ in their habits. Some cite community threads heavily, others lean on publishers or on their own search engine's top results. That's why it's worth checking the actual citations per assistant rather than following a general list.
How do you find the sources that matter for you?
- Collect answers to your buyer questions on each assistant you care about (see how to choose which prompts to track).
- List every cited URL and domain, grouped by question and by assistant.
- Mark which ones mention you, which mention competitors, and which mention neither.
- Rank the gaps: domains cited often, for many of your questions, that mention competitors but not you.
That list is your outreach plan. It is usually short. A handful of domains tends to account for a large share of the citations in a given category.
How do you earn a mention honestly?
What works depends on the source type:
- Comparison articles. Contact the author or editor with accurate, useful information: what you do, who it's for, pricing, and how you differ. Offer a trial or demo for a fair test. Many writers update roundups regularly and welcome corrections and new options.
- Review platforms. Ask real customers to leave reviews, at a natural moment (after a successful onboarding, a renewal). Keep your profile complete and consistent with your site.
- Community threads. Take part as yourself, openly identified, and only where you can help. Answer questions; don't drop links. Undisclosed self-promotion gets removed and remembered.
- Directories. Make sure you're listed in the main ones for your category, with the same description everywhere.
- Trade press. Offer data, expert comment or a genuinely newsworthy story, not a product announcement.
- Wikipedia. Only if you meet its notability rules, and never by editing your own article. Focus on getting the independent coverage that notability depends on.
What to avoid
- Paying for undisclosed placements in "independent" reviews. Beyond the ethics, many jurisdictions treat it as misleading advertising.
- Fake reviews or sock-puppet accounts. Platforms detect them, and the damage to trust outlasts any short-term gain.
- Mass-produced pages written only to be cited. Assistants and search engines are getting better at ignoring them, and they add nothing for readers.
How long does it take to see results?
For assistants that search the web, a new mention on a frequently cited page can show up in answers within days or weeks of the page being recrawled. Influence on the model's built-in knowledge takes until the next model is trained. Measure before and after, with enough answers to tell a real change from noise. See why one AI answer tells you almost nothing.
Frequently asked questions
Is Reddit really that important for AI search?
In many categories, yes: community threads are cited often, particularly for consumer products and software. But it varies by assistant and category. Check your own citation data before investing heavily in any one platform.
Should I create my own "best X" listicle that includes my product?
You can, if it's honest: clearly from you, fair to competitors and genuinely useful. Assistants may use it, but an independent source naming you carries more weight. See how to write comparison and alternatives pages.
Does a link matter, or just the mention?
For AI answers, the mention and how you're described matter most. An assistant can learn "Brand X is the budget option for freelancers" from a page that never links to you. Links still help classic search ranking.
What if a cited source says something wrong about us?
Ask for a correction, politely and with evidence. Outdated or wrong sources are a common cause of assistants describing a brand incorrectly. See what to do when AI gets your brand wrong.