In one sentence
An AI source is any page an AI assistant reads to build its answer, like a review roundup, forum thread or product page, and the sources it trusts decide which brands it names.
What sources AI assistants use for shopping answers
Assistants draw on two pools. One is training data: pages the model read before release. The other is live search: pages it looks up when the shopper asks, through grounded search.
For shopping questions, the pages that tend to feed answers are the ones that compare and judge:
- Review roundups and "best of" lists
- Forum and community threads
- Brand product and category pages
- Marketplace listings and customer reviews
- Video reviews and their transcripts
AI source vs. AI citation
Every citation is a source. Not every source gets cited. See AI citation.
Why AI sources matter for ecommerce brands
The assistant only knows what its sources say. If the roundups in your category skip you, the answer probably will too. If a forum thread calls your sizing inconsistent, that can show up in the answer as fact.
You can't write the answer directly. You can shape the pages it's built from.
How to find and win the sources AI uses in your category
- Ask your shopper questions and collect every page the answers cite. That's your visible source list.
- Sort it: your pages, competitors' pages, review sites, forums, marketplaces.
- Close the gaps on your own site: answer the questions the cited pages answer, with more specifics.
- Get onto the third-party pages that keep showing up: pitch the roundups, earn reviews, answer in the forums where your customers already talk.
About the author
Jenn Starr, Co-founder, Selix
Jenn has spent 20+ years helping SaaS startups build their marketing and implementation playbooks. She has bootstrapped three marketing agencies and is a 500 Startups alum. At Selix, she leads the work of getting ecommerce brands named in AI answers.
Related terms