83%
What they are
There are two kinds:
- Expert review sites test products and write recommendations, like Wirecutter, RunRepeat or NerdWallet.
- User review platforms collect customer reviews, like G2, Capterra and Trustpilot.
Why AI leans on them
Review sites answer exactly the question AI is asked: which is best, for whom, and why. They also rank well in search, so AI finds them.
In some categories one site dominates. In our study, NerdWallet was cited in 83% of small-business credit card answers, RunRepeat in 79% of running shoe answers, and SafeHome.org in 73% of home security answers.
How to show up
- Find the one or two sites your category's answers cite most.
- For expert sites, pitch the writer and offer a product to test.
- For user platforms, ask happy customers to leave reviews, and complete your profile.
- Reply to bad reviews calmly and fix the cause.
Example
A running shoe brand checks AI answers for "best trail running shoes" and sees RunRepeat cited in most of them. Its newest shoe hasn't been reviewed there.
It sends pairs for testing. Once the review is up, the shoe starts appearing in answers that cite it.
Frequently asked questions
Are user reviews or expert reviews more important?
It depends on the category. Software answers lean on G2 and Capterra style sites, consumer products on expert testers. Check what your answers cite.
Can I pay to be on a review site?
Some sell profiles or ads. Paid placement rarely changes an expert's verdict, and AI reads the verdict.
What if a review site ranks me badly?
Ask what would change their view, fix it, and ask for a re-test. A fair low score is better than no mention.