Discussion starter: an AI search question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- AI Search
Self-serving best-of listicles that rank the publisher first get cited by LLMs a lot. Is that tactic durable, or living on borrowed time?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
I would treat it as living on borrowed time. Self-serving “best of” lists work in AI answers today for a mechanical reason: assistants retrieve pages that rank for “best X” queries and summarise lists easily, and a brand can often rank its own list for its own category. But the format fails almost every test that search engines already apply to reviews, Google’s own AI guidance warns against manufacturing mentions, and there are obvious ways for AI systems to discount a list whose author is also its winner. The durable version is different: a comparison page that says openly whose it is, compares on stated criteria, and admits where competitors are the better choice.
Why the tactic works right now
- Retrieval follows rankings. Many assistants ground answers in search results, and “best X” lists rank well for “best X” prompts because they match the query format.
- Lists are easy to summarise. A ranked list with short descriptions is exactly the shape an answer needs.
- Authorship is often not weighed. A summary that names five tools rarely tells the user that one of the sources was written by the tool ranked first.
- Volume compounds. When several brands each publish lists, the same names appear across many pages, and assistants read repetition as consensus.
Why it is fragile
Google’s guidance on reviews has for years asked review content to show evidence of first-hand testing, explain what sets products apart, cover benefits and drawbacks, and help readers decide. A list where the publisher ranks itself first, with thin notes on everyone else, is close to the opposite. Google’s May 2026 guide to optimising for generative AI features also says that seeking inauthentic mentions is not helpful. And analyses of the May 2026 core update, such as Aleyda Solis’s, pointed to source type and intent as big drivers, with derivative content layers losing ground. None of that names self-serving lists specifically, but it describes the direction of travel.
On the AI side, discounting these pages does not need a breakthrough. The author of a list and the brand ranked first are both visible on the page. An assistant can learn to weight independent lists above vendor lists, or to state “according to the vendor’s own comparison”. Users asking for recommendations have every reason to prefer the former, and so do the platforms whose credibility depends on them.
The reputational risk is already here
Even if the rankings hold, the pattern is easy to spot and easy to write about. Buyers who click through to find the recommending source is the recommended vendor feel misled, and some will say so publicly. Journalists and community moderators have covered brands gaming AI recommendations. A tactic that turns into a story about your brand manipulating assistants costs more than the citations were worth.
What a durable comparison page looks like
Self-serving listicle Transparent comparison Titled as an independent “best of” list Titled as the vendor’s comparison (“How [brand] compares with X and Y”) Publisher ranked first by default Criteria stated up front; outcome follows the criteria Competitors described in one generic line each Competitors described fairly, with the cases where they are the better fit No sign of testing Evidence: what was tested, when, and on which plan No date or update history Dated, updated when products change A transparent comparison page earns a different kind of trust. It can still rank for “[brand] vs [competitor]” and “[competitor] alternatives”, and it gives assistants accurate facts about you to work with. It will not dominate “best X” prompts the way a disguised list sometimes does today. That is the trade: less short-term visibility, much lower risk.
If you already have self-serving lists live
- Rewrite the framing. Make authorship obvious in the title and introduction.
- Add genuine criteria and honest verdicts, including where competitors win. If you cannot find any case where a competitor is better, the page is not credible.
- Remove fake neutrality signals, such as invented award badges or “editor’s choice” labels.
- Watch what happens to citations. Track the prompts the list was being cited for. If citations drop, you have learned how much of your AI visibility rested on the disguise.
- Invest in the sources that cannot be disguised: independent reviews, analyst coverage, customer reviews and community reputation.
Hypothetical example:
A project management vendor publishes “The 10 best project management tools for agencies” and ranks itself first. Assistants cite it for several months. A competitor’s customers start sharing screenshots of the list alongside the vendor’s name in a community thread, and the thread begins to rank for the vendor’s brand. The vendor rewrites the page as “How [brand] compares for agencies”, with criteria, honest verdicts and a note on which competitors suit larger teams better. It loses some citations for generic “best tools” prompts but keeps them for comparison prompts, and the community thread stops growing.
Measure how much of your AI visibility rests on your own pages
Before deciding what to change, find out how exposed you are. For each prompt in your tracking set where an assistant recommends you, record the sources it cites or links. Then sort the prompts into three groups: those where independent sources support your inclusion, those where your own pages are among several sources, and those where your own list is the only cited source that puts you there. The third group is the visibility that disappears if assistants start discounting vendor-written lists. If it is a small share, the rewrite costs little. If it is most of your AI presence, you have a concentration risk that leadership should know about, and the priority is building independent coverage before the rewrite rather than after it.
How to judge durability for your own category
Run the same set of “best X” prompts monthly across the main assistants, and record which sources are cited and whether they are vendor-written or independent. If the share of vendor-written sources in citations starts falling, the window is closing in your category. If it stays high, the tactic is still working there, but the risks above do not go away. My recommendation is to build the transparent version now, while you can still compare the two, rather than scrambling after a policy change or a bad story forces the decision.
Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.