Run Each Prompt Multiple Times Generation involves randomness and retrieval can return different pages between runs, so a single answer is a sample. Three runs per prompt is the practical minimum and five is better where the stakes are high.

Write it once, covering the category question, the problem question, the comparison question, the competitor question and the branded question. Fifty is a workable minimum. Then freeze it, and if you must add prompts later, add them as a separate cohort so the original series stays comparable.

So attribute it by name every time it appears in a report. A visibility figure presented without saying which tool produced it and how it was sampled will eventually be quoted back at you as fact by somebody who did not know it was an estimate, and that is a difficult correction to make in front of a board. get recommended by ai

In most categories the result is the same shape: a review platform, an industry directory, one or two forum threads, a comparison article, occasionally a trade publication, and only then anybody's own website. The competitor is winning on pages neither of you owns.

Prioritise by your own citation data rather than by prestige. A trade directory nobody has heard of that appears in half your category's answers is worth more attention than a well known publication that never gets cited. get recommended by ai

Where a roundup includes you with errors, a factual correction with evidence has a high acceptance rate. Publishers generally do not want to be wrong, and this is the single highest return outreach available in this discipline.

On Third Party Tracking Tools Several tools now offer to monitor this at scale, and they save real time once your prompt set runs into the hundreds. They are worth buying for trend lines and for coverage you cannot manually sustain.

Build the Task List From What You Found The teardown usually produces four workstreams, in this order of cost: fix any access problem, claim and correct your listings on the recurring sources, rewrite your equivalent pages to lead with specifics, and start the slower work of earning coverage on the sources where you cannot simply claim a profile.

Where It Diverges Sharply Traditional SEO optimises for a ranked list. Generative systems optimise for a synthesised answer, and the sources they pull from are not the same set. Ahrefs studied 15,000 long-tail prompts across four assistants in July 2025 and found that around 80 percent of the pages cited did not rank anywhere for the original query, with only about 12 percent appearing in the top ten. Ranking first does not reserve you a seat.

The output is a spreadsheet and it is the most important document in the project. It tells you whether you are named, whether what is said about you is true, who is named instead, and which pages your category's answers are actually built from.

This is also why review volume and recency show up so consistently in what gets cited. A platform with forty recent accounts of working with you is more informative than your own page saying customers love you, and it is treated accordingly.

Rule Out the Mechanical Explanations Before concluding the gap is editorial, confirm you are readable. Check robots.txt for the relevant crawlers, check your server logs for what those agents actually receive, and load your key pages with scripts disabled to see what survives.

Report frequency rather than presence. Being named in one run out of five is a genuinely different situation from being named in five out of five, and a report that collapses both to mentioned has thrown away the useful part.

Ask ChatGPT, Perplexity or Gemini to recommend a supplier in your category and you will get a short list. Three names, maybe five. Your customers are already asking those questions, and the answer they receive does not come from a page of ten blue links they can scroll past. It comes as a recommendation, delivered with confidence, and most people act on it without checking a second source.

The pages that earn citations are consistent across industries: an honest comparison of the options including where you are not the right choice, a plain definition page for the thing you sell, a specifications page with real numbers, and a pricing page that says something concrete.

Being missing from the five pages that generate your category's answers is a complete explanation on its own, and it is fixable without anyone's permission on the platforms that accept claims and corrections.

Publish Your Own Comparison Anyway It will rarely be the most cited source in your category and it is still worth having, for two reasons. It puts a version of your figures into circulation stated correctly, and it is frequently the page journalists and roundup writers use when compiling their own comparisons.

Fix the Prompt Set and Never Casually Change It Your prompt set is the instrument. If you adjust it between runs you are measuring your own edits, and any trend line you draw afterwards is meaningless.