Last week we added outside proof to our own profile on our comparison page and gave our homepage a title that says what we do. In week 7 of our UK B2B Tech PR AI Visibility Index, Perplexity named us in 35 answers, up from 14. ChatGPT read the page more than twice as often and named us no more often than before.
Week 6 showed us that ChatGPT used our B2B tech PR page as a guide to the market and shortlisted other agencies from it. The page describes every agency fairly, which gave an assistant little reason to pick the one that wrote it. We made three changes.
Perplexity cited the comparison page in 39 answers, up from 23, and our homepage in 13, which it hadn't cited at all the week before. It named us in 35 answers and moved us from 12th to 7th in its own ranking. Its answers to our check questions held, which points to what it read as the cause and away from a change in Perplexity itself.
We can't say which of the two changes did more, because both went live in the same week.
ChatGPT cited our page in 40 answers, up from 18, and no source was cited more often by ChatGPT this week. It named us in 28 answers, against 27 a week ago, and moved us from 8th to 10th in its own ranking as it named other agencies more often. Its check answers held too.
ChatGPT seems to treat the page as a list to choose from. Stronger proof on our own profile hasn't changed which agencies it picks, at least in one week. The number of times our name appeared only as a label on a source, which we don't count as a mention, rose from 34 to 42 across the six assistants.
Gemini cited the comparison page again, in 11 answers after none in week 6, and named us in 16 answers, up from 10. It still cited Archetype's page, in 25 answers against 32. Gemini's check answers changed this week, so part of that movement comes from Gemini itself. It didn't cite our homepage at all, which means the new title hasn't changed what Gemini reads yet.
We stayed 9th overall, on 3.8% of the weighted share of mentions, while Hotwire held first by a tenth of a point over Clarity. The comparison page was the most-cited page in the study, in 164 answers, up from 108. Answers citing our site rose from 183 to 232, and once we leave those answers out we fell from 14th to 18th.
That gap is the lesson of the week. The proof we added is our own account of other people's verdicts, and the assistants read it on our site. The next step is getting those verdicts into the trade press, review sites and third-party lists the assistants read without going through us.
This was also Grok's first week on grok-4.3, a newer version of xAI's model. It named us in 56 answers, down from 60, and its check answers held, so the switch doesn't appear to have changed much.
The full ranking and every week so far are on the index page. To see which pages the assistants read and recommend from in your own category, ask us for a free AI visibility snapshot.