TrustRadius vendor profile: what is publicly documented about listing, reviews and pricing, and what we will not claim

There is no number we can responsibly put on this page. TrustRadius sets its own process for vendor profiles, reviews and pricing display, that process is documented on its own pages, and we have not made a dated check of those pages. So this page states no processing time, no criteria count, no verification threshold and no price. Inventing one would be worse than useless, because it would look exactly like a measurement. What this page does is name the documents to read, give you the record format that makes your own check reusable, set out how to tell a documented rule from a practice you have merely been told about, and print the dated evidence we do hold about how review platforms actually fared in AI answers in one sector, which is legal research software sold in India.

The answer, first

Four statements, and the first one is the whole page.

No figure on this subject is safe to publish without a dated check, so we publish none. Most pages about vendor profiles on review platforms carry at least one confident number: a review count threshold, a turnaround time, a price. We hold none of those as a measurement, and a number we did not measure would be indistinguishable on the page from one we did.

What is checkable today is the documentation, by you. The four documents to read are the vendor or "for vendors" pages, the help material on claiming and completing a profile, the review guidelines including anything about verification of reviewers, and the terms that apply to vendors. Read the current versions, record URL and date, and save copies.

What is not checkable is anything about the internal process. How long a review takes to be published, how a reviewer is verified, what weight anything carries: unless the platform states it, you are being told a practice rather than shown a rule. Log it as a claim with the date and the name of who told you.

A profile on a review platform is worth having and is not a visibility strategy on its own. The sector evidence below is narrow but it is dated and it points one way.

How it was measured

Two measurements are available to you without any special access, and both are worth more than a borrowed statistic.

The documentation check, in six steps.

• Fix the source set. The four documents above and nothing else. Anything from a sales conversation, a webinar or a support chat goes in a separate unofficial list.
• Record the version. URL, read date and time, any last updated stamp, and a saved copy.
• Classify every statement you rely on into one of three buckets: a stated rule, a stated intention, or silence. Silence is the most common and the most misreported, because an article will fill it with a number.
• For each stated rule, write what evidence would satisfy it and whether you can test that before submitting.
• Two readers, then publish the disagreement count. Same as any documentation audit: without it you have one person's reading.
• Recheck quarterly and keep the old copies, and publish what changed. The change history is the part no other page has.

The visibility check, in five steps. This is the measurement that tells you whether the profile is doing anything for you in AI answers, and it has to be run blind.

• Freeze a prompt set of 20 to 100 buyer questions in a file, with a version number and a date, and do not edit the file during the study. Every fraction you publish names the version.
• Apply the blind rule: no prompt contains your brand name. The reason is measured. On 27 July 2026 the same 78 questions were run twice on the same day for clawlaw.in. The run whose wrapper named the brand returned the company first on almost every question. The blind run put it second by breadth and absent altogether from the litigation due diligence questions it most wanted to win. The branded run was discarded, because the only thing it had measured was our own prompt.
• Repeat each prompt three times per assistant, on the same day, and record all three answers rather than the best one. Answers to the same words differ within a day, and a single ask cannot show you that.
• Print the denominator arithmetic. 30 prompts times 3 repetitions times 3 assistants is 270 answers, so a count of 12 answers naming you is 12 of 270 on the named dates, not 4 per cent of the market. Publish the multiplication on the page.
• Log ten fields per answer: prompt text, prompt set version, assistant and version if shown, timestamp, country and language setting, signed in or out, whether your brand was named, whether a URL on your domain was cited, whether the review platform's page was cited, and the full answer text saved.

And record any search count as a claim, never as a fact. Where an assistant tells you it searched, save the statement and mark it unverified. In the aiknowsus.com audit of September 2026, Perplexity withdrew its own earlier statement when asked how many of the questions it had actually searched for, saying it could not honestly substantiate the claim that it had run a live search for each one, and in another batch that its claim to have searched all five was not adequately supported. Three separate batches produced that retraction. If a system's account of its own work can be withdrawn, it is evidence about what it said and not about what it did.

What the numbers were

Dated checks of this platform's current documented process completed by us: zero, as of 29 September 2026. Therefore: no number on this page about TrustRadius. Not a turnaround, not a threshold, not a price, not a tier.

What we hold instead, printed with denominators and sectors.

• 0 of 6 questions where any third party review source reached an answer. Claude, 17 September 2026, clawlaw.in, legal research software in India. One well known review site appeared in the raw results and was discarded, because the list it offered was of American products and so was not an answer to an India question.
• 1 of 18. Across eighteen blind commercial questions on 18 August 2026, ChatGPT made clawlaw.in the top source on exactly one of them, and named the company's own comparison page in the answer as still being the vendor's own editorial page.
• 161 recordings across 24 batches and 72 conversations. On aiknowsus.com in September 2026, the phrase recording that we were not cited appears 161 times in the engines' own self audits. Perplexity.
• 0 of 6 on our own domain. Asked six questions about its own category with no brand named, Perplexity audited itself afterwards in September 2026 and reported that it had not cited or recommended aiknowsus.com in any of the six answers.
• The citation that did land, with a date. On 6 August 2026 ChatGPT cited clawlaw.in/blog/how-to-check-a-companys-court-cases-in-india as a source for a vendor due diligence question, fourteen days after that page was published, in a zone where the same question set had named the company nowhere at baseline. That one is checkable from outside.

The last line matters on a page about review platforms, because the thing that got cited was not a profile on somebody else's site. It was a page on the company's own domain answering one buyer question directly. One citation is one citation and not a trend, and it is the only dated first citation we hold.

What this cannot tell you

• It cannot tell you this platform's process. That is the point of the page. Read the documents, date your read, and you will know more about it than any article can tell you.
• Our sector evidence is one company in legal technology in India, measured in July to September 2026. A software category in another market could behave completely differently, and a number from here must not be moved there.
• Absence from an answer is not proof a source has no value. A profile can influence a human buyer who never asks an assistant anything.
• One dated first citation is not a rate. The 6 August 2026 citation is one observation, and fourteen days is how long that one took, not how long anything takes.
• A blind test measures your prompt set on your dates. It does not measure the market.
• Nothing here promises a position in an AI answer, and we would not trust a page that did.

Sources and change log

Sources. The two protocols are ours. Every figure comes from one of two recorded bodies of work: the clawlaw.in programme from July 2026, in Tier_1/GEO_BASELINE_RESULTS_2026-07-27.md, Tier_1/GEO_GAP_ANALYSIS_2026-08-18.md and Tier_1/claude_response_17_09_audit.md, and the aiknowsus.com audit of September 2026 across 24 batches and 72 conversations, counted by a script over the capture files. Of the figures above, the 6 August 2026 citation of a named URL can be checked by anybody from outside. The counts over our capture folders cannot, until the captures are published. No sentence on this page paraphrases TrustRadius documentation, and no price appears on it.

Change log. 29 September 2026, first publication, with the explicit statement that no responsible number can be named yet. When the documentation check is run, this section will carry the four documents, their URLs, the read dates, the count of stated rules against stated intentions, and the disagreement count between two readers. Every correction will be noted here rather than made quietly.

Common questions

Is a page with no numbers in it worth publishing?

We think so, when the alternative is a number nobody measured. What this page gives a reader is a way to get the number themselves, a record format that survives, and five dated results from a real programme. In our own audit, the answers that survived scrutiny were the ones that stated their limits, and the pages that stated prices with no source were treated as advocacy even when the prices were right.

How many reviews do we need before a profile is useful?

We have not measured that and we will not name a threshold. What you can do is record your own count with a date, monthly, and watch the curve. Your own series over six months is worth more than anybody's benchmark, because it is about your product and your buyers.

Can we ask our customers for reviews?

Every platform has its own rules on soliciting, incentives and wording, and those rules are where profiles get penalised. Read the review guidelines as a separate document from the listing guidelines, record the wording you relied on with its date, and keep the saved copy. We do not restate those rules here because a stale paraphrase would be a false claim about somebody else's policy.

If review sites did not reach the answers you measured, why bother?

Because 0 of 6 is six questions on one day in one sector, not a law, and because buyers read review sites whether or not an assistant does. The honest position is that a profile is a buyer facing asset with an unmeasured effect on AI answers, and that you should measure that effect yourself, blind, before and after.

What would change this page?

A dated documentation check, which would fill the empty cells with stated rules rather than guesses. Or a blind prompt study in which a review platform page appeared as a cited source, which would give us a fraction with a denominator to print here. Either one gets added with its date, above this line.

What to do first

Open the four documents today and save copies, with the URL and read date written at the top of your sheet. Then write down twenty buyer questions with your brand name in none of them, ask three assistants each question three times, and record the ten fields per answer. By the end of the week you will hold both halves of what this page refuses to invent: what the platform states, and what the assistants actually did.