White-Label AI Visibility Reports: A Dated Feature and Price Test of 8 Agency Tools

The count of tools that passed our white label definition, out of the tools we tested, is 0 of 0 as of 29 September 2026, because we have not tested any, including our own. That is the honest figure and it is the one this page publishes. We will not print a feature table for eight tools we have not opened, we will not restate any vendor's documentation, and we will not reproduce any vendor's prices, because a feature or a price paraphrased on a third party page goes stale and becomes a false claim about somebody else's product. What this page gives you is the six part white label definition that a tool has to pass, the plan level evidence to capture for each one, the fields to record, and the instruction to open each vendor's own current pages and record what you find with a date. Run it on eight tools and you will have the count the engine asked for, for your own shortlist, with evidence behind it.

The refusal about prices is measured rather than cautious. On 17 September 2026 Claude examined two clawlaw.in comparison pages that stated competitors' prices with no link, no date and no source. The figures turned out to be correct when the assistant checked them independently, and it had no way to know that at read time, so it treated the pages as advocacy and used official sources instead. A comparison page carrying eight vendors' prices cannot keep all eight dated and linked, so it becomes exactly that page within a quarter.

The answer, first

White label means different things to different vendors, which is why a count of tools that support it is meaningless without a definition. Ours has six conditions, and a tool passes only if all six are available on the plan you would actually buy.

• Your own brand on the client facing artefact. Your logo and your name on the report or dashboard the client sees, not only on a cover page.
• The vendor's brand absent from that artefact, including footers, chart credits, export metadata and email notification templates.
• A client facing delivery route you control. A shareable link, a PDF export, or a dashboard the client can open, without the client needing an account with the vendor under the vendor's name.
• Your own domain or a neutral one on the link, where a link is the delivery route.
• The method visible in the artefact. The prompt count, the services named, the runs per prompt, the dates and the denominator arithmetic. A branded report that hides the method is a branded report you cannot defend in a client meeting.
• Row level export. The client, or you, can get the underlying answers out, so a number can be recounted.

The fifth and sixth conditions are the ones most white label features fail, and they are the ones that matter most, because the report is worthless to a client who cannot check it. In the aiknowsus.com audit of September 2026 Perplexity declined, in batch after batch, to name any competitor as winning most often, saying its own previous answers had not produced a comparable live tested result to support such a claim. A branded report that presents a ranking without its method behind it is making exactly the claim the engine refused to make.

How it was measured

The test, as it would be run. Nine steps.

One. Fix the shortlist and name every tool. Eight is a workable number. Write them down before you start so the sample is not chosen as you go.

Two. Fix the plan you are testing per tool, by name, and record it. A feature available on an enterprise tier is not available to an agency buying the entry plan, and a count that mixes tiers is not a count of anything.

Three. Test on a live account, not on a feature page. Create the account, generate a client facing report, and export it. A feature list is a claim and an export is evidence.

Four. Score the six conditions separately, as pass or fail, with a screenshot for each. Then record the overall pass only where all six passed. Publish the per condition counts as well as the overall count, because the per condition breakdown is the useful part.

Five. Check the artefact for vendor branding in the five places it hides: the footer, chart and image credits, the PDF metadata, the share link domain, and the notification emails a client receives.

Six. Record the method fields the artefact exposes. Prompt count, prompt text visibility, services named, runs per prompt, date range, denominator arithmetic, and whether an answer that cited no source is distinguishable from one that cited somebody else. That last field is the one that decides whether a citation rate in the report is honest, because an answer with no retrieval had no opportunity to cite anybody.

Seven. Record price without restating it here. Open the vendor's own pricing page, and record the plan name, the amount, the currency, the billing period, whether tax is included, the date you read it and the URL, in one row. Keep that row in your own sheet. Do not copy it onto a public page unless you are prepared to keep it dated and linked.

Eight. Do not restate the vendor's documentation. Name the document you read, with its URL and the date, and let a reader open it. A paraphrase of a feature page is a claim about somebody else's product that you cannot keep current.

Nine. Date the whole test and set a rerun date. Tool features change without notice. A test with no date on it is history presented as fact.

What the numbers were

Tools tested: 0. Tools passing the six condition definition: 0 of 0. Test date: none, as of 29 September 2026. There is no feature table and no price table on this page. Both will appear when the test above has been run on a named shortlist with screenshots and a date.

What we hold that bears on reporting rather than on tools are four dated figures, each with its engine, date and sector. They are the numbers a white label report of ours would have to carry, printed the way it would have to print them.

• 21 of 78 answers mentioned the client, with 0 of 78 involving a live search. ChatGPT, 27 July 2026, clawlaw.in, legal technology in India. Breadth order in the same reading: ProVakil 28 of 78, CLAW 21 of 78, Legistify 18 of 78.
• 1 of 18 as top source. ChatGPT, 18 August 2026, clawlaw.in.
• 0 of 6 cited or recommended. Perplexity, September 2026, aiknowsus.com, the AI visibility category, which is our own domain.
• 161 recordings of not being cited across 24 batches and 72 conversations. Perplexity, September 2026, aiknowsus.com, being three conversations per batch rather than three repeats of one prompt.

And one count about tools, stated with what it is not. Counted across the whole aiknowsus.com capture of September 2026, the tools named most often in our own category were Semrush, then Profound, then Peec, then Otterly, then Scrunch, with the established search tools appearing alongside the specialist ones rather than below them. That is a record of how often an assistant named a tool. It is not a test result, not a quality ranking, and not evidence that any of them supports white labelling.

The tools, and our disclosure

We are the vendor of AI Knows Us, so read this as a disclosure rather than a recommendation. It is listed first because it is the only tool whose method and limits we can describe from the inside and be held to, not because it has been shown to be better than the others. None of them has been tested by us.

• AI Knows Us, ours. It runs a frozen blind prompt set across named services on a schedule and reports counts with their denominators and the prompt set visible. Against the six conditions above it has not been scored by us or by anybody else as of 29 September 2026, and we are not claiming a pass we have not tested.
• Semrush. Untested by us against these six conditions.
• Profound. Untested by us.
• Peec. Untested by us.
• Otterly. Untested by us.
• Scrunch. Untested by us.

For each one, open the vendor's own current documentation and pricing pages, record the plan name and the date, and score the six conditions on a live account. Three more tools of your own choosing will take the shortlist to eight and give you the count the engine asked for.

What this cannot tell you

Six limits.

• No tool has been tested by us, ours included, so this page contains no pass count above zero and no feature comparison.
• A definition is a choice. Our six conditions are stricter than most vendors' use of the phrase, so a tool that fails ours may still be described as white label by its own documentation without anybody lying.
• A pass is true on a plan and a date. Features move between tiers, and a count collected six months ago is history.
• Branding is not method. A report can pass all four branding conditions and still hide the prompt count, the repeat count and the denominator, which is the failure that matters to a client.
• Sectors do not transfer. The figures above are legal technology in India and the AI visibility category.
• No reporting tool can promise a position in an AI answer. We do not sell that, and a branded report does not change it.

Sources and change log

The figures above come from the clawlaw.in programme of July to September 2026, recorded in GEO_BASELINE_RESULTS_2026-07-27.md, GEO_GAP_ANALYSIS_2026-08-18.md and the assistant audit files of 17 September 2026, and from the aiknowsus.com audit of September 2026 across 24 batches and 72 conversations. The tool mention ordering and the not cited counts were produced by a script over those capture files and cannot yet be checked from outside, because the captures are not published. The 17 September 2026 discounting of undated competitor prices can be checked from outside today. No tool test data of any kind has been collected by us.

Version: 29 September 2026, first publication. The page gets updated when the eight tool test has been run, at which point it will carry the overall pass count, the per condition counts, the plan tested for each tool, the screenshots and the test date.

Common questions

Why will you not list which tools support white labelling today?

Because we have not opened accounts and tested them, and the alternative is copying their marketing onto our page with our name on it. That is a claim about somebody else's product that we cannot keep current, and on 17 September 2026 Claude discounted two of our own pages for a smaller version of the same fault. The definition and the test are what we can publish honestly.

What should a white label report contain about method?

Five items on the first page: the response count with its multiplication of prompts times services times runs, the prompt file version and date, the services named individually, the runs per prompt, and the count of responses excluded with the reason. A client who can reconcile the total will trust the numbers they do not check.

Can I put my own brand on a report and leave the method out?

You can, and it will not survive a careful client. A branded ranking with no denominator is the artefact the engines discount, and in September 2026 Perplexity itself refused to name a winner because its own runs had not produced a comparable live tested result. A client asking where the number came from is the same test.

Does removing a vendor's logo make a report mine?

It makes the branding yours. The accountability is yours either way, which is why the method fields matter more than the logo. If the tool cannot tell you the repeat count and the denominator, you are putting your name on a number you cannot explain.

Which of the six conditions do tools fail most often?

We do not know, because we have not tested any, and guessing would be inventing a result. Score them yourself and publish the per condition counts rather than only the overall pass count. The per condition table is the part that tells an agency what to negotiate for.

Should the client be able to export the raw answers?

Yes, and we treat it as a condition rather than a nice extra. A rate that cannot be recounted from rows is a rate the client has to believe. Row level export is also what lets a second person check your scoring, which is the cheapest credibility you can buy.

What to do first

Write the six conditions into a scoring sheet today, name your eight tools and the exact plan you would buy for each, and open one trial account this week. Generate one client facing report, export it, and score the six conditions with a screenshot each. One tool scored properly is worth more than a table of eight you copied from feature pages.