Peec AI and Rankscale are the only two vendors here whose own documentation names a branded-versus-unbranded split; Profound and Semrush's AI Toolkit group prompts but don't label it that way. A branded question (does ChatGPT know us) and an unbranded one (who does it name for the category) measure two different things.
This question sits inside a bigger one: before paying for any AI visibility tool, know exactly what it counts and what it leaves out. A tool that blends branded and unbranded answers into one score is hiding the more useful of the two numbers.
A branded question tests whether AI already knows who you are; an unbranded one tests whether it would pick you without being asked for by name, which is the harder and more commercially useful test. Ask ChatGPT what a named product costs and it almost always has something to say about it, correct or not, because the brand is in the prompt. Ask it for the best option for a given budget and team size, with nobody named, and the answer reveals who the model actually reaches for.
Proof: Jasper's framing states that branded prompts name the company being measured, so the brand will almost always appear, while unbranded prompts measure whether a buyer can find the brand at all (Jasper — branded vs non-branded prompts). Rankscale makes the same case differently: branded visibility is the floor, non-branded visibility the ceiling, because that is where new pipeline comes from (Rankscale — branded vs non-branded prompt groups, read November 2026).
Why it matters: a tool that reports one blended number can make a brand that only wins when asked for by name look healthier than it is.
Of the vendors with a documented prompt-grouping feature, only two name "branded" and "unbranded" as a label a user can see. The rest let you group or filter prompts in overlapping ways without using those words, as read in November 2026.
| Tool | Names the branded/unbranded split? | What it actually does |
|---|---|---|
| Peec AI | Yes | Automatically tags prompts by intent, so a user can see which are branded, unbranded or commercial |
| Rankscale | Yes | A documented module for setting up prompt groups, framed explicitly as branded vs non-branded |
| Profound | Not documented | Groups prompts by topic, tag, persona or region; suggests "broad coverage" and "strategic depth" prompt sets |
| Semrush AI Toolkit | Not documented | Reports a mention share across what its knowledge base calls "non-branded industry queries", with no branded/unbranded toggle found in the toolkit itself |
Read this as a snapshot: every one of these vendors ships features monthly. Check the vendor's own documentation before you buy on this feature alone, the way this cluster's full comparison asks you to check every claim here.
Peec AI separates two things most dashboards blend: whether a prompt is branded, unbranded or commercial, and whether your brand is named in the answer at all versus just used as a source. Its own visibility-tracker page describes the first as automatic, done on the prompt itself rather than left for a user to label by hand.
Proof: Peec AI — AI visibility tracker, read November 2026.
Rankscale's documentation treats branded visibility as the baseline a brand should already clear, and non-branded visibility as the harder number that actually grows the pipeline. A brand that fails its own branded prompts has a more basic problem than one that merely scores low on unbranded ones.
Proof: Rankscale — setting up prompt groups: branded vs non-branded, read November 2026.
Do this: if you only run one set first, run the unbranded one. It is the one competitors can't buy their way out of by being named correctly.
Not as a named label, as far as their own documentation shows, though both let you approximate it. Profound's prompt tracker groups prompts by topic, tag, persona or region, and suggests a "broad coverage" set for the category alongside a "strategic depth" set for your own brand, a different split that can be made to do similar work by hand. Semrush's AI Toolkit measures a mention share spanning what it calls "non-branded industry queries", but no page we found documents a branded/unbranded filter a user can switch on. Check each vendor's current documentation before counting on either.
Honest read: "lets you group prompts" is not the same claim as "labels branded vs unbranded for you." Ask the vendor directly which one it does before you buy.
Start from language buyers already typed, not from questions invented at a desk. Reflexa's own stance since its 6 October 2026 rebuild is to run no prompt set at all, because a question nobody asked tells you about the question, not about your business, so this is a build-it-yourself job, not a feature to buy from us.
Bottom line: the split is a method, not a product feature. Run it by hand in an afternoon, with nothing paid.
None of this requires buying a tool first:
Peec AI and Rankscale are the two vendors whose own documentation names the split outright. Profound and Semrush's AI Toolkit let you group prompts in overlapping ways but don't document a branded/unbranded label.
A branded question names you or your product directly and mainly tests recall. An unbranded question describes the buyer's situation with nobody named, and tests whether you'd get picked without being asked for.
No. Reflexa stopped running any prompt set on 6 October 2026, when it retired the Recognition, Competitors and Sources screens that asked engines invented questions, and checks the structural reasons a page goes unread instead.
Yes. Pull the real phrases buyers type from Search Console and your own site search, split them by whether they name you, then ask ChatGPT, Claude and Perplexity both sets with web search on.
Enough to cover the handful of ways buyers phrase it, across three engines, repeated over a few weeks rather than one run. Keep the branded set short; the unbranded set needs more variants, since category, budget and use case each change the answer.
The free check reads your site the way ChatGPT, Claude and Perplexity read it, no card, about three minutes, evidence included.
Reflexa Agent, the product: Reflexa Agent vs the AI visibility tools · pricing · priser på svenska