Blog · 24 September 2026 · By HeardOf

Can you see the answer, and who got named instead? What nineteen AI visibility tools' pages say

Can I see what ChatGPT actually said about my brand?

Sixteen of nineteen AI visibility tools say so on their own pages, read on 24 September 2026: fifteen in so many words — full, exact, actual, complete, unabridged or raw response — and one, SE Ranking, as responses you can review. Three say nothing about it on the pages we read: Evertune, Goodie and Bluefish.

The question a buyer types is "see what ChatGPT says about my brand", and a visibility score is one step removed from it: a rate, over prompts, of answers that named you. This post reads the same nineteen vendors as what nineteen of them say about how they get their answers, on the same day and the same page sets, and asks three things of each: does the product show the full answer text, does it show brands you did not declare, and does it show the URLs each answer cited. A negative means not stated on the pages we read, which are listed by URL, time and status in the research file; the rule for what counts is at the end, and the table carries a quote for every cell.

Four of the fifteen, in their own words. Athena's docs: opening a row "displays the full, unabridged text of the AI's response, the exact sources cited, competitor details, and qualitative attributes." Conductor's docs list what opens with a response, including "The full text of the response." and "The citations in the response." Semrush's knowledge base: a "View full response" control on any prompt opens "a detailed view of the AI-generated answer." Ahrefs' methodology post, updated 26 February 2026 and so labelled historical: "We store the raw responses, and users can then search this corpus".

SE Ranking is counted apart because its page uses none of those words: its tracker page says "Review AI responses to discover how your brand is framed", and SE Visible's FAQ offers "snapshots of how your brand appears in AI answers, along with the context and sentiment of each mention". Whether it shows the whole text is not on the pages we read.

Will the tool tell me who got named instead of me?

Eleven of nineteen say their product lists or surfaces brands you did not add, as read on 24 September 2026. Five say they show which brands were mentioned, who was recommended instead, or a competitor set found automatically, without saying whether the names come from the answers or must be declared first; three say nothing on the pages we read; none says the opposite on the pages read.

The eleven say it in different shapes. Peec's docs: position is calculated "including all detected brands, not just your tracked competitors", and "We detect brands mentioned alongside yours and suggest them when they appear at least twice." Profound's metrics docs: "Any brand named in the answer counts, not only your configured competitors, unless you have excluded it under Configure Competitors in brand settings." and, of its Answers API: "Every brand Profound detected is listed, including any you have excluded under Configure Competitors." Conductor's docs describe a toggle between "Entire market" and "Tracked competitors" so you can "discover new competitors vying for the same visibility in AI search responses". Athena's docs describe a Discovered competitors tab of "AI-surfaced brands" that "aren't currently being monitored". Otterly's help centre, in an article dated 21 May 2026 and so labelled historical: the brand report's More Detected Brands view "shows all brands that have already appeared in AI answers for your tracked prompts — even ones you haven't added as competitors yet." Scrunch and Rankscale name a suggested or auto-identified competitor set; AmICited, Semrush, Writesonic and Similarweb list every brand in an answer, in Similarweb's words "in the order of recommendation". Each sentence is quoted in the table.

Otterly's brand-detection article, dated 1 December 2025 and historical, says that since 27 November 2025 "we look for a brand name match based on text" against the brand and competitor names you define. That describes how a name is matched, not which names are listed; the article of 21 May 2026 quoted above, historical as well, is the one that says the list includes names you did not add. Otterly is the vendor in HeardOf vs Otterly.

The five in between are Goodie, whose Brand Command page promises "a plain verdict: which claims are false, who's getting the recommendation instead of you, and which parts of your story you already own"; Surfer, whose competitors dashboard lets you "Identify emerging competitors that are gaining more mentions or higher positions than you"; BrightEdge, whose home page says AI Hyper Cube shows "which brands are mentioned or cited"; Evertune, whose measure page says "We surface, score and rank your competitors automatically"; and Ahrefs, whose help centre says "Brand Radar will automatically find your brand name, as well as commonly associated competitor brands and websites." None of those five sentences says whether the names come from the answers or must be declared first, so none is counted either way. Bluefish says nothing about it; HubSpot defines competitor visibility as "how often each tracked brand appears"; SE Ranking's tracker page says it "checks for linked and unlinked references to brands, products, businesses, or personal names" and takes "up to five competitors", which does not say whether a sixth is listed.

HeardOf lists no brand you did not name, per our post on which engines we check as read on 19 September 2026; the section below says what that post says and no more.

Can I see which URLs each answer cited?

Eleven of nineteen say the cited URLs or sources are shown with the answer they belong to, as read on 24 September 2026; seven describe citations only as domains or pages ranked across all answers; one, Bluefish, says nothing on the pages we read.

Three of the eleven. Peec's docs put a "Sources sidebar: All URLs the AI model referenced or cited when creating the response" on every chat and distinguish it from "Citations: Sources explicitly referenced within the AI response text". Otterly's response detail, in a help article dated 24 June 2026 and so historical, lists "Citation links used by the AI engine (when available)". AmICited's academy puts "the Sources cited (N) list" "to the right of each response, alongside the full answer text". The other eight are quoted in the table.

The seven are not saying they hide anything; their pages describe the aggregate and stop there. Evertune's FAQ: "Content Analytics reports every domain and URL AI models return as a source". Ahrefs' help centre: "identify top cited pages and domains". HubSpot: "which domains, pages, and content types are being referenced in AI answers". BrightEdge runs the other direction — enter a URL "and AI Hyper Cube shows every AI answer that cites it as a source" — which is a list of answers per URL, not of URLs per answer. Goodie, Surfer and SE Visible are quoted in the table.

What does each tool's own page say?

Fifteen of nineteen say in so many words that the answer text is shown and one more says responses can be reviewed; eleven say they list brands you did not add; eleven say the cited URLs are shown with each answer — counts re-derived from the cells below, read 24 September 2026.

Does this mean no free tool shows who came out instead?

It does not settle that, because these are the pages of products, not the output of a free checker. What a free checker returns for a fixed set of brands is a separate measurement we have not run, and this post makes no claim about it.

What the pages do settle is narrower: on 24 September 2026, eleven of nineteen products said on their own pages that they list names you did not add. On the vendors' own account, a buyer can get that from more than half of the tools we track, and, per our post on which engines we check as read on 19 September 2026, cannot get it from ours.

What does HeardOf show?

Per our post on which engines we check, as read on 19 September 2026, HeardOf scores you and up to three competitors that you name, and brands the engine mentioned that you did not name sit in the stored answer text and are not counted. For Google AI Overviews, the post says, as of 23 September 2026, that we keep the overview block's text and the links it references; it does not say whether either is shown.

It says the answers are stored, not whether their text is shown to you; it says scoring stores whether you were named or cited, not where, and nothing about showing the URLs an answer cited; on the second question it is a no, in its own words. That describes what we ran on 19 and 23 September 2026, from our own systems, and is not a roadmap statement.

What counts as a statement, and how was this read?

A statement counts when it is on the vendor's own domain — a product, feature or pricing page, docs, a help centre, a knowledge base, an API reference, a changelog, or a methodology post whose subject is the product — and says what the product shows or stores; a statement about an API, an MCP connector or an export counts, and the cell says which surface it is. Testimonials, figures inside mock-ups, and a sentence about a competitor's product do not count.

Full answer text is a yes in so many words when a page uses full, exact, actual, complete, unabridged or raw of the response, or names the response text as a thing the user sees; a yes as responses viewable when a page says you can view or review AI responses without those words. Brands you did not declare is a yes when a page names brands beyond the ones you add, and a yes on one page settles it; a middle value when a page says it shows which brands were mentioned, who was recommended instead, or a competitor set found automatically, without saying whether the names come from the answers or must be declared first. A page that describes matching the names you define is quoted, and is not counted against a page that lists names you did not add. Cited URLs per answer is a yes when a page says the sources a given response cited are shown with that response; aggregate only when the pages describe citations as ranked domains or pages across all answers, or from a URL to the answers citing it.

The read was a first pass on 24 September 2026 between 17:37 and 17:41 UTC with curl and a browser user-agent, over the same page sets as the methods post — 1,601 requests, 1,567 of them answered 200 — then 108 product, feature, pricing, docs and help pages rendered in a headless browser with JavaScript between 17:41 and 18:02 UTC, because a page can draw text a script never receives; rendering changed no cell, no rendered page drew a body its served HTML lacked, and one page, Otterly's home, could not be rendered on three attempts and was read as served. A second pass between 18:19 and 18:21 UTC added twelve requests — three HubSpot knowledge-base articles, two Otterly help articles re-fetched and seven Evertune platform pages; the cells changed in that round rest on pages the first pass had already read. Every request and every rendered page is listed with its URL, time and status in the research file. The same sites as before could not be read: SE Ranking's help centre, Similarweb's support centre and Athena's status page answered 403; Goodie's help subdomain and Rankscale's docs subdomain do not resolve; the six HubSpot knowledge-base addresses tried first answered 404, and the three articles found afterwards changed no cell. A page whose printed date is more than 90 days before the read is labelled historical. At least four vendors carry text addressed to AI readers — Peec's ai-instructions page, AmICited's file for language models, Otterly's footer link inviting AI to learn about it, and the banner on Writesonic's HTML docs pages — and all were read as data. Prices are in what nineteen AI-visibility tools charge and are not repeated here. The three questions this post asks are two of the eight in eight questions to ask an AI visibility tool before you buy, which puts them in a buyer's order and copies every count from our live posts.

Everything this post says about our own system comes from our own post on which engines we check, as updated 23 September 2026, and carries that provenance in the sentence that says it; you cannot check it. Everything else carries a vendor's page and the day it was read, and the research file carries the time.

Eight questions to ask an AI visibility tool before you buy, and what nineteen tools' own pages say to each