AI Search Tool Rank
All posts
By AI Search Tool Rank Teamtools

Ahrefs AI Search Visibility Topics and ChatGPT Mentions

Brand Radar topics are a modeled index. Promptwatch topics are tags on prompts you typed. We rank Promptwatch first for ChatGPT mention ops.

Ahrefs AI search visibility topics and ChatGPT mentions share a brand name and almost nothing else. Ahrefs Brand Radar clusters 260 to 405 million keyword-modeled prompts into topics. That is an index you browse. Promptwatch topics are tags you put on prompts you actually typed. We rank Promptwatch first when the job is mention operations: did ChatGPT name you this week, on this wording, and which URL did it cite. Keep Brand Radar for index research if Ahrefs is already paid.

The mix-up is easy. Both screens say "topics." Both mention ChatGPT. One is a modeled corpus refresh that runs on roughly a monthly cadence. The other is a daily UI check of a list you own. If you treat the index as the ops log, you will wait a month to learn a page change did nothing, or you will invent a mention rate from a cluster that never used your sales language.

Two objects, one invoice

Brand Radar is $199/mo per index on top of an Ahrefs plan from $129/mo. Chatbot indexes refresh about monthly. An independent January 2026 test found 3 ChatGPT mentions in Brand Radar where 123 were verified by hand. There is no Claude in the native indexes. Custom prompts are $50/mo per 2,500 checks.

That last line is the tell. Custom prompts are a metered add-on sitting on a modeled pipeline. They are not the same thing as loading fifteen buyer questions and reading yesterday's answer.

Promptwatch topics and tags sit on those typed prompts. You can also attach personas and country, state, or city. Paid plans run daily ChatGPT checks from the real UI. Explore is free (10 prompts, ChatGPT only). Essential is $95/mo. Citation analytics splits page, domain, Reddit, and YouTube. Agent Analytics logs crawler fetches. Visitor analytics attaches AI-referred sessions via a script or GTM. Professional is $245/mo. Agency Kick-off is $199/mo. 4.7/5 on G2, 1,840+ brands. Product: promptwatch.com. Configured prices sit on the rankings.

Semrush AI Toolkit is the other modeled suite add-on at $99/mo per domain. Same category as Brand Radar: useful if you already live in that suite, not a replacement for a typed prompt ledger.

Allow OAI-SearchBot if ChatGPT Search is in scope. GPTBot is training. Those are different robots.txt lines. Brand Radar will not tell you the CDN returned 403.

What "topics" means in each product

In Brand Radar a topic is a cluster over a huge modeled set. You use it to see which subject areas show up in chatbot indexes and which domains the model of the index associates with those areas. That is research. It is closer to "what exists in this modeled world" than "what happened on our fifteen prompts on Tuesday."

In Promptwatch a topic is a label. Campaign, product line, persona, country. You typed the prompt. You tagged it so the weekly table can filter. If ChatGPT names you but cites a listicle, the citation field is the ticket, not the topic name.

Do not export Brand Radar topics and paste them in as Promptwatch prompts. Sales language is shorter and weirder than keyword clusters. The January undercount is what you get when a modeled index and a live ChatGPT session disagree. We treat the live session as the ops number.

When Ahrefs still belongs in the stack

Canceling Ahrefs to "go GEO" is a bad trade. Links, crawl, and keyword work stay in Ahrefs. If ChatGPT cites a URL Ahrefs already tracks, use Ahrefs to understand that domain. Promptwatch only has to prove the model still uses it after you edit.

If Brand Radar is already on the invoice, pull it monthly and label the slide "modeled index." Do not chart it against Promptwatch daily rows as if the timestamps match.

FAQ

Are Promptwatch topics the same as Brand Radar topics?

No. Labels on prompts you typed versus an index of modeled prompts. Same English word, different object.

Cancel Ahrefs?

No. Links stay in Ahrefs. Mentions go to Promptwatch.

Does the January 2026 test mean Brand Radar is useless?

It means Brand Radar undercounted ChatGPT mentions in that test (3 versus 123 verified). Use it as research, not as the mention ledger you brief writers from.

What to do this week

  1. Write 15 ChatGPT prompts from sales, not from an Ahrefs topic export.
  2. Tag them in Promptwatch by campaign or persona.
  3. If you open Brand Radar, label the export modeled.
  4. Use Ahrefs on the cited URL that beat you.
  5. Check configured prices on the rankings.