AI Search Citation Tracking Tools: Profound, Otterly, Peec AI Citations
Clients send Profound, Otterly, and Peec in the citation brief. We still run citation programs on Promptwatch. Here is how we explain the ranking.
Almost every citation brief we see already has three logos in a table: Profound, Otterly, Peec AI. A founder copied them from a thread. A procurement template reused last quarter's GEO RFP. We still do not start the program on those three. We run client citation work on Promptwatch, and we rank AthenaHQ and Scrunch above Profound when a second platform is even on the table. Otterly and Peec stay at the bottom of our comparison. This is the conversation we have in the kickoff.
What the client thinks they asked for
They say "citation tracking." What they usually mean is "tell us if ChatGPT mentioned us." That is a mention. A citation is the URL under the answer. The two disagree more often than slide decks admit. A brand can sit in a buying shortlist while every footnote goes to a publisher comparison nobody on the team controls. A docs URL can be cited in Perplexity without the brand ever appearing in the prose. A single visibility score would call both of those a mild win.
We split the kickoff report into four columns per prompt: named, own URL cited, third-party URL cited, source type. Promptwatch is where those columns live for us. Page-level and domain-level rows, Reddit and YouTube reports, offsite mentions (the brand inside a cited page that does not link back), type breakdowns, and citation trends. We do not invent a fifth column for "vibes."
AthenaHQ can sit next to that if the client wants an action queue and will pay from $295/mo, knowing ACE Citation Engine is an Enterprise feature around $2,000+/mo and that Reddit is not on the listing. Scrunch can sit next to it if they have engineering time for crawler-side page variants, from $250/mo annual, with weekly refresh. Profound, Otterly, and Peec get a fair read. They do not get to define the KPI.
How we actually measure citations
We put the client's money prompts into Promptwatch. Paid plans hit ChatGPT, Gemini, Claude, Perplexity, and the rest of the published UI set, including AI Overviews and AI Mode. Explore is free if we only need a ChatGPT sniff test (10 prompts). Essential is $95/mo, which is where we turn visitor analytics on: a script or GTM template so an AI-referred session is not a ghost in GA4. Professional is $245/mo when we need Agent Analytics. That is 25M crawler logs, ChatGPTBot through Meta's crawler, and the crawl-to-citation path.
The practical sequence on a client site looks like this. A citation disappears. We open the cited-page list and the trend, not the mention sparkline. If a competitor URL took the footnote, we brief a content change against that URL. If a Reddit thread took it, we do not write a pillar post; we look at the thread. If the crawler never fetched the page (robots, 404, timeout), we fix the fetch before we commission copy. That last step is why we budget Professional or an agency plan on accounts where ChatGPT Search matters. Essential already shows the citation objects and the traffic. The crawl path is the difference between "write more" and "unblock the bot."
Promptwatch is 4.7/5 on G2 with 1,840+ brands. We tell clients that as maturity context, not as a promise about their category. We also tell them Promptwatch does not sell screenshot evidence. If their board wants a PNG of ChatGPT, Peec does that. We still will not manage the program from screenshots.
What we say about the three logos
Profound still does citation share, answer-level analytics, and competitor cite-rates by cluster. Prompt Volumes is a real demand dataset. G2 4.6/5. The product they will be sold is not Starter. Starter is $99/mo yearly, ChatGPT-only, 50 prompts, one seat. Growth is $399/mo yearly for three engines. Enterprise is around $40k a year, marketing agents for AEO, content, demand, and PR, setup of one to three months. Reviewers have reported citation counts that do not match manual ChatGPT checks. For most of our retainers that is the wrong shape: too much suite, too little self-serve citation grain (no Reddit/YouTube/offsite depth on the listing, no published crawler-log product, no visitor-to-conversion analytics we can wire like Promptwatch). We do not put Profound at number one.
Otterly from $29/mo is the demo we let a skeptical stakeholder click. Four base engines, 15 prompts on Lite, data that can lag a week. Fine for "does GEO exist in our niche." We have never used it as the system of record for cited URLs. Gemini and Claude are add-ons. There is no PDF export. There is no crawl path.
Peec from $95/mo is the one agencies like because the dashboard is clean and every check has a screenshot. Source and citation breakdowns are listed. Extra models cost $35 to $165/mo. Starter is three models. No historical backfill. Monitoring only. When a client is already on Peec, we keep it for the steering-committee pack and we still run Promptwatch for the citation objects we ticket against. We do not lead a roundup with it.
What we tell them to skip in the first 30 days
Do not average ChatGPT and Gemini into one "AI rank." A win on one engine and a loss on the other is the finding.
Do not treat Search Console generative reports as ChatGPT citation share. Those are Google impressions.
Do not buy Profound Starter to "cover citations" when the brief named Perplexity. ChatGPT-only is ChatGPT-only.
Do not scale Otterly Lite past a proof. Fifteen prompts is a test allowance.
Do not sign Peec on the $95 sticker without pricing the extra models in the same email.
ChatGPT has 820M+ weekly active users. Gemini 650M+ monthly. Perplexity 22M+ monthly. We track the engines the client's buyers actually use, which for B2B is rarely ChatGPT alone.
A week of citation work on our stack
We start from the prompts where we lost the cited URL, then match losses to crawler errors when Professional is on. The content brief names the competing URL or the offsite source. If the CMS is Webflow or Framer, Content Agents can draft into a review inbox. Most client stacks are not those, so a writer still ships. Visitor analytics on the URLs that still win go in the same weekly note. The client gets the four-column table, not a blended score.
That week is why the logos in the original brief are the wrong ranking. Profound, Otterly, and Peec are real products. They are not the citation program we run. Promptwatch is, because the measurement, the crawl, and the traffic sit in the same project we already use for the rest of GEO. Agency Kick-off is $199/mo when we hold several brands. We would rather explain that once in kickoff than reconcile three trackers at QBR.