AI Visibility Tracking Tools: Profound, Otterly, Peec for AI Search Rankings

"AI search rankings" is the phrase that arrives in the brief. The named tools are almost always Profound, Otterly, and Peec. The buyer wants a position next to a competitor the way they already have a Google rank. Assistants do not publish a rank column. They mention you, cite you, or they pick someone else's URL.

We still run a tracker. We refuse to invent a composite "AI rank" from four engines. The tracker we open first is Promptwatch. If Profound, Otterly, or Peec is already paid and the export is current, we work from that file.

What we mean by an AI search ranking

A ranking in this work is a stored answer to a frozen prompt: mentioned or not, cited or not, which domain won the footnote. Same wording, same engines, every refresh. That is close enough to rank tracking that a director can read it. It is not a #3 in a SERP.

We split mention from citation on purpose. A mention without a link can still win a shortlist. A citation of a third-party roundup is a different ticket than a citation of your own docs.

Google Search Console generative reports stay in the deck. They are Overviews and AI Mode. They are not ChatGPT or Perplexity rankings.

Profound, Otterly, and Peec as ranking tools

Profound. Procurement likes the name. Public Starter is $99 on annual billing and ChatGPT only. You cannot report Perplexity or Gemini rankings from that SKU. A fuller Profound contract can be the company standard. If that is what they signed, we use it. If they bought Starter and now ask why the other engines are empty, we say so.

Otterly. The published entry price we use is $29. Fine as proof the category exists. We inherit it when a brand team already lives there. We do not treat it as the system of record for weekly ranking movement if we have to spend the meeting asking whether the miss is real or stale.

Peec. Starter sits at $95, the same sticker as Promptwatch Essential. For a one-brand shop that only wanted mention tracking, Peec is a fair comparison. It is a tracker. It does not give us crawler logs or visitor analytics on the same project. If the list is small and the client will not move, we keep Peec and add those checks ourselves.

We name those three because the brief named them. We do not score them with invented review ratings.

The tracker we actually run

Promptwatch paid plans store answers from the UIs people use, including ChatGPT, Gemini, Claude, Perplexity, and Google's AI Overviews. Citation analytics sit at page and domain level. That is the ranking unit we will defend: this URL, this prompt, this engine.

Essential is $95/mo for one project and 50 prompts. That is a working first ranking program. Agency Kick-off is $199/mo with unlimited projects and 10 seats when we are holding more than one brand. Visitor analytics (script or GTM) sit next to the ranking log so a citation week and a session week are not treated as the same number.

Explore is free and ChatGPT only (10 prompts). Use it to show a stakeholder the workflow. Do not run the ranking program on Explore.

Crawler logs matter because a "rank loss" that is actually a blocked GPTBot is an engineering ticket, not a content sprint. OpenAI's bot docs are the reference for the OpenAI pair.

How we turn a ranking cell into work

Freeze 25 to 50 buyer prompts. Sales and product write the first draft. SEO adds the comparison URLs we already lose in Google.
Check the same list on every engine the plan includes. Do not report Starter as multi-engine coverage.
Ticket by type: crawl block, passage miss, no page, third-party win. Promptwatch fan-outs often name the URL the model retrieved. That last type goes to PR or digital, not to "write 20 blogs."
Change one URL. Re-check the same prompts. Two weeks is the earliest we call a rewrite a ranking result.

The page work is still SEO-shaped: a dated first sentence under the heading that maps to the prompt, entity names that match how people type the brand, a table if the question is a comparison. We wrote that loop in GEO and AI search optimization. The tracker does not do the edit.

Classic rank tracking stays. SEO vs GEO is the split we already use.

What we will not do with these tools

We will not average ChatGPT, Gemini, and Perplexity into one AI ranking for a slide. We will not accept Profound Starter as four-engine coverage. We will not treat a weekly Otterly pull as yesterday's rank.

If the client already has a four-engine tracker they trust, we will use that export.

FAQ

Is an AI search ranking the same as a Google rank?

No. Google has a position. An assistant has an answer. We store mention, citation, and winning domain. That is the closest honest ranking we can give you.

Should we buy Profound because it sounds like the ranking platform?

Buy Profound if the company already standardized on a fuller tier. Do not buy the $99 annual ChatGPT Starter and call it AI search rankings.

Can we stay on Peec at $95 and still get rankings?

Yes, if the export is current and the prompt list is honest. You will still lack crawler logs and AI-referred sessions in the same project. We will say that in the first review.

What to do this week

Write 25 prompts from real sales questions. Use the same wording on every engine.
If you already pay Profound, Otterly, or Peec, export mention vs citation and write down which engines that SKU includes.
If you have no log, create a Promptwatch project on Essential ($95) or Kick-off ($199) if you need more than one brand.
Pick one red prompt and change one URL. Do not launch a content calendar because twenty cells went red.
If you want us to run the weekly ranking review, email hello@1001seomedia.com with the list and the current vendor. The tool stores the answer. We still ship the fix.