AI Search Visibility Monitoring Tools: ChatGPT and Perplexity Brand Mentions, Citations, and Ads Monitoring

Most RFPs for AI search visibility monitoring now ask for three columns: brand mentions, citations, and ads. They ask for all three on ChatGPT and on Perplexity. That is six cells, and in October 2026 only five of them describe something that exists.

ChatGPT has mentions, citations, and a live ad product. Perplexity has mentions and citations. Its ad program is gone. Perplexity stopped taking new advertisers in October 2025, according to Adweek, and in February 2026 its executives told reporters they were moving away from ads entirely (WIRED). So if a vendor demo shows you a "Perplexity ads" chart, ask what it is counting.

This guide is a checklist for buying one tool that fills the five real cells. Promptwatch is our default pick, and the reasons are specific. They are below.

What each column has to capture

Brand mentions. The tool runs your prompts on a schedule and records whether your brand is named, where it sits in the answer, and how the answer talks about you. Minimum fields: prompt, engine, date, position, sentiment, competitors named alongside you.

Citations. The tool records which URLs the answer cites. Good tools go to the page level, not just the domain, and break out Reddit, YouTube, and third-party pages that mention you without linking. Without page-level citations you can see that you lost, but not what to fix.

Ads. The tool captures the sponsored unit that sits under a ChatGPT answer. OpenAI's own help center describes what that unit contains: advertiser name, favicon, title, copy, landing page, and an image (Ads in ChatGPT: The Basics). A monitoring tool should store those fields against the prompt that triggered them, so you can see which of your buyer questions are being bought by someone else.

Note what OpenAI does not offer. Advertisers get aggregated reporting on their own ads, such as views and clicks (Ads in ChatGPT). None of OpenAI's ads help pages describe a public ad library where you can look up a competitor's ChatGPT ads. If you want to see rival ads, you need a tool that watches real answers.

The buyer checklist

Print this and score each vendor in the demo.

Real interface capture. Ads render in the ChatGPT product, below the answer. Ask whether the vendor collects from the user interface or from an API call. Promptwatch states it monitors real UIs, not just APIs.
Ads tied to prompts. An ad without the prompt that triggered it is trivia. You want a by-prompt view: which tracked questions attract ads, how many, and when the latest one appeared.
Organic and paid side by side. The useful question is "we are cited organically on this prompt, and a competitor is also buying the slot under it." That needs both data types in one project.
Perplexity handled honestly. Mentions and citations on Perplexity, yes. An ads column on Perplexity, no.
Location. OpenAI lets advertisers target by country, by region through location IDs, and by platform (targeting docs). A UK monitor and a US monitor can see different ads. Ask what geography each monitor runs from.
Daily cadence. Ad rotation is fast. Weekly snapshots miss campaigns.
A plan that actually includes the ads feature. This is where trials fail. More on it below.

How the shortlist scores

Promptwatch. Fills all five real cells. Prompt tracking with sentiment, share of voice and competitor benchmarking covers mentions on every tracked engine, including ChatGPT and Perplexity. Citation analytics go to page level, with Reddit, YouTube, and offsite-mention reports. Ads Radar captures sponsored ads inside ChatGPT Search, Google AI Overviews, and Google AI Mode answers to your tracked prompts, with title, description, image or price, advertiser, domain, target URL, and the prompt that triggered it (Ads Radar docs). The By Prompt table shows each prompt's ad count next to your own average visibility on it, which is the paid-versus-organic view in one row.

It also covers what the other columns cannot explain on their own. Crawler logs (Agent Analytics) show whether ChatGPT's and Perplexity's bots fetched the page you wanted cited. Visitor analytics show whether any of it sent traffic that converted.

Gauge. The catalog lists Gauge from $599/mo with ChatGPT Ads tracking, citation intelligence, and six engines including ChatGPT and Perplexity. Gauge's own feature page also says you can connect an OpenAI Ads account to review your campaign spend and clicks (Gauge ChatGPT ads). The catalog notes no published crawler-log product and no visitor analytics that attribute AI traffic to conversions.

LLM Pulse. LLM Pulse lists "ChatGPT Shopping and AI ads tracking" from €49/mo across five engines, including ChatGPT and Perplexity. The default cadence is weekly, which matters for ads. The listing does not say which engines the ads tracking covers.

AthenaHQ and Scrunch. AthenaHQ (from $295/mo) and Scrunch AI (from $250/mo annual) both cover mentions and citations across many engines. Neither catalog entry documents an ads capture feature. Good for the first two columns, not the third.

Profound. Profound starts at $99/mo annual, ChatGPT-only on Starter, with three engines on Growth and the full set on Enterprise contracts around $40k a year. No ads capture is listed in our catalog.

The plan gate that sinks trials

Promptwatch changed its plan gating this month. Per its llms.txt, checked on 2026-10-21, Ads Radar and Shopping Insights are included from the Business plan at $579/mo, and on every agency plan: Kick-off $199/mo, Growth $399/mo, Scale $799/mo. Essential ($95/mo) and Professional ($245/mo) do not include Ads Radar. On those plans the page shows sample data behind an upgrade overlay.

That matters for procurement. A brand team that starts a seven-day trial on Essential to "test ads monitoring" will test mentions and citations, which is fine, but it will not see a single captured ad. Decide up front which column you are trialing. If ads are the reason for the purchase, budget for Business, or for Kick-off if you are an agency.

How to run the evaluation

Pick 30 to 50 buyer prompts. Make most of them commercial: "best X for Y," "X vs Y," "X pricing." Promptwatch's own ad frequency report found ads on 20.1% of search-backed ChatGPT responses over the 90 days to mid-August 2026, with the daily rate mostly between the low 20s and high 30s after July 1. A short evaluation on informational prompts may simply show zero ads and prove nothing.

Run the same prompts on ChatGPT and Perplexity for two weeks. Then score the vendor on three questions. Can it show me, per prompt, who was mentioned and which page was cited? Can it show me the ads under my ChatGPT answers, with landing pages? Can it tell me why I lost a citation, down to whether the crawler fetched my page?

The first question every tool here can answer. The second narrows the list to tools with an ads capture feature. The third is where Promptwatch is usually the only one left, because it has the crawler logs and visitor analytics the trackers do not.