Best tool to track Copilot mentions: Microsoft hands you your own citation data for free, which changes what a tool is for on this surface

Before you buy anything for Copilot, open Bing Webmaster Tools, because Microsoft already reports your Copilot citations to you. Its AI Performance report, in public preview, covers "Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations". It gives Total Citations, Average Cited Pages, page-level citation activity, visibility trends over time, and the grounding queries, which Microsoft describes as the "key phrases the AI used when retrieving content that was referenced in AI-generated answers". Copilot is the only one of these surfaces where the operator hands a site owner that data directly, and it is free. That single fact reorders this purchase.

And one thing you should know before reading our recommendation: we do not read Copilot. Our own engine list is ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews. On this surface we are not a candidate, so the shortlist below is other people's products.

What Microsoft gives you, and what it says it does not

The report is worth reading on its own terms rather than as a free sample of a paid tool, and Microsoft is unusually clear about its boundaries:

• Total Citations counts how often your content appears as a source in AI answers, and Microsoft states it does not indicate placement.
• Average Cited Pages is defined as "the average number of unique pages from your site that are displayed as sources in AI-generated answers per day", and Microsoft states it "does not indicate ranking, authority, or the role of any page within an individual answer".
• Grounding queries are sampled, not complete: "The data shown represents a sample of overall citation activity", with a note that the metric will be refined as more data is processed.
• Page-level citation activity gives counts per URL, which is the half of the picture most paid tools charge for.
• In October 2026 Microsoft added a Topics Portfolio, in beta, inside the same report.

Read those limits carefully, because they define exactly what a tool is still for on this surface. Microsoft tells you when your pages were used. It does not tell you whether Copilot named you in the answer, whether it named a competitor instead, or what the answer said. Being cited and being named are two different findings, and this report measures only the first.

So the tool question on Copilot is narrower than on any other surface

You are not buying citation detection, because you already have it. You are buying three things the operator's report does not carry: the competitor column, the prompt set you chose rather than the queries that happened to reach you, and the stored text of the answer itself. Ranked on whether Copilot is available at the cheapest tier with a published price:

Vendor	Copilot at the cheapest published tier?	What the page says
Otterly.AI	Yes, in the base four on every published tier	"Tracking of 4 AI Search Engines: ChatGPT, Google AI Overviews, Perplexity, MS Copilot", with "Daily tracking frequency" and unlimited team members
Scrunch	Yes, in the Core four	"4 LLMs supported ChatGPT Perplexity Google AIO Copilot", with 125 unique prompts and 5,000 responses a month
Rankscale	Yes, on every plan card, and more deeply than anybody	"Microsoft Copilot" on each plan's engine list, "Copilot GUI" in its credit calculator, plus local and map pack tracking, sponsored ad measurement and "Bing Copilot Shopping" analysis
Ahrefs Brand Radar	Yes, named with the other six platforms	Standard metering of "1 check per platform, location and update"
Profound	No, Enterprise only	"Microsoft Copilot" sits in the Enterprise capability list of "Up to 9 Answer Engines tracked". The free trial covers three engines and not this one
Depra AI	No, an Enterprise add-on	"Every other model as an add-on: Microsoft Copilot, Claude, Google AI Mode and more (contact us)"
Peec AI	Only if it is one of your three slots	"Choose 3 models" on Starter, Pro and Advanced; "Up to 13 LLM models tracked" at Enterprise
Semrush	No	Its own FAQ answers "What LLMs does Semrush track?" with "ChatGPT, Gemini, Perplexity, SearchGPT, Google AI Mode, and Google AI Overviews". Copilot is not in that answer
AI Knows Us	No	Our engine row reads ChatGPT, Gemini, Perplexity, Claude, AI Overviews. Copilot is not one of our five

The recommendation, and it is not us

• Rankscale is first on the stated criterion, and by a clear margin. It is the only vendor in this set that publishes Copilot coverage beyond a mention: local and map pack results, sponsored ad share of voice, and shopping card analysis, all naming Copilot specifically, and all described as included on every plan. If Copilot is your surface, that is the deepest published treatment of it anywhere in this category.
• Otterly.AI is second and is the simpler purchase. Copilot is one of four engines on its cheapest published tier, checked daily, with unlimited team members. Otterly is also the vendor that treats Copilot as a given and Gemini as an add-on, which is the opposite of everybody else and makes it the natural pick for a Microsoft-weighted buyer.
• Scrunch is third, with Copilot in its Core four and the largest published prompt allowance at an entry tier in this set.
• Ahrefs Brand Radar is a reasonable fourth if you already own Ahrefs, with the clearest published metering rule in the trade.
• Peec AI works if Copilot is one of only three surfaces you need.
• Profound and Depra both cover Copilot and both put it behind a conversation, Depra as an add-on on top of its top tier.
• Semrush and AI Knows Us do not read Copilot. For us that is simply the answer, and the honest version of this page says so rather than arguing around it.

Otterly and Semrush are the pair the assistants name together most often in our own readings, and they differ on exactly two engines, one of which is this one. That pairing is worked through here, on one of 147 live pages of ours in this category with no durable inbound link, counted across 2,050 stored bodies.

Questions about tracking Copilot specifically

If Microsoft reports my citations free, why buy a tool at all?

For the three things the report leaves out. It tells you when your pages were used as sources. It does not tell you whether Copilot recommended you, whether it recommended a competitor on the same question, or what words it used. It also reports the queries that reached you rather than the buying questions you care about, so it cannot show the questions where you are absent. Absence is the finding that changes what you publish next.

Is the AI Performance report complete?

Microsoft says the grounding query data "represents a sample of overall citation activity" and that it will refine the metric as more data is processed, and the report is in public preview. Treat it as a strong directional first-party signal and read the trend rather than any single day's figure.

Why do some vendors skip Copilot entirely?

Because surface choice is a bet on where buyers are. Semrush's published list runs to six surfaces and reaches further into Google's than anybody's, with AI Mode and AI Overviews both named, and Copilot left out. Otterly made the opposite bet and sells Copilot in its base four with Gemini as an add-on. Both are coherent, and the right one depends on whether your buyers work in a Microsoft shop.

Does Copilot use the Bing index?

Microsoft reports Copilot citations through Bing Webmaster Tools, alongside AI-generated summaries in Bing, which tells you where the plumbing runs. The practical consequence is useful: the work that gets you indexed and crawlable in Bing is the work that puts you in range of this surface, and Bing Webmaster Tools is where you verify it. Microsoft also added support for the data-nosnippet attribute, which is the control to know about if you want a passage left out.

How do I check whether Copilot names my business right now?

Ask it your own buying question without your brand name in it, then read the footnotes. Do that for five real questions and write the names down. Then open the AI Performance report and compare: the questions where you were cited and the questions where a competitor was named are two different lists, and the gap between them is your brief.

Can a tool show me competitor Copilot citations?

Competitor comparison is the main thing you are paying for on this surface, since Microsoft's report covers only your own property. Rankscale publishes competitor benchmarking plus Copilot-specific ad and local reporting, Otterly publishes unlimited brands and competitors on every tier, and Scrunch publishes 5 competitors on Core. That column is the difference between knowing you were used and knowing whether you won.

The one surface where the operator did the first half of the job

Every other engine in this category makes you hire somebody to find out whether you were used as a source. Microsoft publishes it, for free, per page, with the phrases that pulled you in. So the sensible order on Copilot is to read your own grounding queries first, find the three questions where your pages are doing the work, and only then pay somebody to tell you who else is being named on those same questions. That order costs nothing to start and it narrows the tool decision to a single column, which is a better position than any vendor's comparison table will put you in.