Copilot Tracking

    Copilot answers inside the working day. Does it name you?

    AnswerRank runs procurement and shortlist questions through Copilot, records which vendors it puts forward, and traces every answer back to the indexed page it came from.

    ChatGPTGeminiPerplexity
    Copilot
    Google AI Overviews
    AR
    Copilot, work context
    ILLUSTRATIVE
    “We're on Dynamics 365 — which expense management tools integrate with it?”
    Three options come up most often for Dynamics 365 environments. Vendor A has a certified connector on AppSource [1]. Vendor B supports it through Power Automate [2], and Vendor C is often chosen by mid-market finance teams [3].
    1appsource.microsoft.com
    2learn.microsoft.com
    3g2.com
    You aren't in the set.
    Your connector exists, but no indexed Microsoft-surface page says so.
    WHERE THE ANSWERS CAME FROM · 60 QUESTIONS
    Microsoft-owned surfaces41%
    AppSource listings, Learn docs, partner directory
    Review platforms28%
    G2, Capterra, TrustRadius
    Vendor sites23%
    Of which yours: 3%
    Everything else8%
    QUESTIONS WE PUT TO COPILOT
    “Expense tool with a certified Dynamics 365 connector”3rd
    “Vendor A alternatives for a 500-seat finance team”absent
    “Is Vendor A SOC 2 compliant and EU-hosted?”1st
    “Cheapest per-seat expense software for SMBs”absent
    You hold the compliance questions and lose every comparison question — a documentation problem, not an awareness problem.
    What we measure

    Copilot answers at work, which changes what counts

    The questions are longer, the stakes are procurement-sized, and the sources skew towards documentation. Four readings, kept separate.

    %
    Shortlist rate
    How often you make the set of vendors Copilot puts forward, across a fixed list of buying questions, run against run.
    ✓
    Answer accuracy
    Whether Copilot states your pricing, integrations, certifications and hosting correctly — the details that decide a procurement conversation.
    ⌘
    Source mix
    The split between Microsoft-owned surfaces, review platforms and vendor sites in your category — and how much of it you control.
    ⇄
    Head-to-head outcome
    What Copilot says when a buyer names you against a specific rival, including which of you it recommends and on what grounds.
    How it works

    Why Copilot picks the vendors it picks

    Copilot is grounded in a web index and answers next to the work — inside Edge, Windows and Microsoft 365. That gives four levers, all of them practical.

    Start trial →
    01
    Present in the index it reads
    Copilot grounds answers in Microsoft's web index rather than a private crawl. Indexing problems and blocked pages remove you before any judgement is made.
    02
    Documented, not just described
    Work questions ask for specifics: seat pricing, SSO, data residency, which systems you connect to. Claims that live only in a sales deck cannot be quoted.
    03
    Visible on the surfaces it trusts
    Marketplace listings, partner directories, platform documentation and review platforms carry disproportionate weight in software categories.
    04
    Safe to recommend
    Answers given at work stay conservative. Missing compliance pages, unclear ownership or stale changelogs read as risk, and risk gets skipped.
    B2B marketing teams
    Evidence for a channel nobody reports on yet
    A shortlist rate per buying question, the vendors named ahead of you, and the documentation gap behind each miss.
    Microsoft ecosystem vendors
    Your marketplace listing, read back to you
    See what Copilot actually says about your connector, certifications and supported versions, and where the description is out of date.
    Agencies
    A retainer with something to show each month
    Client workspaces, fixed question sets, and a record of which fixes moved which answers.
    FAQ

    Questions we get asked

    Is Copilot worth tracking separately?+
    In B2B, usually yes. It answers where people work — in Edge, in Windows, alongside Office files — so the questions lean towards evaluation and procurement rather than casual browsing. The vendor sets it returns often differ from ChatGPT's for the same category.
    Which questions do you run?+
    Buying questions in the phrasing a buyer would actually use: integration requirements, alternatives to a named incumbent, seat pricing, compliance and hosting, migration effort, and direct comparisons. Sixty per category is a workable starting set.
    Does our Bing ranking decide this?+
    Being indexed matters; ranking position matters less than most teams assume. A page can rank modestly and still be quoted if it answers the question plainly, while a top-ranked landing page full of positioning language gets passed over for a documentation page.
    Copilot says something wrong about our product. Now what?+
    We trace the claim to the page it came from. In most cases it is your own outdated documentation, a stale marketplace listing or an old review — all fixable at source. Then we re-run the question and confirm the answer changed.
    Can you see what Copilot says inside a customer's tenant?+
    No, and neither can anyone else. We measure the public, web-grounded answers — the ones every prospect sees before your name enters a procurement document.
    What does the first month look like?+
    Run one gives a baseline shortlist rate, the vendors appearing in your place, every factual error worth disputing, and the source mix for your category. Run two separates real gaps from noise. You end with an ordered list of fixes, most of them documentation rather than campaigns.

    Somebody is scoping vendors right now. Find out who Copilot handed them.

    One domain, sixty buying questions, first results the same day — shortlist rate, the vendors named ahead of you, and the pages behind each answer.