Home / Features / ICP & Buyer Personas · Beta

Your average hides who you lose.

ICP & Buyer Personas reads your scan, finds the roles asking AI about your category, and recounts every number one buyer at a time. No guessing, no language model inventing people.

See my buyer personasSee it in the product
20 buyer rolesUp to 4 personasBuilt in codeSame scan, same audience
Buyer
30%

All buyers average 30%
    Win them →
    Example · invented brand Vela, numbers from a sample scan
    The same 54 answers, sorted by buyer

    One dot is one AI answer. Green recommended you.

    Pick a buyer and every other answer fades. The green and red that were mixed together in one number separate into who you win and who you lose.

    Recommends youDoes not
    30%

    Example · invented brand Vela, numbers from a sample scan
    In the product

    Who to target. Who to talk to. What to do next.

    The Audience tab, in three steps. Every sentence comes from your own stored evidence.

    1Who to target

    Mid-size banks and credit unions that need audit-ready compliance without more spreadsheets.

    Company size
    200 to 2,000 employees
    Where
    North America · United Kingdom
    They buy when
    an audit or new regulation is coming
    They compare you with
    Kestrel · Pipewise
    Built from vela.example and 54 AI answers ·

    2Who to talk to

    12 questions are asked by every buyer: brand, category and alternatives.

    3Your next moves

    1. 1
      Win the Compliance leadAI recommends Kestrel in 11 of their 18 answers. Three open actions would change that.
      Open actions →
    2. 2
      Add questions Finance leaders askOnly 12 answers speak to Finance. Add buyer questions that say “for a CFO”.
      Add buyer questions →
    3. 3
      Earn the pages AI cites for IT leadersPipewise is cited from three pages you have no answer for.
      See sources →
    Example · invented brand Vela, numbers from a sample scan
    How personas are built

    Evidence picks the buyers. Not a language model.

    A fixed catalog of 20 buyer roles is scored on your own data. A role becomes a persona only when independent evidence agrees. The same scan always gives the same audience.

    1
    Your site and brand profileIndustries, regions, the words you use for your buyer.
    2
    Your buyer questions“…for a compliance team”, “…for a CFO”.
    3
    The roles AI names in its answersWho the answers are written for.

    Two of three agree, the role clears a minimum score, and it becomes a persona. Up to four.

    Per-buyer trend

    Watch the one buyer you worked on move.

    Every persona gets its own trend line across your weekly scans, so a win with one buyer does not disappear inside the average.

    After publishing for one buyer
    +20 pts

    Illustrative example with an invented brand. It is not a promised result. Each line is recounted from your own stored answers on every scan.

    Impact

    From one blended score to a plan for each buyer.

    Without buyer personas
    30%
    • One number for 54 answers. It cannot say who is skipping you.
    • Fixes aimed at everyone, so they reach no one.
    • A win with one buyer is hidden by the others.
    • Sales and content guess who to speak to.
    With ICP & Buyer Personas
    4 buyers
    • Each buyer has their own rate, rival and trend line.
    • The Action Board ranks work by answers each buyer-fix can change.
    • Every win is tied to the buyer you aimed it at.
    • Sales and content get the questions, the sources and the angle for each role.
    Content

    Write for the buyer you lose

    The playbook lists the questions that buyer asks and the pages AI cites. You publish the page AI is missing.

    Outreach

    Know who to talk to first

    The Compliance lead decides and the Finance leader signs off. The role badge shows who holds which part of the deal.

    Reporting

    Show progress per buyer

    "Recommended to Compliance leads in 31% of answers, up from 11%" is easier to defend than a flat average.

    Advantages

    Why it beats guessing your audience.

    1

    Measured, not imagined

    Personas come from your site, your buyer questions and the roles AI names. Nothing is invented.

    2

    Repeatable

    No language model decides. The same scan gives the same audience, so you can trust a change.

    3

    Re-scopes every tab

    The Buyer lens changes the whole dashboard to one role: rates, rivals, sources and actions.

    4

    Tied to action

    "Win them" opens the actions for that buyer, and Creative Studio builds the asset from the same facts.

    5

    A trend per buyer

    See whether the buyer you worked on moved, even when the average did not.

    6

    A playbook with sources

    Questions they ask, what they weigh, what to say, and the sources to earn.

    Why it is different

    A persona deck describes buyers. This one measures them.

    A persona deck or workshopICP & Buyer Personas
    Source

    Opinions and a workshop

    Your site, your buyer questions and what AI answered

    Numbers

    None

    Recommendation Rate, recounted for each buyer

    Stays true

    Out of date by next quarter

    Recounted on every weekly scan

    Used for

    A slide that sits in a folder

    Re-scopes the whole dashboard and decides the next action and asset

    PersonaCompliance lead

    AI recommends Kestrel in 11 of 18 answers.

    Action BoardWin them · 3 actions

    Ranked by how many answers each can change.

    Creative StudioCarousel, video, landing page

    Built for that buyer, from the same counted facts.

    Next scanSame questions again

    Their trend line shows whether the answer moved.

    Raise it

    How to improve your ICP & Personas: From the number to the fix.

    MeasureRecommendation Rate

    Recommendation Rate is the share of AI answers to real buyer prompts in which an engine recommends your brand as an opti…

    Explore →    MeasureIntent Coverage

    Intent Coverage spreads every scan's prompts across six buyer intents, from category discovery to brand evaluation, and …

    Explore →    UnderstandCompetitor Tracking

    AI competitor tracking measures how often AI engines mention and recommend each competitor on the same buyer prompts as …

    Explore →    FixCreative Studio

    Creative Studio is the MentionX asset builder. It starts from a gap on your Action Board, writes a strategy brief, appli…

    Explore →
    ICP & Personas FAQ

    ICP & Personas FAQ: Questions about buyer personas.

    How personas are built, how many you get, and what the lens changes.

    A buyer role, such as finance, compliance or marketing, that your scan shows asking and being answered about your category. MentionX infers up to four from your own evidence and shows the questions each one asks.

    No. The audience view is built in code from your stored scan data, so the same scan always gives the same audience. A language model is not asked to guess who your buyers are.

    Up to four inferred from the evidence. A role appears only when it clears a minimum score and independent evidence agrees, so a thin scan can show fewer.

    A selector in the dashboard that re-scopes every tab to one persona's questions and recounts every KPI for that buyer. Clear it to return to the full scan.

    Intent Coverage groups questions by what the buyer is trying to decide, across six intents. Personas group the same questions by who is asking. They answer different questions, so read both.

    Yes. It is in Beta and available on every plan once a scan has completed.

    Start free

    Which buyer is AI skipping you for?

    Enter your URL and market. After your first scan, the buyer lens shows your numbers one persona at a time.

    All features