The method

A win-loss analyst, run as software.

It’s the same work a win-loss firm does — interview the buyers, synthesise the reasons, recommend the moves — only now it’s self-serve, runs on every deal, and takes days instead of weeks.

  1. 01

    Scope the study

    You talk it through with an AI analyst — your product, your competitors, and where you suspect you’re losing — and the assessment takes shape live as you do. It draws on a win-loss framework to propose the dimensions worth assessing and the hypotheses worth testing; you shape the scope, adding, dropping, and re-weighting what gets explored. The AI writes the actual questions so they stay neutral and never lead the buyer.

    Try the designer live →
  2. 02

    Preview with a test interview

    Before any buyer is contacted, you take the interview yourself, exactly as they will experience it. If a topic is missing or off-target, your feedback flows straight back into the design conversation and the scope revises in front of you. Nothing reaches a buyer until the study asks the right things.

  3. 03

    Interview the buyers

    You add your won, lost, and churned contacts — or connect your CRM so every closed deal flows in automatically. Each buyer gets a private, single-use link. Everyone answers a shared core (the same dimensions and a few standardised questions, so results are comparable), plus follow-ups scoped to what their deal and role let them speak to — a churned customer on onboarding, a lost prospect on the bake-off. Interviews are conversational, short, and never leading. Buyers answer in their own words, typed or spoken.

  4. 04

    Synthesise the evidence

    Responses are summarised per segment, then reduced into a prioritised view of why you win, lose, and churn — cut by outcome, competitor, and segment, including what winners say that losers don’t. The synthesis is evidence-bound and preserves dissent, so minority views aren’t averaged away. Your initial hypotheses are tested against what buyers actually said, so blind spots and confirmations both surface.

  5. 05

    Protect every respondent

    Before you see anything, a confidentiality guard enforces a minimum-N gate on every segment cut: any group too small to protect a person is suppressed or merged. Individual responses are never shown — unless a buyer explicitly opted in to be quoted — and identity is never joined to response content in anything you can see. It’s exactly why buyers tell the AI the truth they wouldn’t tell your rep.

  6. 06

    Draft the revenue actions

    Finally, a planner turns the guarded readout into a sequenced set of moves — pricing and packaging, sales enablement, product, retention — each tied to a specific finding and a success measure. It draws on curated playbooks to shape the how, but the evidence sets the what: every recommendation traces back to something the buyers said. Both deliverables take questions: ask follow-ups and the answers come only from the guarded readout, never raw interviews.

·You control
  • The product, competitors, and the questions worth exploring.
  • Which dimensions are in scope, and their emphasis.
  • Which buyers are interviewed, and when the study launches.
·The AI handles
  • Writing neutral, non-leading questions.
  • Adapting each interview to the buyer’s deal outcome and role.
  • Synthesis, confidentiality, and the drafted actions.

Ready to scope yours?