• BIAS
  • ASSUMPTION
  • EVIDENCE
  • HUMAN

How to audit customer discovery data with your AI

Connect your research notes to your AI client to strip away bias. Your agent turns messy interview transcripts into a clean list of what you actually know versus what you're assuming.

Ask AI about this page

Short answer

How can I use AI to audit my customer discovery research?

Your AI scans your interview transcripts to flag leading questions and unverified claims. You get a structured report that separates hard evidence from your own biases. This breakdown shows exactly where your data is weak.

Where every issue lands.

The audit outcomes

The AI categorizes your research into four specific buckets to show you what's solid and what's shaky.

BIAS

Flagged leading questions

The AI identifies questions that nudged users toward specific answers. You'll see exactly which phrasing compromised your data.

ASSUMPTION

Unverified claims

It pulls out statements you've treated as facts that no user actually confirmed. This prevents you from building on thin air.

EVIDENCE

Verified insights

The AI extracts direct quotes and behaviors that support your hypotheses. These are the points you can actually trust.

HUMAN

Required follow-ups

You get a list of gaps where the data is too thin to act on. These are the specific areas where you need to run more interviews.

The workflow

What your AI does when the data arrives.

The AI doesn't just summarize; it interrogates your notes to find the truth.

  1. Ingest research

    Your agent reads through your raw interview transcripts or research documents to build a baseline of the conversation.

    validate_customer_discovery

  2. Detect bias

    The AI looks for linguistic patterns in your questions that suggest a desired outcome rather than an open inquiry.

    validate_customer_discovery

  3. Cross-reference claims

    It compares your stated hypotheses against the actual words used by the customers in the transcripts.

    validate_customer_discovery

  4. Generate report

    The final output organizes these findings into a validation report that highlights the risks in your current understanding.

    validate_customer_discovery

Try it

Copy these to start.

Use these prompts once you've connected the MCP.

Starting points

These are opening lines. Swap out the bracketed text with your specific file names or folder paths.

Customer Discovery Prover Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_FcoU6IdAx3q5sWDJpDPGsPfVRCZ7YEkM4nbyPExm/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Customer Discovery Prover capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "customer-discovery-prover-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_FcoU6IdAx3q5sWDJpDPGsPfVRCZ7YEkM4nbyPExm/mcp"
    }
  }
}
Copy into chat04

  • Audit the interview notes in my Google Drive folder to find leading questions.

  • Check these Notion research docs for unverified assumptions about our target persona.

  • Review the transcripts in my local research folder and flag any biased phrasing.

  • Analyze the customer feedback in this spreadsheet to separate evidence from assumptions.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Customer Discovery Prover once, then ask.

Click the link to connect the MCP to your AI client. Once it's live, you can start auditing your files immediately.

Connect Customer Discovery Prover to your AI

FAQ

How this task behaves

  • 01

    Can the AI rewrite my interview questions for me?

    No. The AI identifies biased questions in your existing data, but it does not modify your original files or draft new ones.

  • 02

    Does the AI delete my raw interview notes?

    No. The AI only reads your data to perform the audit. It has no capability to delete or overwrite your original research.

  • 03

    What kind of files can I audit?

    You can audit any text-based research data like transcripts, notes, or spreadsheets that your AI client can access.

  • 04

    Will the AI tell me if my hypothesis is wrong?

    Yes. By comparing your assumptions against the actual user data, it will flag where your hypothesis lacks supporting evidence.

  • 05

    Can I use this with any AI client?

    Yes, as long as you are using an MCP-compatible client like Claude, Cursor, or Windsurf.

  • More questions about Customer Discovery Prover? The Connector page answers them. See everything the Customer Discovery Prover Connector can do

Connect Customer Discovery Prover to Claude, Cursor, ChatGPT & more