AI Features

Documentation Index

Fetch the complete documentation index at: https://docs.founder-sherpa.com/llms.txt

Use this file to discover all available pages before exploring further.

Founder Sherpa integrates AI throughout the discovery process to help you write better hypotheses, parse transcripts, generate questions, and analyze evidence at scale.

​

Problem statement quality check

When creating or editing a project’s problem statement, two layers of quality feedback are available:

​

Real-time heuristics

As you type, a quality indicator classifies your statement as weak, moderate, or strong. The heuristic checks for:

• Statement length and specificity
• Vague phrases (“improve things”, “better experience”)
• Missing specifics (numbers, roles, time/money references)
• Whether you’re describing a problem vs. a solution
• Whether the statement is phrased as a question

​

AI quality check

Click Check with AI for deeper analysis. The AI returns:

• A quality rating with reasoning
• Specific improvement suggestions
• An optional rewrite you can accept with one click

The AI uses your project name as context to ground its suggestions.

​

Hypothesis quality check

Similar to the problem statement checker, each hypothesis has an AI quality check button that evaluates:

• Whether the hypothesis is specific and testable
• Clarity of the target persona and expected behavior
• Suggestions for strengthening the statement

The AI receives your project name, problem statement context, and persona name to provide relevant feedback.

​

AI-suggested interview questions

While building interview guides, click Suggest Questions to generate targeted interview questions. The AI considers:

• The persona you’re researching
• The hypotheses you’ve selected to test
• Your project’s problem statement

Each suggestion includes:

• The question text
• A rationale explaining why it’s relevant
• A suggested category

You can add suggestions to your Question Library or insert them as custom notes in the guide.

​

Transcript parsing

When you import a transcript, AI processes the full text to extract:

• Evidence items — Key insights, quotes, and observations organized by hypothesis
• Interviewee identification — Who was being interviewed, with fuzzy matching against existing interviewees
• Question detection — Questions asked during the interview, matched against your question catalog

For long transcripts, the system automatically chunks the content and processes each section, then merges the results.

​

Hypothesis validation

From the project’s Hypotheses tab, you can validate individual hypotheses against collected evidence:

1. Select one or more hypotheses using the inline checkboxes
2. Click ✦ Validate with AI
3. The AI analyzes all linked evidence and returns an inline result panel with:
   • A validation verdict (supported, contradicted, or inconclusive)
   • Key evidence patterns and themes
   • Gaps in your research that need more interviews

Selection persists across filter changes, so you can select hypotheses from different filters and validate them together.

​

Batch analysis

From the project’s Insights tab, run batch analysis across all collected evidence. The AI:

• Summarizes patterns across interviews
• Identifies which hypotheses have strong support or contradiction
• Highlights gaps where more evidence is needed
• Generates an audit report of your discovery progress

This is the “Refine & Repeat” step of the Discovery Journey — helping you decide what to investigate next.