⚡ Promptolis Original · Business & Strategy
🔍 Mom Test Customer Discovery
Real customer discovery without leading questions — Rob Fitzpatrick Mom Test framework with interview structure, sourcing, and pattern recognition.
Why this is epic
Most 'Mom Test Customer Discovery' prompts online produce generic, template-quality output. This one is structured like production-grade prompt engineering — role definition, principles, input schema, output format, auto-intake.
Research-backed: Real customer discovery without leading questions — Rob Fitzpatrick Mom Test framework with interview structure, sourcing, and pattern recognition.
Designed for practitioner-level depth, not generalist skim. Works across Claude, ChatGPT, and Gemini with consistent quality.
The prompt
Example: input → output
Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.
Business stage: Idea stage — no product built Target customer: Small law firms (5-15 lawyers) drowning in document review The problem: They spend 40%+ of billable time reading contracts + e-discovery documents, much of it routine Your hypothesis: An AI tool that pre-summarizes documents + flags unusual clauses would save 10-15 hours/week per attorney Interview count: 15 interviews before building anything Access: I know 2 attorneys personally; would need to cold-outreach to get to 15 Design my customer discovery.
Your Customer Discovery Plan — Small Law Firm Doc Review
Interview Structure
Each interview should be 30-45 minutes, structured roughly:
Minutes 0-5: Rapport + setting expectations
'Thanks so much for your time. I'm [name], exploring how small law firms handle document-heavy work. I'm NOT selling anything — I'm in research mode, trying to understand real workflows. I'll ask a bunch of questions about your day-to-day. Is it OK if I record for note-taking? (Share the eventual transcript or notes.)'
Minutes 5-15: Context + their world
'Tell me about your firm — size, practice areas, how you got here.'
What does a typical week look like?'
What aspects of the work do you enjoy most?'
Minutes 15-30: The pain zone (don't mention solutions yet)
'Walk me through the last big document-review project you did.'
'Where did that work become painful or tedious?'
'How much time did it take? How do you bill that?'
'What have you tried to make it faster or less painful?'
'What do you wish you could hand off to someone else?'
Minutes 30-40: Willingness to invest
'What did you spend on tools or help for that kind of work last year?'
'Who at your firm decides what software or services you adopt?'
'What has made you adopt new software recently — what broke through vs what didn't?'
Minutes 40-45: Referrals + close
'Who else at your firm or in similar firms would be interesting to talk to about this?'
'Is there anything I should have asked that I didn't?'
'Can I follow up if I have more specific questions later?'
Specific Questions to Use
Past-behavior anchors (gold standard):
- 'Walk me through the LAST time [specific scenario] happened.' [Specific, past, concrete]
- 'What did you try? What happened? What did you do next?'
- 'How often does that happen?'
Pain-specific probing:
- 'What's the hardest part of [their described process]?'
- 'What gets you frustrated with the current way?'
- 'What would have to change for [pain point] to go away?'
Investment signals:
- 'What have you already tried to solve this?' [Their workarounds reveal pain intensity]
- 'What tools or services have you paid for in the last year for this kind of work?'
- 'How much is [the problem] costing you per month in lost time?'
Network signals:
- 'Does this problem come up in conversations with other attorneys?'
- 'Who else shares this problem? Other firms like yours?'
Specific Questions to AVOID
- ❌ 'Would you use a tool that does X?' (Will always get 'probably yes' — worthless)
- ❌ 'If I built you X, would you pay for it?' (Same problem)
- ❌ 'What do you think of [my product idea]?' (Contaminates with your solution)
- ❌ 'How would you design a better system?' (Customers are not product designers)
- ❌ 'Would [specific feature] be helpful?' (Leading, contaminated)
- ❌ 'What are your biggest frustrations?' (Too abstract — probe specific situations instead)
How to Interpret Responses
Strong signal of real pain:
- They've already tried multiple solutions (paid tools, hired help, specific workarounds)
- They can quantify the cost in hours or dollars
- They describe it with emotion (frustrated, furious, 'makes me want to quit')
- They've talked about it with colleagues
- They have specific workflows they've developed AROUND the problem
Weak signal (don't over-weight):
- 'Yeah, that's kind of annoying, I guess.' [Mild, no workaround attempts]
- 'We've just learned to live with it.' [Acceptance, not motivation]
- 'I don't have time to deal with that.' [Deprioritized — unlikely to adopt solution quickly]
False positive (watch for this):
- 'Oh yeah, that would be great!' when you pitch your idea [politeness bias]
- 'I'd definitely pay for that!' in abstract hypothetical [future-hypothetical bias]
- Excitement about YOUR enthusiasm, not the problem [social mirroring]
Sourcing Interviewees
You have 2 attorneys personally — that's the foundation.
Path to 15 interviews:
1. Warm referrals from your 2 contacts (aim for 3-5 additional)
- 'Hey, I'm exploring how small firms handle doc review. Do you know 2-3 attorneys at firms like yours who'd give me 30 min? Just for research, not selling.'
2. LinkedIn cold outreach (aim for 7-10)
- Target: attorneys at 5-15 lawyer firms, associate level or partner, specifically in transactional law or litigation (heavy doc review)
- Message template:
> Hi [Name], I'm [your name], researching document-review workflows at small law firms. Not selling anything — genuinely trying to understand the day-to-day challenges. Would you have 30 min in the next 2 weeks for a conversation? Happy to share what I learn from others.
3. ABA Small Firm Resource Center or state bar associations — sometimes they have communities of small-firm lawyers open to research
4. Paid interviews: $100 gift card can unlock interviews. For 15 attorneys × $100 = $1,500. Worth it to compress timeline from 6 weeks to 3.
5. Attend small firm legal tech conferences (ABA TECHSHOW, LawNext) — great for density of potential interviewees in 2-3 days
Pattern Recognition Framework
After 5 interviews:
- Are they describing the SAME pain in different words? Or different pains entirely?
- Are the specific past-behavior examples they give similar?
- What WORKAROUNDS are they using? Are any common?
After 10 interviews:
- Do you see a specific segment within 'small law firms' that's more pained than others? (e.g., litigation > transactional? 5-10 lawyers > 10-15?)
- Are they already paying for something adjacent? Where is that budget sitting?
- Is there a natural 'who decides' pattern? (Managing partner? COO? Tech chair?)
After 15 interviews:
- You should be able to write a precise Ideal Customer Profile (ICP)
- You should know the SPECIFIC problem (not generic 'doc review')
- You should know what they've tried and WHY it failed
- You should have 3-5 direct quotes that capture the pain
What You're Looking For (to justify building)
GO signal:
- 10+ out of 15 describe the specific pain with similar details
- Multiple have tried paid solutions that didn't work
- They can quantify cost ($5K-$50K+ per attorney per year in lost time)
- They naturally ask 'wait — do you have something for this?' (pull signal)
PIVOT signal:
- Pain is real but different from what you thought
- The actual decision-makers aren't who you targeted
- Budget exists but in a different category than you expected
NO signal:
- Pain is mild, accepted
- They've learned to live with it
- Budget consistently low or non-existent
- Your 2-3 'enthusiastic' interviewees are outliers, not the pattern
Key Takeaways
- Mom Test: ask about past behavior, not hypothetical future behavior. 'Last time X happened' > 'Would you use a tool that does X?'
- Don't pitch your solution in first 15 minutes. Contaminates the interview.
- Silence is a tool. Wait 5 seconds after good answers — customers fill silence with deeper info.
- Look for evidence of investment (paid solutions tried, workarounds built, time spent). This indicates real pain.
- 15 interviews is the right target. 5-10 interviews reveal the pattern; 15 refines it.
- Sourcing: warm referrals first, cold LinkedIn second, paid ($100 gift cards) if needed to compress timeline.
- GO signal: 10+ of 15 describe same pain, quantify cost, some pull signal (they ask about your solution).
- Pivot signal: different pain than expected, different decision-maker, different budget category. Adjust your hypothesis.
- Document everything within 1 hour. Memory fades fast. Record + transcribe + notes.
Common use cases
- Professionals who need structured thinking on this topic, not vague advice
- Practitioners making specific decisions with real stakes
- Anyone tired of generic AI responses to domain-specific questions
- Users wanting depth over breadth — one thing done well, not 10 things done poorly
Best AI model for this
Claude Opus 4.7 for nuanced interpretation. Any LLM for interview structure.
Pro tips
- Paste your real situation (with specific numbers and context), not generic 'help me with X' framing. The prompt rewards specificity.
- If the prompt asks auto-intake questions, answer them fully before expecting output — incomplete inputs produce incomplete outputs.
- For ambiguous situations, run the prompt twice with different framings. Compare outputs. Often reveals the right path.
- Save the outputs you value. Iterate on them across sessions rather than re-running from scratch.
- Pair with a human expert for high-stakes decisions — the prompt is a first-draft tool, not a final authority.
- Share what worked back with us (promptolis.com/contact). Helps us refine future versions.
- The research citations inside the prompt are real — look them up if a specific claim matters for your decision.
Customization tips
- For consumer (not B2B) customer discovery, interviews are shorter (20-30 min) and often done via online panels or user interview services (User Interviews, Respondent.io) at $50-100 per interview. Consumers usually less willing to give free time than professionals.
- For non-English-speaking target customers, work with a native speaker for interviews — translated interviews lose significant nuance. Budget for interpreter $50-100/hour or recruit bilingual team member.
- For highly technical markets (developers, IT, engineers), adjust questioning: these audiences sniff out leading questions faster than most. They're more likely to volunteer specific detail. Spend less time on rapport, more on specific workflow questions.
- For enterprise customer discovery (large company decisions), expect long cycles: 6-12 months for customer decisions even if pain is clear. Interview multiple people at each target company — decision is committee-based.
- For existing customers (if you have early product), the dynamic shifts to discovering WHY they chose you vs alternatives, WHAT almost made them leave, WHAT would make them evangelize. Different question set.
- For pre-product (idea stage, no prototype yet), resist the urge to show wireframes or mockups until Minute 40+ of interview. Once you show, the interview converts from discovery to pitch feedback.
- For interviews that aren't producing signal (after 5-7 interviews, no clear pattern), the problem is often your target persona — you're interviewing the wrong people. Regroup and hypothesize a narrower or different segment.
- If interviewing is consistently producing lukewarm 'that's nice' responses across 10+ interviews, the pain is probably not strong enough for a business. Better to discover this at $0 investment than at $50K investment. Honest assessment > persistence.
Variants
Default
Standard flow for most users working on this topic
Beginner
Simplified output for users new to the domain — less jargon, more foundational explanation
Advanced
Denser output assuming practitioner-level baseline knowledge
Short-form
Compressed output for quick decisions, under 500 words
Frequently asked questions
How do I use the Mom Test Customer Discovery prompt?
Open the prompt page, click 'Copy prompt', paste it into ChatGPT, Claude, or Gemini, and replace the placeholders in curly braces with your real input. The prompt is also launchable directly in each model with one click.
Which AI model works best with Mom Test Customer Discovery?
Claude Opus 4.7 for nuanced interpretation. Any LLM for interview structure.
Can I customize the Mom Test Customer Discovery prompt for my use case?
Yes — every Promptolis Original is designed to be customized. Key levers: Paste your real situation (with specific numbers and context), not generic 'help me with X' framing. The prompt rewards specificity.; If the prompt asks auto-intake questions, answer them fully before expecting output — incomplete inputs produce incomplete outputs.
Explore more Originals
Hand-crafted 2026-grade prompts that actually change how you work.
← All Promptolis Originals