3 phases · 7 prompts. PREP before each call. SYNTHESIZE after each call. SHIP the recap email + Cohort 2 brief when you're done. Every prompt uses the expert-panel pattern — paste, fill in your context, ship.
Phase 1 · BEFORE the interview
Walk in prepared, not generic
P01⏱ 5 min · run before each call · save the output
Personalized Pre-Interview Briefing
When to use5 minutes before each 1:1. Generates 2-3 personalized questions to weave into the standard 11 — so the student feels SEEN, not surveyed.
# ROLE
You are my world-class research panel for prepping a 30-minute customer interview. Channel:
- **Rob Fitzpatrick** (The Mom Test) — to keep questions behavioral, never hypothetical
- **Indi Young** (qualitative research expert · Listening Deep) — to spot what to listen FOR in this specific student
- **Teresa Torres** (Continuous Discovery Habits) — to translate insights into product decisions
- **Erika Hall** (Just Enough Research) — to keep it pragmatic, not academic
# CONTEXT
Based on what you already know about me: I run Build Sprint · 30 Days to Ship at Network School. Cohort 1 just finished (25 builders, 80% shipped). I'm interviewing every student 1:1 to design Cohort 2. The standard 11-question protocol covers: opening, outcomes, what worked, what to fix, future intent, close.
# THIS STUDENT
Name: [STUDENT NAME]
What they shipped: [1-SENTENCE DESCRIPTION OF THEIR PROJECT]
Where they got stuck most: [RECALL FROM COHORT CHAT · WHAT FRUSTRATED THEM]
Conversion status: [PAID USERS YES/NO/IN-PROGRESS]
One thing I remember about them: [PERSONAL DETAIL · OPTIONAL]
# YOUR JOB
Ask me 3 quick clarifying questions about this student before generating the briefing. Then give me:
1. **Open with this:** A personalized opening that's warmer than "How are you?" — something that references their specific journey
2. **2-3 personalized questions** to weave into the standard 11. Each must be (a) behavioral not hypothetical, (b) specific to THIS student's project, (c) likely to surface a testimonial-grade quote
3. **The trap to avoid with this person** — what topic might they over-praise on, or under-share? Where might they get diplomatic when I need them blunt?
4. **One surprise question** I should ask if there's time — the one that might shake loose something unexpected
# QUALITY BAR
Indi Young would say: "The job of an interview is to learn what's TRUE for this person, not to validate your hypothesis." Hold me to that. If my framing is leading or hypothetical, push back.
Output as a 1-page brief I can read in 90 seconds before the call.
Expert panel
Rob Fitzpatrick · Mom TestIndi Young · Listening DeepTeresa Torres · Continuous DiscoveryErika Hall · Just Enough Research
Phase 2 · DURING the interview
When they go vague, probe surgically
P02⏱ 30 sec · open in a second tab · paste mid-call
Live "Mom Test" Probe Generator
When to useThe student gave a vague answer ("It was great, I learned a lot"). You need a follow-up question in 10 seconds. Paste their answer, get 3 ways to probe deeper without sounding interrogative.
# ROLE
You are my real-time interview coach. Channel:
- **Rob Fitzpatrick** (The Mom Test) — for the 3-deep "why" technique
- **Michael Bungay Stanier** (The Coaching Habit) — for the "what else?" question that opens doors
- **Chris Voss** (Never Split the Difference) — for mirroring and labeling techniques that surface truth
# SITUATION
I'm mid-interview with a Build Sprint Cohort 1 student. They just gave me a vague answer and I need a sharper follow-up RIGHT NOW.
# QUESTION I ASKED
"[QUESTION I JUST ASKED]"
# WHAT THEY SAID (vague, not actionable)
"[THEIR ANSWER · PASTE VERBATIM]"
# YOUR JOB
Give me 3 follow-up questions I could ask in the next 10 seconds. Each must:
1. Be SHORT (under 12 words)
2. Reference a SPECIFIC WORD they just used (mirror technique)
3. Force them to a CONCRETE example, not another opinion
4. NOT be hypothetical, NOT be leading, NOT pitch the product
Format as:
**Option A (concrete probe):** "..."
**Option B (mirror probe):** "..."
**Option C (silence-then-...):** "..."
Pick your favorite and tell me why in one sentence.
Expert panel
Rob Fitzpatrick · Mom TestMichael Bungay Stanier · Coaching HabitChris Voss · Never Split the Difference
Phase 3 · AFTER each call · within 24 hours
Synthesize while it's fresh
P03⏱ 5 min per interview · run on transcript or notes
When to useRight after the call ends · while the body language is still in your head. Paste the transcript or your notes. Get a structured digest you can drop into your master tracker.
# ROLE
You are my world-class qualitative-research synthesis panel. Channel:
- **Teresa Torres** (Continuous Discovery Habits) — to translate raw stories into product decisions
- **Indi Young** (Listening Deep) — to spot underlying purposes the student didn't name explicitly
- **Steve Portigal** (Doorbells, Danger, and Dead Batteries) — to spot the surprise nobody asked about
- **April Dunford** (Obviously Awesome) — to spot positioning language I should steal for marketing
# CONTEXT
I just finished a 30-min feedback interview with a Build Sprint Cohort 1 student. I'm running this immediately while it's fresh.
# STUDENT
Name: [STUDENT NAME]
Project shipped: [WHAT THEY BUILT]
# TRANSCRIPT OR NOTES
[PASTE FULL TRANSCRIPT OR YOUR NOTES HERE]
# YOUR JOB
Output a 1-page synthesis with EXACTLY these 6 sections:
## 1. The headline (1 sentence)
The single most important thing this student told me. If I read only one line, this is it.
## 2. Sean Ellis answer + reason
A (very disappointed) / B (somewhat) / C (not really) + the verbatim reason they gave.
## 3. Testimonial-grade quotes (max 3)
Verbatim quotes I could use on a landing page. Each must be: specific, emotional, and ideally include a number or named outcome. Tag each with [PERMISSION NEEDED] so I remember to follow up.
## 4. What they LOVED (specific to sessions/prompts/people)
The 2-3 things they named as the highest value. Be specific — not "the program" but "the Day 7 pricing prompt."
## 5. What they want CHANGED (the one big lever)
Their ONE change suggestion + my read on whether this is additive ("more X") or replace ("instead of X").
## 6. The surprise
The one thing they said I wasn't expecting. The pattern-break. If you don't see one, say so explicitly — don't manufacture.
# QUALITY BAR
Steve Portigal: "Research is the act of being surprised, not confirmed." If everything in this synthesis confirms what I already knew, you missed the surprise. Look again.
When to useAfter all 25 interviews are complete. Paste all 25 synthesis outputs from P03. Get the ranked themes, ranked changes, and the patterns no single interview revealed.
# ROLE
You are my world-class research synthesis panel for cross-interview pattern detection. Channel:
- **Teresa Torres** — opportunity solution trees, ranking insights by impact × confidence
- **Erika Hall** (Just Enough Research) — pragmatic prioritization; what to ACT on vs ignore
- **Tomer Sharon** (Validating Product Ideas) — the rigor of multi-source triangulation
- **Christensen / JTBD school** — to spot the "job" the program is actually being hired for
# CONTEXT
I run Build Sprint at Network School (30-day ship-your-AI-product program). Cohort 1 just finished. I interviewed [X] of 25 students. I'm now synthesizing across all of them to design Cohort 2.
# DATA
[PASTE THE 25 P03 SYNTHESIS OUTPUTS · OR THE FULL TRANSCRIPTS · OR PASTE A NOTION DB EXPORT]
# YOUR JOB
Output a 2-page strategic brief with EXACTLY these sections:
## 1. Sean Ellis PMF score
Count A / B / C answers. Calculate % "very disappointed" (A). Is it above the 40% PMF bar? What does the rationale data say about WHY?
## 2. Top 5 themes (ranked by frequency × intensity)
For each theme:
- The theme (1 line)
- # of students who raised it
- Representative verbatim quote
- Whether it's CONFIRM (matches my hypothesis) or SURPRISE (didn't expect)
## 3. Top 5 testimonials (ranked for marketing impact)
The 5 quotes I should use on the Cohort 2 landing page. Each one: specific outcome, named person, emotional resonance. Tell me which persona each quote sells to.
## 4. The JTBD finding
What "job" did students actually hire Build Sprint for? Be precise — not "to learn AI" but "to finally finish a side project I've been stuck on for 8 months." This is your Cohort 2 positioning.
## 5. Top 3 changes for Cohort 2 (prioritized)
The 3 changes the data strongly supports. For each:
- The change
- Evidence (# of students + verbatim)
- Effort to implement (low/med/high)
- Risk if I DON'T do it
## 6. The 1 surprise I should investigate further
The pattern that emerged that I didn't go looking for. The thing that might be a bigger opportunity than Cohort 2 itself.
## 7. What to KEEP (don't change)
The 3 things students named as protected — never change these.
# QUALITY BAR
Teresa Torres: "If your insights don't drive a decision, they're not insights — they're just observations." Every line in this brief should drive a Cohort 2 decision. If a section is decorative, cut it.
Expert panel
Teresa Torres · Continuous DiscoveryErika Hall · Just Enough ResearchTomer Sharon · Validating IdeasClayton Christensen · JTBD
P05⏱ 10 min · send to all 25 alumni
Cohort Recap Email · "here's what you taught me"
When to useAfter P04 is done. Closes the feedback loop with respect, shows the cohort their input mattered, and quietly seeds Cohort 2 referrals. Sent to all 25 students.
# ROLE
You are my writing partner for a high-stakes alumni email. Channel:
- **Patrick Collison** (Stripe founder letters) — for warm + substantive + specific
- **Andy Raskin** (The Greatest Sales Deck) — for narrative arc that makes the reader feel like a protagonist
- **Joanna Wiebe** (Copyhackers) — for conversion-grade copy without being salesy
- **Jason Fried** (Basecamp) — for under-300-word discipline
# CONTEXT
I run Build Sprint at Network School. Cohort 1 just finished (25 builders). I interviewed every one and now I'm sending a "thank you + here's what changes" email to all 25. Goal: close the loop, build loyalty, plant the Cohort 2 seed (without pitching).
# WHAT I LEARNED (from P04 cross-synthesis)
[PASTE P04 OUTPUT · OR THE KEY SECTIONS]
# CONSTRAINTS
- 250-350 words max (under-300 is ideal)
- Use my voice: warm mentor, motivational, direct, contractions, no jargon
- Use "you" and "we" naturally
- NO pitch for Cohort 2 anywhere · just plant the seed: "if you know a builder who'd thrive in this..."
- Sign as Megan
- Subject line < 60 chars · should make them WANT to open
# YOUR JOB
Write the email. Structure:
1. **Subject line** (3 options to choose from)
2. **Preview text** (under 90 chars)
3. **Open** — name the moment ("I just finished 25 interviews. Here's what you taught me.")
4. **The 3 things you taught me** — bulleted, specific, ATTRIBUTED ("Sara said..." with permission, or anonymized)
5. **The 1 thing changing in Cohort 2** — concrete, not vague
6. **The 1 thing staying sacred** — the thing they protected
7. **Close** — gratitude · the soft Cohort 2 seed · sign-off
# QUALITY BAR
Jason Fried: "If it can be 200 words, don't make it 400." Patrick Collison: "Be specific or be quiet." Cut anything that doesn't earn its space.
End with: 3 subject lines I should A/B test, and your pick + why.
When to useAfter P04 cross-synthesis. Translates feedback into a concrete Cohort 2 product spec: which of the 30 sessions to KEEP exactly, REDESIGN, KILL, and what to ADD.
# ROLE
You are my world-class curriculum + product strategy panel. Channel:
- **Bob Moesta** (JTBD architect) — what job does Cohort 2 need to do better than Cohort 1?
- **Marty Cagan** (Inspired) — outcome-driven product decisions, not feature-driven
- **Will Larson** (engineering leadership) — what's the smallest set of changes that gets the biggest result?
- **Sarah Tavel** (Benchmark Capital · hierarchy of engagement) — what behaviors should Cohort 2 produce that C1 didn't?
# CONTEXT
Build Sprint Cohort 1: 30 sessions across 30 days, 25 students, 80% shipped, [X]% Sean-Ellis-very-disappointed. Cohort 2 starts [DATE]. I want it to be measurably better, not just iterated.
# COHORT 1 STRUCTURE
[PASTE LIST OF C1 SESSIONS · TITLES + 1-LINE DESCRIPTIONS · OR LINK TO YOUR CURRICULUM DOC]
# WHAT FEEDBACK TOLD ME
[PASTE P04 OUTPUT — especially the top themes + top 3 changes]
# YOUR JOB
Output a Cohort 2 redesign brief with these sections:
## 1. Cohort 2 thesis (1 sentence)
The ONE thing Cohort 2 is built around that C1 wasn't. Should be testable.
## 2. The session matrix (KEEP / REDESIGN / KILL / ADD)
For each C1 session, classify:
- **KEEP** · evidence it worked · don't touch
- **REDESIGN** · evidence it half-worked · what to change
- **KILL** · evidence it didn't earn its slot · what to do with the time
- **ADD** · brand new session · evidence from feedback · what it covers
Output as a table.
## 3. The 3 structural changes (not session-level)
Things bigger than any single session — cohort norms, comms cadence, pricing, format. Each: change + evidence + risk.
## 4. The 1 sacred thing
The thing students protected. Make it MORE central in Cohort 2, not less.
## 5. The Cohort 2 outcome promise
The specific result Cohort 2 grads will have that C1 grads didn't. Should be measurable in a follow-up interview.
## 6. The risks
Top 3 ways this redesign could backfire. For each: early warning signal + mitigation.
# QUALITY BAR
Marty Cagan: "If you're not changing the OUTCOMES you produce, you're not really redesigning — you're just rearranging." Every change must connect to a different outcome.
Expert panel
Bob Moesta · JTBDMarty Cagan · InspiredWill Larson · Eng LeadershipSarah Tavel · Hierarchy of Engagement
P07⏱ 10 min · paste transcripts · get landing-page copy
Testimonial Mine · turn quotes into marketing assets
When to useAfter P03/P04 are done. Mines all transcripts for ready-to-publish marketing quotes — by persona, by format (LinkedIn / landing page / sales deck), by emotional register. The Cohort 2 marketing engine.
# ROLE
You are my marketing + brand panel. Channel:
- **April Dunford** (Obviously Awesome) — to extract positioning language buyers actually use
- **Harry Dry** (Marketing Examples) — to spot the ONE quotable line that converts
- **Justin Welsh** (LinkedIn solopreneur) — to repackage quotes as scroll-stopping social posts
- **Joanna Wiebe** (Copyhackers) — to turn raw quotes into landing-page H1s and proof modules
# CONTEXT
I just finished 25 feedback interviews for Build Sprint Cohort 1. I have permission to use most quotes (tracked per-quote). Now I'm building the Cohort 2 marketing engine. I need to mine the transcripts for every quotable asset.
# DATA
[PASTE ALL 25 TRANSCRIPTS · OR EACH STUDENT'S MOST QUOTABLE SECTIONS · OR P03 OUTPUTS]
# PERMISSION STATUS
[PASTE A LIST: "Sara — full permission · Jake — anonymize · Maya — full permission with name only" etc.]
# YOUR JOB
Output a structured Testimonial Bank with these sections:
## 1. Landing-page H1 candidates (top 5)
Direct quotes (or barely edited) that could be a HERO HEADLINE on the Cohort 2 page. Each: the quote, attribution, persona it targets, why it works.
## 2. Proof modules (top 10)
The specific outcome quotes — named project, named dollars, named timeline. Format ready to drop into a 3-column proof grid.
## 3. LinkedIn post hooks (top 5)
The quotes that would stop scroll on LinkedIn. For each: the quote, the 1-line setup I'd write above it, the CTA below.
## 4. Objection-handler quotes (by objection)
Mine for quotes that defuse the 4 most common Cohort 2 objections:
- "I don't have time"
- "I already know AI"
- "I work alone, I don't need a cohort"
- "Is this worth $X?"
## 5. The persona quote bank
Group quotes by persona type (solopreneur, intermediate dev, non-technical founder, etc.). The Cohort 2 page can show quotes that match the visitor's persona.
## 6. The single quote I should put EVERYWHERE
If I could only use ONE quote on the Cohort 2 page, podcast, sales deck, and LinkedIn — what is it and why?
# QUALITY BAR
April Dunford: "If you can't use a customer quote as-is on your page, it means you're not actually selling what they're buying." Find me the quote that proves I am.
Format the output as a clean Notion-pasteable doc with clear sections.