South Summit 2026 · Madrid · IE University · June 5

From idea to product in 3h.

Today you validate a real idea, build a business case, and ship a working MVP — then pitch it live. All you need is your laptop with Wi-Fi and an idea you would genuinely want to turn into a company.

3h · 2 blocks + live pitches after 💻 Laptop + Wi-Fi · nothing to install 🎯 5 finalists pitch live

The sprint in three blocks

Block 1 · 90 min

Think the idea through

0:00 → 1:30 · Validation · Market · Business model
  • Joe Haslam · theory7'
  • Quico Machín · demo13'
  • Individual work50'
  • Group share20'
 Block 2 · 90 min

Build the MVP

1:30 → 3:00 · Brief · Build · Pitch · Submit
  • Joe Haslam · theory7'
  • IE University · intro5'
  • Individual work55'
  • Group share15'
 Final pitch · 20 min

5 selected ideas

3:00 → 3:20 · Live pitches
  • Selection3'
  • 5 pitches × 3 min15'
  • Wrap-up & assignment2'
Before you start

A quick setup, then we begin

2 min

Nothing to install. Block 1 runs in a browser tab. Block 2 runs in Lovable — the team will walk you through it live.

Before you start

3 steps

2'
  1. Open claude.ai in your browser. Sign in to your account. If you don't have one, creating it takes under 2 minutes.
  2. Open a new conversation. In the model selector (top) choose Claude Sonnet.
✓ Can you see the chat with a blinking cursor? You're ready. The build tool for Block 2 will be introduced live.
Garry Tan, President & CEO of Y Combinator
Garry Tan
President & CEO, Y Combinator

💡 The methodology behind Block 1 — gstack

The prompts you'll use today are based on gstack, created by Garry Tan, current President and CEO of Y Combinator — the accelerator behind Airbnb, Stripe, Reddit and OpenAI. 100,000+ stars on GitHub, free and open source. The prompts reproduce its methodology exactly: design doc, 6 forcing questions, CEO scope-cut.

🔗 github.com/garrytan/gstack

Why this matters: a great habit with AI is to look for a purpose-built skill for the exact job in front of you, rather than a blank prompt. We chose gstack for Block 1 — it's the right fit for idea validation — and pair it with our facilitation experience. In Block 2, Lovable takes that validated brief and turns it into a real product.

Block 1

Think the idea through

90 min · 0:00 → 1:30

Validate real demand, map the competition, size the market and build a business case. No code yet. Today, thinking is the work. Your output: a Founder Brief that drives Block 2.

JH
Theory
Joe Haslam
Professor of Entrepreneurship · IE Business School

Why most startups fail on the idea, not on execution

7'

Joe opens with the most expensive rule in entrepreneurship: "Interest ≠ demand. Money and panic are the only real signals." A quick framework to tell a true idea from a pretty one.

QM
Demo
IE University
IE University · South Summit 2026

How we validated our idea "MarketPilot" in 50 minutes

13'

We'll show you live how we applied the 3 prompts to a real idea and produced a Founder Brief. You'll see the 6 forcing questions answered, the competitor table, the business model, and the final brief. That's exactly what you'll do next.

Individual work

Your turn · 4 steps, 50 minutes

50'

Work in silence. Each prompt reproduces a gstack skill — copy, paste into Claude, respond.

A Validate the idea — 6 forcing questions 20 min

Paste this whole prompt into a new Claude conversation. Replace the last line with your idea and send it.

Act as a Y Combinator partner running an "office hours" session in Startup mode. Apply the anti-sycophancy rules: don't flatter, don't say "interesting", don't soften. Push vague answers until they are specific. Specificity is the currency. Ask me these 6 questions ONE BY ONE. Wait for my answer after each before moving to the next: 1. Demand reality: What is the strongest proof that someone wants this — not interest, not sign-ups, but genuine panic if it disappeared? 2. Status quo: How do they solve it TODAY, badly? (spreadsheets, tools held together with tape, expensive agencies) 3. Desperate specificity: Name the real human. Job title. Company. What gets them promoted or fired? 4. Narrowest wedge: The smallest thing someone would pay for THIS WEEK? One feature. Days, not months. 5. Observation: Have you watched someone use it? What surprised you? 6. Future: In 3 years, is this MORE essential? Why? Push rules: - If I answer "everyone" or "it's for X companies" → push: "that's a category, give me the concrete human". - If I answer "people say it's interesting" → push: "liking it isn't paying; has anyone done it?". - If I answer "nothing like it exists" → push: "so how do they cope with it today?". When the 6 are done, generate a Design Doc in markdown: Problem · Demand evidence · Status quo · Target user and wedge · Assumptions · 2-3 approaches with pros/cons/effort · Recommendation · Real-world task for this week. My idea is: [WRITE YOUR IDEA HERE IN ONE SENTENCE]
✓ Answers that pass
  • "3 marketing managers begged for this; one pays €200/month for a Notion+Zapier hack that breaks"
  • "Marta, 35, sole marketer at a 45-person SaaS, judged on leads and posts/week"
  • "One week of content ready to publish, €29/month"
✗ Answers the AI will reject
  • "Everyone needs this"
  • "Mid-sized healthcare companies"
  • "An all-in-one AI platform"

B Market and competitor research 12 min

Stay in the same conversation. Ask Claude to research the landscape.

Now research the competitive landscape for my idea. Use web search for real data. Give me, in a table: 1. The 5 most relevant direct competitors in Spain and Europe, with their current price. 2. Where each one falls short (one concrete weakness, not a generic one). 3. The real gap that ALL of them leave open — where my wedge can fit. 4. An approximate TAM/SAM/SOM for my target user in Spain (numbers, not vague ranges). Be conservative, not optimistic. If there's no data, say so and suggest how to estimate it.

C Business model and exit thesis 13 min

Based on my Design Doc and the competitor research, propose the business case: 1. Pricing model: with a concrete number in € (monthly, per seat, freemium, …). 2. Realistic 12-month projection: customers and MRR month by month. Conservative. 3. The 3 biggest operating costs with an approximate number (€/month). 4. Key metrics: target CAC, expected LTV, and break-even threshold. 5. Exit thesis: 3 companies that could acquire this in 3-5 years, with a concrete reason for each. Be a skeptical analyst, not an enthusiastic founder.

D Generate your Founder Brief 5 min

Your Block 1 deliverable — a visual one-pager that captures everything you've validated. Share it via URL, print as PDF, and use it as the brief you'll hand to Lovable in Block 2.

Take everything we've built in this conversation and generate a visual Founder Brief as a self-contained HTML Artifact in a new window (not inline in the chat). Structure it as a clean, print-ready one-pager with these sections: 1. Cover: project name · wedge in one sentence · today's date 2. The Problem: status quo (3 bullet points, concrete and specific) 3. The User: name · role · company size · what gets them promoted or fired 4. The Wedge: the one thing someone pays for this week · price 5. Demand Evidence: the strongest proof point from our conversation 6. Market: TAM / SAM / SOM with numbers, not ranges 7. Business Model: pricing · 12-month MRR projection (3 milestones) 8. Competitor gap: compact table — 3 competitors + the gap we own 9. This week's experiment: one concrete, verifiable action before Friday Design: polished and professional. Navy blue (#000066) header, white background, Montserrat font via Google Fonts. The Artifact must be fully self-contained with all CSS and HTML inline — no external dependencies.
✓ What you'll have: a shareable Artifact you can download as HTML and use as the build brief for Lovable in Block 2. Save the HTML file — it's your portable version of the brief.
📄 To export it as a PDF: open the saved HTML file with a double-click (it opens in your browser), then right-click → Print → Save as PDF. That gives you a clean PDF you can share or attach anywhere.
⏱ Time check: 35 min into this sub-phase, we'll remind you out loud that 10 minutes are left. If you're stuck on the first question, type: "Take my best hypothesis and move on; I'll validate it with real customers this week."
Optional · Block 2 prep

Generate your Lovable build brief

+5 min

Only if you finish Step D early. This generates the structured text brief you'll paste into Lovable at the start of Block 2. Do not build anything here — building is Block 2's job.

★ Lovable build brief Optional

Paste this in the same conversation. Claude will generate a structured brief — copy the output and keep it ready for Block 2.

Based on everything we've validated, generate a structured build brief — not an Artifact, just plain structured text I can copy and paste into a product builder. Use this exact format: PROJECT NAME: [name] ONE-LINE WEDGE: [what someone pays for this week, in one sentence] TARGET USER: [name · role · company size · what gets them promoted/fired] THE PAIN: [current status quo — how they solve it today, badly] THE MAGIC MOMENT: User inputs [X] → sees [Y] in [time] 3 KEY SCREENS (maximum): 1. [screen name]: [what it does in one sentence] 2. [screen name]: [what it does in one sentence] 3. [screen name]: [what it does in one sentence] BRAND TONE: [3 adjectives — e.g. direct, data-forward, no jargon] DESIGN VIBE: [reference — e.g. Linear, Stripe, Notion] PRICING: [model + price] Keep every field to 1-2 lines. Be concrete, not generic.
⚡ Keep this text. In Block 2 you'll open Lovable, paste this brief, and the Lovable team will guide you from there. The building happens entirely in Block 2 — don't start now.
Group share

Share your wedge in 60 seconds

20'

In a quick round, each participant reads out loud: their wedge + the price + the concrete user. We challenge 2-3 with quick questions to push specificity. Hearing other ideas helps you sharpen yours before Block 2.

⚡ Before Block 2

Prepare your Founder Brief for the build

Block 2 is guided live by the Lovable team. Your Founder Brief (the one-pager from Block 1 Step D) is your reference document. Have it ready in one of these forms:

Option A — Browser window

Keep the Founder Brief Artifact open in a browser tab. You can reference it during the build phase without copying/pasting.

Option B — HTML file (backup)

Download the Founder Brief HTML and save it locally. If the Artifact link expires, you have a backup to share or reference.

Bottom line: Both work. Option A is simplest. Option B is safer. Pick whichever suits you.

Block 2

Build the MVP

90 min · 1:30 → 3:00

You'll turn your Founder Brief into a working product. Joe opens with the core principle, the IE team frames the approach, and then the build session begins — guided live.

JH
Theory
Joe Haslam
Professor of Entrepreneurship · IE Business School

The art of cutting: why an MVP should hurt

7'

Joe explains why most founders build too much and show too late. The rule: "One screen that does one magic thing beats ten half-finished screens." The brief you hand over should be ruthlessly specific.

QM
IE team
IE University
IE University · South Summit 2026

From Founder Brief to product — what good looks like

5'

A quick look at what your build brief unlocks and what a strong MVP output from this session looks like — so you know what to aim for in the next 55 minutes.

LV
Build partner
Lovable team
Live session facilitation · product build

Tool intro + how this session works

7'

The Lovable team introduces the platform and explains how the build session is structured. They'll guide you through every step from here — raise your hand if you get stuck.

Individual work

Build · pitch · submit — guided live

55'

The Lovable team leads this section entirely. Your only references here are the pitch prompt and the submission step — everything else follows their live instructions.

★ Prepare the 60-second pitch 8 min — when the build wraps

When the Lovable team signals the build is closing, switch to Claude and use this prompt.

Help me write a pitch of exactly 60 seconds for my MVP using this template: "For [concrete user] who [concrete problem], [project name] is a [category] that [key benefit]. Unlike [status quo], we [the differentiating wedge]." After that sentence, structure the rest of the minute as: - What I'll show in the live demo (which action I'll perform on screen) - My demand evidence + the price I'll charge - The concrete experiment I'll run THIS week to land a real customer Make it concise, direct, no jargon. End with energy.

★ Submit your idea to the platform 5 min — before 2:55

To compete for the 5 final pitches, submit before minute 2:55. You need: wedge + project name + your app's public URL + your name.

🔗 Submission platform: southsummit-ie-2026.netlify.app/enviar.html?s=sesion-2
Group share

We share screens, we share results

15'

Participants with something working share their screen for 30-45 seconds each. Not the final pitch — it's proof the whole flow worked. We note the most promising ones for the 5 finalist spots.

Final

5 live pitches

20 min · 3:00 → 3:20

The IE team announces the 5 finalists selected from submissions. Each has exactly 3 minutes to demo their product live and sell the idea.

Pitch rules

3 minutes · strict timer

BlockTimeWhat to do
Opening0:00 – 0:30Your wedge in one sentence. User + problem + differentiator.
Live demo0:30 – 1:45Share your screen. One single magic action. Nothing extra.
Model and market1:45 – 2:30Price · the gap you fill · demand evidence.
The experiment2:30 – 3:00What you'll do this week to land a real customer.
📺 Screen sharing: connect your laptop to the projector via HDMI or AirPlay. Have your Lovable app already open and maximised.
⏱ Timer: at 3:00 we cut you off. No extensions. A clean 2:45 beats a rushed 3:30.

Selection criteria for the 5 finalists

  • Clear wedge — understandable with no prior context
  • Real demand evidence — not "I think…"
  • A working product — the app is usable, not a wireframe
  • Coherent business model — a number, not a vague range
  • Founder track record — is this person already doing something about it?
Bonus

Choosing the right model

for Block 1 and the pitch

Claude is your Block 1 thinking partner. For Block 2, the build tool will be introduced live — Claude returns only for the pitch step at the end.

Claude model guide

PhaseModelWhy
B1 · Validation + market + business + Founder BriefSonnetNeeds judgement to push vague answers and research competitors
B2 · Pitch prep (when the build wraps)HaikuShort text — faster and more direct for a 60-second script
💡 Rule of thumb: use Sonnet for reasoning and research in Block 1. Switch to Haiku when writing the pitch — it's faster for short, punchy text.