Make the case with evidence, then show it working
Tonight is demo day. You make an honest business case, test it against its toughest critic, and show your product to people who matter.
By the end of tonight you can…
- Write an evidence-led business case: problem, evidence, solution, traction, economics, risks, ask
- Size your market bottom-up from people you can reach and what they pay
- Keep a risk register with likelihood, impact, mitigation and an owner
- Write a plain-language responsible-AI statement, or explain why you chose not to use AI
- Present an eight-minute demo and plan loop 3 with a weekly operating rhythm
OUTPUT · Business case + responsible-AI statement + eight-minute demo
A business case in seven parts
- PROBLEMWho struggles, when, and what it costs them, in one paragraph
- EVIDENCEWhat you observed: interviews, retention, quotes, each with its base
- SOLUTIONWhat the product does, shown at the first value moment
- TRACTIONWho uses it, who stays and who pays or commits, as counts with bases
- ECONOMICSPrice, cost per user including AI, margin and market size, labelled
- RISKSThe top five from your risk register, with mitigations
- ASKWhat you want from this audience, and what it will let you prove next
Every claim points to evidence in your Product Book. If it doesn’t, it is an assumption: label it.
From “a huge market” to an evidence-led case
“Tutoring is a huge market. If we win just 1% of it, TutorDesk will be worth millions.”
Illustrative: 9 of 12 busier tutors still booking in week 4; 6 signed pilots at £8 a month; AI cost about £0.24 per tutor a month. Ask: two community introductions, not money yet.
Ask
- What does “1% of a huge market” assume without saying?
- Is a signed pilot as strong as a payment? What is missing?
- Why might asking for introductions beat asking for money now?
- Which number here would you check first, and how?
Size the market bottom-up
- REACHABLEPeople you can name a route to: say 2,500 tutors in communities you can join
- BEST FITThe share that matches your best-fit customer: 40% with 15+ learners = 1,000
- PRICEWhat each pays a year, from your pricing test: £8 × 12 = £96
- MULTIPLY1,000 × £96 = £96,000 a year: the ceiling through today’s channels
- REALISTIC SHAREYear one at 5%: 50 tutors × £96 = £4,800. Every input labelled
A small number you can defend beats a big one you can’t. Show every input and where it came from.
A risk register: five columns, one owner each
- RISKCause and effect: “exam season ends, so busier tutors pause and stop paying”
- LIKELIHOODLow, medium or high, with the evidence behind the rating
- IMPACTWhat it would cost if it happened: revenue, trust, time or safety
- MITIGATIONWhat you are doing now, and the signal that triggers more action
- OWNEROne named person and a review date. At this stage, usually you
Include the AI feature’s risks. A risk with no owner is a worry, not a plan.
Red-team your business case with extended thinking
Here is the business case for [MY PRODUCT]. I’ll present it on demo day to mentors, potential customers and investors, and I want to find its weakest points first so that I can fix them or state them openly. <case>[PASTE BUSINESS CASE]</case> Read it as a sceptical investor would, then as a busy [TARGET CUSTOMER] would. List the five weakest points, ranked by how much they would damage trust. For each, quote the sentence, say why it is weak, and say what evidence or plainer wording would fix it. Recalculate every number and show your working. If a weakness depends on something the case doesn’t say, say so rather than assume. Done means the ranked list plus the three questions this audience is most likely to ask.
Check the response
- Every number was recalculated, with the working shown
- Each weakness quotes the exact sentence it attacks
- Fixes add evidence or honesty, not stronger wording
A responsible-AI statement in plain words
- WHAT IT DOESReads a parent’s message and proposes a free slot for the tutor to approve
- ITS LIMITSCan misread dates or miss a constraint: right in 17 of 20 eval cases
- THE DATASends only the message and the tutor’s free slots; parents are told AI reads it
- PEOPLE IN CONTROLNothing is booked until the tutor approves; they can edit, ignore or switch it off
Write it for a parent, not an engineer. Internal-tool path: cover that tool, or say why no AI.
The loop never ends: MVPAI as a weekly rhythm
- MAPTalk to two users weekly
- VALIDATEPick this week’s one bet
- PLANSet this week’s objective
- ASSEMBLEShip small: tests, evals
- IGNITEFlag it, read the numbers
Loop 3 starts on Monday. A weekly rhythm keeps you honest once nobody is marking.
Finish the case and rehearse your demo
A business case a sceptic would trust, and an eight-minute demo that shows the product working, live.
- Fix the red team’s top three weaknesses with evidence or honest wording (12 min)
- Write your responsible-AI statement, or your reasons for not using AI (8 min)
- Script the demo: problem, live demo, traction, economics, risks, ask, next (8 min)
- Rehearse in pairs with a timer; your partner notes one unclear claim (12 min)
Business case and responsible-AI statement: product/12-business-case.md
Final capstone: eight-minute demos and business case
- The demo shows the first value moment and the AI feature or internal tool live, in eight minutes
- Retention and willingness to pay appear as counts with bases, against thresholds set in advance
- Economics show price, cost per user including AI, and bottom-up size, labelled real or illustrative
- AI evidence: eval score against threshold, injection tests, fallback; or a reasoned no-AI decision
- The risk register names top risks with owners, and the red team’s hardest point is answered
- Next steps name loop 3’s first question, the ask, and the result that would change the plan
NEXT · Loop 3 starts on Monday: run the weekly rhythm and send your guests an update in four weeks.
© 2026 MVP Studios Ltd. MVPAI™ and the MVP Founders Programme materials are original works of MVP Studios Ltd. All rights reserved. No copying, redistribution or commercial use without written permission. Course terms