AI Builder Challenge

What a winning submission looks like

The judging criteria, walked through with an example entry, so you know what to build and what to hand in.

The example entry on this page is fictional.

How the score is weighted

Revenue carries most of the score. The other three criteria count for 10% each.

  • 70%Revenue & Commercial Viability
  • 10%Problem-Solution Fit
  • 10%AI Utilization
  • 10%Story & Documentation

Example entry (fictional)

MenuLens

AI menu translation for small restaurants in Sofia

Team
Two builders: one develops, one sells
Timeline
Started in May 2026; first paying customer in week 5
Problem
Tourists skip small restaurants whose menus are only in Bulgarian. Owners have no time or budget for professional translation, and their menus change every week.
Solution
Owners photograph their menu and MenuLens turns it into a QR menu in six languages. The owner approves every translation before it goes live.
Pricing
€15 per restaurant per month, after a 14-day free trial
Revenue
9 paying restaurants by November 2026; 7 still paying after their second month
AI tools
Claude for interview notes and translations, Cursor for development, v0 for the first UI, ChatGPT for outreach emails

The four criteria, with the example

Revenue & Commercial Viability

70%

Demonstrated revenue from paying customers. Product-market fit and scalability potential.

What the rules ask for

  • Revenue from paying customers. Free trials and free users don't count.
  • A clear value proposition and pricing.
  • Proof of every payment: invoices, payment-processor screenshots or bank records.
  • Signs the product can grow beyond its first customers.

In the example entry

  • Stripe screenshots of all 9 subscriptions, with customer details redacted, plus two invoices.
  • Explains why owners pay: €15 a month costs less than one lost table of tourists.
  • Shows which trial users converted, which didn't, and why.
  • Maps the next 30 restaurants on the same streets and what it would take to sign them.

Weak spots to avoid

  • Counting free-trial sign-ups or friends' token payments as revenue.
  • A pricing page with no evidence that anyone has paid.
  • Big projections instead of real transactions.

Problem-Solution Fit

10%

Clear identification of a real problem and effectiveness of the solution provided.

What the rules ask for

  • A real problem, clearly identified.
  • Evidence that the solution actually solves it.

In the example entry

  • Notes from 12 owner interviews on one page, including the quote that changed the plan.
  • Owners report that staff no longer translate every dish by hand for tourists.

Weak spots to avoid

  • A solution looking for a problem.
  • Describing features without the problem they fix.

AI Utilization

10%

Strategic use of AI tools throughout ideation, design, development, and marketing.

What the rules ask for

  • Strategic use of AI across ideation, design, development and marketing.
  • AI used meaningfully in at least one stage (a requirement to qualify).

In the example entry

  • A short table: stage, tool, what it did and what it saved.
  • Honest about limits: owners review every translation because the AI mistranslated regional dish names early on.

Weak spots to avoid

  • A long list of tools with no explanation of how they were used.
  • Claiming AI did everything without showing what it did.

Story & Documentation

10%

Quality of documentation, honesty about obstacles, and clarity of lessons learned.

What the rules ask for

  • Clear, complete documentation.
  • Honesty about obstacles.
  • Lessons learned, stated plainly.

In the example entry

  • A dated timeline from idea to first payment, with screenshots.
  • The pivot: they first built a consumer app for tourists, nobody paid, so they sold to restaurants instead.
  • Three lessons in plain language.

Weak spots to avoid

  • A polished pitch deck with no evidence behind it.
  • Hiding what went wrong. The rules reward honesty about obstacles.

What to submit

Every submission needs these six items. Here is how the example covers each one.

  1. 1

    Product description and link

    A live link to MenuLens and a two-paragraph description.

  2. 2

    Proof of revenue from paying customers

    Stripe screenshots and invoices, with customer details redacted.

  3. 3

    Documentation of your build process

    A dated timeline with screenshots.

  4. 4

    Which AI tools you used and how

    The stage-by-stage tool table.

  5. 5

    Obstacles faced and lessons learned

    The tourist-app pivot and three lessons.

  6. 6

    Preferred educational institution for the donation

    The vocational school one of the builders attended.

What disqualifies or weakens an entry

  • Building on a product that existed before the competition. It must be built from scratch between April 2026 and December 1, 2026.
  • No paying customer. Even €1 counts, but €0 does not.
  • Not using AI in any stage of development.
  • Submitting after the deadline: December 1, 2026 at 11:59 PM.
  • Forgetting to redact customer details in payment screenshots.

Ready to build yours?

Registration is free and open until December 1, 2026.

Register nowRead the full rules

MenuLens is a fictional example written to illustrate the published judging criteria. It is not a real entry, and following it does not guarantee a result. The Rules page is the official reference for eligibility and judging.