finlit · The Capital Mind presents
Three weekends. One brief. Use AI to design, model, and pitch a real financial solution — then defend it live to industry judges. Top 20 teams present at the global final.
"You have one week. Use AI to design, model, and pitch one financial solution that solves a real problem for people your age — anywhere in the world. Show your AI Workbook: every prompt used, every output accepted, every output rejected, and why. Then defend it live."
The full brief is revealed at the Weekend 1 launch session. Every team hears it at the same moment. Build window opens immediately after.
This is not a quiz or a classroom exercise. You get a real brief, real AI tools, and one week to produce something a judge will interrogate live. No prior finance or coding experience required.
A mentor and industry expert reveal the brief live. Rules, AI Workbook format, judging criteria — all explained. Build window opens the moment it ends.
You have seven days. Build your solution, stress-test your model, complete your AI Workbook. Submit before Friday midnight.
Top 20 announced. Semis on Weekend 2. Top 20 all present at the global final on Weekend 3 — two live rooms, one panel, one winner.
Enrol your class, year group, or finance club as a block. Schools with five or more registered students receive a dedicated teacher dashboard, progress updates, and a school leaderboard at the finals.
A single $60 entry fee per student covers the brief, the AI Workbook, judging and the global final. No software costs, no subscriptions, no school licence. Schools with sponsored seats can register fee-waived cohorts.
Mentors from the finlit Harvard programme are available during office hours in the build week. Students who complete the competition receive a participation certificate.
Schools with finalists are named in all press coverage and at the live finals event. The finals are live-streamed — parents, teachers, and press all watching.
The full brief is revealed at the launch session. These four tracks frame the problem space. Teams choose one and go deep.
Design a tool that helps a young person manage money across two currencies, two countries, or two financial systems. The founding use case of modern neobanks — built by your generation.
Build an AI-powered budgeting or savings model for a student on a fixed income. Must include a stress-test scenario — what happens when income drops 30%?
Build an explainer that helps a peer evaluate a financial product — BNPL, crypto, or subscription trap — before signing up. Surface the real cost and the hidden risk.
Design a feature for a digital bank's under-18 product that does not yet exist. Include a product brief and a year-one adoption model.
The competition runs across three consecutive weekends. Submissions open and close in between. Every deadline is hard.
Every team submits a full prompt log — every AI interaction, every output accepted, every output rejected, with one-line reasoning. Judges score it. It is 25% of your grade.
Every AI interaction from research to pitch is documented. The Workbook is submitted alongside your brief and model. Judges read it.
Each output is marked Kept, Modified, or Rejected with one-line reasoning. Judgment is graded — not just output quality.
A team that catches one AI error and corrects it will outscore a team that accepted 14 outputs without question. Every time.
Prompt 9 of 11 · Tool: ChatGPT-4o · Phase: Modelling — "Build a BNPL repayment schedule for a $200 purchase at 0% headline rate. Calculate the true APR if a $5 late fee applies to each missed payment."
Four criteria. 25% each. Scored across your submission and your live pitch. AI Workbook quality is worth as much as the model itself.
Correct formulas. Stated assumptions. Honest stress-test that reveals something real — not an optimistic scenario.
Which outputs did you trust and why? Which did you discard? One corrected AI error outscores ten uncritical acceptances.
Is the problem real? Does the intended user exist? Does the solution address the actual mechanism — not just a symptom?
Can a judge with no background follow your argument in five minutes? Is the problem named? Is the ask clear?
No cut after the semifinal. Every top 20 team presents at the global final. Two parallel rooms, one panel, one winner announced live.
Individual entry. Grades 5–12. Any country. $60 per student. Cohort 1 starts 29 Aug 2026.
Open the application →Schools with 5+ registered students get a teacher dashboard, real-time submission tracking, and their school named at the finals. No admin overhead.