Apply for Next Intake

Muhammad’s gift-genie

DiscoverSpot problems worth solving, talk to real people and choose the one idea worth building.
ValidateSize the market, study competitors and prove people want it before building.
You are hereBuildBuild one complete journey for one user, by hand first, then sped up with AI.
LaunchPosition it, reach real people and get the first users on a live product.
SellMake the first sales ask, learn the numbers and run a growth experiment.
Goal: GrowLead under pressure, understand investment and risk, and pitch on Founder Day.
Muhammad14 years oldTBC, UK

The problem

Buying a gift for someone often turns into an hour of scrolling online shops with no real starting point, especially for people who find it hard to decide, and it usually ends in either a generic default gift or no gift bought at all.

The solution

GiftGenie asks four quick questions, who the gift is for, the occasion, their interests, and a budget, then returns a short list of specific, buyable gift ideas instead of an endless scroll of options.

Monetisation service

Most revenue comes from an affiliate commission (5-6%) on gifts bought through a suggested retailer link, the same model real gift-recommendation and shopping-discovery apps use. Free for unlimited searches with affiliate links on each suggestion. A £1.99/month Plus tier removes ads between questions and saves an occasion calendar so reminders arrive a few days before birthdays.

Minimum projected ARR

£23,000Year 2
£112,000Year 4
£340,000Year 6

Muhammad’s story

Ashfi is, by his own description, genuinely bad at deciding what to buy people, and GiftGenie started as a way to solve that specifically for himself before he realised plenty of other people have the exact same problem every birthday and holiday.

In the Make It a Real App phase, Ashfi kept the whole flow to four short questions on purpose, since the point of the app is to end a long, indecisive search quickly rather than replace it with a different kind of scrolling. He used AI to help build the quiz screens and results layout while deciding the actual questions and their order himself.

The hardest part was picking exactly four questions instead of ten. Every extra question made the results slightly more accurate, but also made the app feel like exactly the kind of long decision it was supposed to avoid.

Tech stack
pythonclaudesnackreact

What Muhammad’s parent thinks

“TBC”

Book a Free Consultation