Pick a depth. Each prompt opens in your AI pre-loaded with the lesson. Click a row to preview the prompt.
The single most common reason people never start freelancing is 'I don't have a skill anyone would pay for,' and it is almost always wrong. You undercount because the things you're good at feel easy to you — the friend who always asks you to fix their resume, the spreadsheet you built for your last job that everyone copied, the fact that you can write a polite email that gets a refund. Those are sellable. This task forces a complete inventory before you judge anything, because you cannot pick the best option from a list you never wrote down. Do this badly and you'll default to the one obvious skill and miss the one that's actually in demand; do it well and you walk into the next task with fifteen to twenty candidates instead of one.
Work through four buckets and write everything, no filtering yet. Bucket one, jobs: every role you've held and the actual tasks inside it — 'answered support tickets,' 'made the weekly sales deck,' 'scheduled the social posts.' Bucket two, hobbies: 'edit my own travel videos in CapCut,' 'run a Dungeons and Dragons group,' 'bake and photograph cakes.' Bucket three, favors people ask you for: this bucket is gold, because it's demand that already found you — 'friends ask me to proofread cover letters,' 'my mum asks me to set up her phone,' 'people ask how I got my Instagram to 5k.' Bucket four, things you learned recently: 'I figured out Notion,' 'I can prompt Midjourney decently.' Then hand the raw list to an AI to find adjacent skills you skipped. A real prompt: 'Here is my skills list: [paste]. For each item, suggest two adjacent services someone might pay for that I probably already can do or could learn in a week. Be specific, not generic.' If you wrote 'made the weekly sales deck,' it might surface 'pitch-deck cleanup for founders' and 'turning ugly PowerPoints into clean templates' — services with real buyers you'd never have named yourself.