AI Guides › Playbooks
By Nigel Guy · 7 min read
Most people choose a side hustle by looking at what is loudest on their feed, then spend three months building something nobody has agreed to pay for. It feels productive because there is always a logo, a channel or a course outline to tweak. The routes that pay quickly are usually the ones where a person you already know hands you money for a specific job, and those are rarely the glamorous ones.
The rule: choose the route where someone you can already reach has a problem you can fix this month, and sell one small, priced job before you build anything.
This guide covers four common routes. Speed to first payment below is a judgement, not a measured figure, because it depends entirely on who you know.
| Route | Who pays you | Usual speed to first payment | Main catch |
|---|---|---|---|
| AI help for small businesses | A local business owner with a repetitive task | Fastest, if you have a network | You are selling trust, and you may handle their data |
| Building websites | A business that needs a page or a fix | Fast, but crowded | Price pressure; tools make the build easy, so the build is not what you are paid for |
| Teaching AI | A workplace, club, or group of beginners | Moderate | Needs a room, an audience and a clear outcome |
| Faceless content | A platform, later, and indirectly | Slowest | No customer until you have an audience |
The last row matters. YouTube's own help pages say partner-programme eligibility needs 1,000 subscribers plus either 4,000 qualified public watch hours in 12 months or 10 million qualified Shorts views in 90 days. Check the current page for the UK before you plan around it. Its policy on "inauthentic" content (renamed in July 2025) also names mass-produced content made from generic AI templates as something it does not reward. Faceless content is not impossible, but it is the one route with no customer on day one.
Score each route from 0 to 3 on five questions, using only facts about you. Do not score what you hope to become.
Add them up. A route scoring under 8 out of 15 is not your first route, however good it sounds. If two routes tie, pick the one with the higher Reach score. Reach beats everything else early on.
| Days | Action | Output |
|---|---|---|
| 1-3 | Run the scorecard. Pick one route only. | One route, written down |
| 4-7 | Define one small paid job with a fixed price and a clear finish | A one-sentence offer |
| 8-14 | Contact your three reachable people individually | Three conversations, not three broadcasts |
| 15-24 | Do the first paid job or run the first paid session | A delivered result and a message asking for a short, honest comment |
| 25-30 | Review what took longer than expected and reprice | A second offer, adjusted |
A fixed-price small job beats an hourly "AI consultant" label. "I will turn your weekly customer enquiries into a drafted-replies spreadsheet for a set fee" is a job. "AI consulting" is a category.
Priya works shifts at a hospital, has about five free hours a week, and manages a friend's bakery Instagram. Her scorecard: small-business AI help scores 12 (she knows three owners, has done the bakery work, hates cold selling but can message friends). Faceless content scores 5 (no reach, nothing to show yet). She picks the first route, offers one job: "Drafted replies and a weekly order summary from your messages, set up in one session for a fixed fee." She messages the bakery owner and two others. Nothing here is a forecast; it shows how the scorecard narrows the choice before any money is spent.
Paste this into your AI chat tool. Fill in every square-bracket placeholder; the more specific you are, the less it will guess.
Role: You are a practical advisor helping me choose a side-income route in the UK. You are not a cheerleader.
Context about me:
- Skills and past work: [SKILLS_AND_PAST_WORK]
- Hours I can give each week, and when: [HOURS_PER_WEEK_AND_WHEN]
- Money I can afford to risk: [BUDGET_IN_GBP]
- How comfortable I am asking for money (1-5): [SELLING_COMFORT]
- People I could contact in the next two weeks (roles, not names): [WARM_CONTACTS]
The four routes: AI help for small businesses, building websites, teaching AI, faceless content.
Goal: a ranked recommendation I can act on this week.
Steps:
1. If any input above is blank or vague, ask me for it before continuing.
2. Score each route 0-3 on Reach, Proof, Time fit, Selling comfort and Cash risk. Show the scores in a table with one line of reasoning each.
3. Rank the routes and say which one I should not start yet, and why.
4. For the top route, describe one small job with a clear finish that I could sell first, and suggest a price range in GBP with the reasoning. Say which parts of the price are a guess.
5. Draft the first message I would send, and to which type of person. Keep it under 90 words and make it ask a question rather than pitch.
Constraints: no flattery, no income promises, no invented statistics. Label anything you are assuming as "assumption".
Self-check before answering: confirm that every score uses only facts I gave you, that the price is flagged as an estimate, and that the message contains no claim I could not back up.
Fill in: the five bracketed items about you. Then argue with the scores; you know your contacts better than the model does.
These are UK rules as published on GOV.UK at time of checking.
This is not tax advice. If your situation is not simple, ask an accountant or use HMRC's own guidance.