The most common place people get stuck with AI income isn't a skills problem. It's a starting problem. They've done the research, they understand the tools exist, they believe the opportunity is real — and then they do nothing, because they don't know what the first actual step looks like.
This article is specifically about that first step. Not the grand strategy. The thing you do this week.
Start with what you already know
AI income works best when it combines with existing knowledge you already have. If you've spent five years in marketing, AI makes you a faster, more capable marketing professional. If you've done any writing, any admin work, any customer service — there's a version of AI-assisted freelancing that builds on that.
The mistake is trying to start from scratch. "I need to become an AI expert and then I'll find clients." That's backwards. The people who pick things up fastest lead with their existing knowledge and add AI as a layer on top.
Pick one service type and do it badly once
This sounds like odd advice but it's the most useful thing you can do. Pick a simple, specific AI-assisted service — a blog post, a product description, a social media caption set — and produce one. Not for a client. Just to see what the workflow actually looks like from start to finish.
You'll hit problems. The AI output won't be quite right. You'll edit it and it'll take longer than you expected. That's fine — and it's genuinely useful, because now you know what the real time cost is. People who've never done this tend to either massively underestimate or overestimate how much work is involved.
The portfolio problem (and how to actually solve it)
New freelancers worry about portfolios. They think they need five previous clients before they can approach a sixth. This isn't true — it's just a common anxiety that stops people starting.
The practical solution is to create three to five speculative samples. Write a blog post for a type of business you'd like to work with, even if they're not a real client. Create a sample email sequence. Write product descriptions for a hypothetical product. These are real, demonstrable work — the fact that they weren't commissioned doesn't matter in an initial pitch.
Where to find the first client
In the UK freelance market, the practical starting points are:
- Your immediate network. Small businesses owned by people you know, or people once removed. Most small UK businesses don't have anyone managing their content — and if they know you and trust you vaguely, that removes a lot of the initial sales friction.
- Freelancer platforms. PeoplePerHour is UK-focused and worth starting with. The rates aren't high at first — but the first project isn't about the money, it's about getting the workflow proven and getting a review.
- Local business groups. Chambers of commerce, LinkedIn local groups, small business Facebook groups. Small business owners talk to each other and they're usually looking for affordable help with things they don't have time to do themselves.
What to charge for your first project
The answer is lower than you'd like, and that's okay. First projects are about proving the workflow, getting feedback, and getting a written review you can use. Charge enough that it's not embarrassingly cheap, but don't price yourself out of the first engagement.
For a UK freelancer starting out, something in the range of £50–£100 for a standalone content piece is reasonable for an initial project. Once you have the first review and the workflow proven, you can price for actual value.
The mindset shift that makes this work
AI income isn't a product you sell — it's a capability you have. You're not selling "AI writing". You're selling good, timely, well-edited content. The AI is how you produce it efficiently. That framing makes pitching easier and positions you as a professional rather than someone who's just pressing buttons on a chatbot.