Reskilling for freelancers: how to stay ahead of AI and hold on to your clients

AI can now write copy, generate images, and draft code in seconds, leaving many freelancers worried they’ll be outcompeted by tools that work faster and cost less. 

For many, the answer to ensure consistent work is to upskill, but that raises its own questions: what should you learn, how do you fit it around paid work, and how do you avoid wasting time and money on the wrong thing? 

The experts we spoke to say you don’t need to quit work or chase every new trend. Here’s how freelancers can choose the right skills to build, learn them on live projects, and avoid the costliest upskilling mistakes.

Unsure about where to reskill? You may already have the answers

The temptation is to follow the trend, but the experts we spoke to say the best direction is usually already sitting in your client list.

Careers expert Drew Povey from Drew Povey Consultancy told us that freelancers should “look for the overlap between three things: what are you already good at, what are clients increasingly willing to pay for, and what genuinely interests you”. That intersection, he says, “is much more sustainable than chasing every new tool or job title”.

Your invoices are another great place to find clues. Dat Ngo, a Certified Public Accountant at Vetted Prop Firms, told us that before looking at any trend report, you should ask yourself, “Which clients pay fastest, argue least, and come back?”

Start with adjacent skills

Build toward adjacent skills those clients already need. Reskilling into something with no overlap with your existing book of business means starting your client acquisition from zero at the same time you’re starting your learning curve from zero, and that’s two cliffs at once.

Dat Ngo Certified Public Accountant at Vetted Prop Firms

Your clients’ requests are another clue. Connor Gillivan of Trio SEO recommends picking a direction based on “what your current clients keep asking you for and don’t have anyone to solve”. Harvey Dhillon, founder and CEO of UK accountancy firm Zmartly, agrees: “Start with what your current clients pay for. That is usually the cheapest place to reskill.”

Don’t have a long invoice history to go down? Andrew Gosselin, CPA at Save My Cent, says to start by asking directly. Identify the clients who paid the highest effective hourly rate, “not the biggest totals”, he tells us, then ask what adjacent skill would let you sell more to them.

Learn on the job: how to build new skills while you keep earning

With many freelancers facing income volatility and living paycheque-to-paycheque, taking time out of work to develop their craft may sound like a luxury they can’t afford. However, the experts we spoke to explain that you don’t need to pause your income to reskill.

As Dat Ngo explains: “Keep 70 to 80 percent of your paid work running while you train in the margins, and hold at least three months of essential expenses in cash before you reduce anything. 

Matt Rouif, CEO and co-founder of Photoroom, takes it one step further, asserting that the best place to train is inside paid work. “The learning happens inside paid work, on a task where you already know what good looks like,” he tells us. Pick one job you do every week and use it to stretch into the new skill.

You can also learn on a live account. Connor’s advice to freelancers is: “Don’t quit anything.” Instead, offer one existing client the next layer of work, such as strategy, alongside your normal work for a quarter. 

“You learn on a live account, you get a case study, and you usually end up raising your rate with that same client afterwards,” he says. That’s a far better use of three months, he adds, than “a certification nobody asks about”.

Know when to use AI and when to trust your own judgement

Careers expert Drew Povey says the most effective way for freelance creatives to “AI-proof” their careers “is not to compete with AI on speed or volume”. 

Instead, they should become “the person who knows how to use AI intelligently while bringing the judgement, creativity, taste, relationships and specialist expertise that technology cannot easily replicate”.

That judgement is what clients pay for. As Matt from Photoroom explains: “Clients can generate a thousand images themselves now. What they can’t do is work out which of those thousand is actually usable.” 

Photoroom tested 4,250 AI-generated product images and found the strongest model passed a full product-fidelity check only 29% of the time. “Spotting the other 71%, and knowing why each one failed, is the job,” he says.

It’s a view shared by people who use the technology every day. Software engineer Quentin Claudet, who has reskilled in AI, tells us the technology hasn’t stolen his job, as he still feels “useful in the process, in the way of thinking”.

He admits leaning on it too heavily can be slippery slope, however. His advice to freelancers navigating AI?  “Don’t get lost in the tool either. Just use it, and use it well.”

The takeaway from the experts is that AI is best treated as a fast first draft, rather than a finished product. Let it handle the volume, but keep the decisions about quality, taste, and what a client actually needs for yourself. Ultimately, the skills worth building are the ones that make your judgement harder to replace.