If you’re running a side hustle or small business in 2026, you’re probably paying for more AI tools than you actually use. One for writing, one for images, one for scheduling, maybe another for analytics. At £15 or £20 a month each, the total creeps up quickly.
A survey of 2,000 US AI users found the average subscriber pays for four AI products at roughly $66 a month, and 53% admit they cancel and restart tools depending on the month. That’s subscription chaos, and it’s costing more than most people realise.
A Simple Cost-Per-Use Test
The quickest way to tell if a subscription is worth keeping is to divide what you pay by how often you open it. A £20 writing tool you use twice a month? That’s £10 per use. A £15 image generator you use every day? About 50p. The maths speaks for itself.
Pull up your bank statement and list every AI tool you’re paying for. Note how many times you used each one last month. Anything above £5 per use probably isn’t pulling its weight.
Then check for overlap. A lot of AI platforms now bundle features that used to need separate tools. Your writing assistant might handle summarisation too, or your project management app might already have built-in scheduling. If two tools do the same job, drop one.
Where Most of the Waste Sits
The biggest culprit is usually the tool you signed up for during a hectic week and forgot about. Annual subscriptions make it worse because they auto-renew quietly. A Zylo report found that companies waste an average of $17 million a year on unused or redundant software. For a solo founder the scale is obviously smaller, but the habit is identical. Even £50 a month in forgotten tools adds up to £600 a year.
Set a calendar reminder every quarter. If you haven’t opened something in 30 days, cancel it. You can always re-subscribe if you actually miss it.
The Cost You Don’t See on the Invoice
There’s another cost that won’t show up on your bank statement. Many AI tools use the data you upload, your prompts, documents, files, to train and improve their models. If you’re pasting in client briefs or financial projections, that information could end up absorbed into a system you don’t control.
The UK’s National Cyber Security Centre (NCSC) has published guidance on AI and cyber security that flags data security as something managers and business owners need to get their heads around before adopting AI tools. It’s a short read and covers the basics well.
Most people don’t think twice about where their files live because the tools make it so easy to drag and drop. But if your cloud folder is connected to a platform that reserves the right to use uploaded content for model training, anything you store there is fair game. Check the terms of service for every tool that touches your business files.
One way to reduce this risk is to keep sensitive files separate from the AI tools you use daily. Encrypted online storage gives you somewhere to store business documents, contracts and anything confidential without exposing it to platforms that might feed it into their training pipelines.
What a Lean AI Stack Actually Looks Like
Most founders and freelancers can cover everything with three or four tools. One strong generalist for writing and research, one specialist tool for your core work, and a scheduling or automation tool to tie it together.
Before adding anything new, ask yourself two things. Does an existing tool already do this? And what happens to your data once you upload it? If something you already pay for covers it, you don’t need it. If the data policy is vague or buried deep in the terms of service, think twice.
Fewer Tools, More Control
New AI tools launch every week, and each one promises to save you hours. But the founders getting the most out of AI aren’t the ones with the longest subscription list. They picked a few tools that genuinely fit their workflow and cut everything else.
Run the cost-per-use test once a quarter. Kill what you’re not using. And keep your most valuable business data away from platforms that treat it as training material.
