Use vibe coding to escape the SaaS you've been stuck with

There's a tool in your stack that everyone complains about. You know the one. It costs more than it should, does less than it claims, and nobody can remember why you bought it. In one business I spoke to recently, it was an expenses platform: £180 a month, used for a single approval workflow that could have been a spreadsheet.

The reason it stays is the same every time. Replacing it feels like a project. Someone has to evaluate alternatives, someone has to migrate the data, someone has to train the team. The cost of leaving looks higher than the cost of staying.

Vibe coding has changed that calculation.

What vibe coding actually is

Vibe coding means using AI code assistants — tools like Cursor, GitHub Copilot, or Claude — to write functional software from plain-English descriptions. You describe what you want. The AI writes the code. You test it, refine it, and iterate.

You don't need to know how to code. You need to know what you want the software to do.

That's a different kind of skill, and it's one most founders and operators already have.

The experiment worth running

Pick the weakest SaaS in your estate — the one with the worst value-for-money profile. Now ask a simple question: what does this tool actually do for us, day to day?

Strip it back to its core function. Most tools that cost hundreds of pounds a month are doing one or two things that actually matter. Everything else is feature bloat you're paying for and not using.

Spend an afternoon describing that core function to an AI coding tool. See what it builds. Don't aim for production software — aim for a working prototype that does the job.

If the prototype works, you've answered the most important question: is this replaceable? The answer is often yes. And that makes the conversation about cancelling much easier.

What this is not

This is not a recommendation to vibe-code your way out of every enterprise system. Some tools are complicated for good reasons. Some carry compliance obligations, data residency requirements, or integration dependencies that make a quick replacement genuinely risky.

This is a recommendation to run a cheap experiment before resigning yourself to paying for something you don't like.

The worst outcome is you spend a day building something that doesn't work and you learn your SaaS is actually earning its fee. That's still useful information.

Looking for a reason to cancel a SaaS contract? Get in touch