Every time a tool adds a new tier, emails appear in my inbox with the same question: “Should I upgrade?” And every time, my answer starts with: what are you actually using the current one for?

Last week it was ChatGPT (we’re currently in August 2026 at the time of writing). OpenAI has split its Business plan into two seat types: a Standard seat at £20 (ish, depending on billing) and a Premium seat at five times the price, built for people who are hitting usage limits and need the extra capacity. It’s a sensible move on their part. But the reason I want to talk about it isn’t really about ChatGPT at all. It’s about what this kind of split reveals.

Because I see the same pattern over and over, in businesses of every size. Someone is on the Premium plan of six different tools in the business tech stack. A project management system, an email platform, a scheduling tool, an AI assistant, a CRM, maybe a design tool too. And when you actually look at how much of each one is being used, it’s often somewhere around forty percent. Not because the business doesn’t need support. Because the upgrade felt like progress, and cancelling or downgrading felt like admitting something wasn’t working.

I understand the instinct. When you’re building something and money is coming in, it’s tempting to believe that more capability in your tech equals more growth. Better tools, more features, and the top tier surely means you’re taking your business seriously. But a tool doesn’t make your business more serious. How you use it does.

What OpenAI’s split invites you to do

The thing I like about this particular update is that it’s honest about something most subscription pricing tries to hide: not everyone on your team needs the same thing.

A seat split like this is a small gift in disguise, because it nudges you to ask a question. Who on your team is genuinely a heavy user, running long sessions, hitting daily limits, relying on the tool for real output? And who is dipping in a few times a week for something straightforward? Those are two very different needs, and they don’t need to be paying for the same plan.

Most software doesn’t ask you this up front. It just has one price and one set of features, and you either pay for it or you don’t. So the decision gets made once at signup, and then it sits there unquestioned for a year or two. Nobody goes back and checks whether the assumption is still correct and everyone needs everything!

The audit nobody wants to do

If you haven’t looked properly at the tools you use in a while, here’s what I’d suggest, and it doesn’t need to take longer than an hour.

Pull up your bank statement or your card transactions and list every recurring business charge. Not from memory. From the actual statement, because memory is where subscriptions go to hide. Next to each one, write down what it’s for and whether anyone, you included is using it. Then, and this is the bit people skip, write down how often.

You’ll usually find three categories. Tools that are genuinely earning their keep and probably deserve the tier you’re paying for. Tools that are being underused, where a lower tier or a different plan would cover exactly what you need. And tools you’d forgotten you were paying for altogether, which is more common than anyone likes to admit.

That third category is the one to pay the most attention to, as it’s not really about the money, though the money is nice to get back. A forgotten subscription is usually a decision that was made in a rush, under the belief that this tool would solve something, and then never revisited once the initial problem had moved on or been solved a different way.

The real cost isn’t just financial

The cost of an oversized tool stack isn’t just the subscription fees. It’s the mental load of managing systems you don’t fully need, the time spent learning features you’ll never use, and the false sense of infrastructure that comes from having a lot of tools without necessarily having the right ones, set up the right way.

A business with three tools that are properly integrated and used will run smoother than a business with nine tools half used and never quite talking to each other. That’s operations 101, but it’s easy to lose sight of when every new release comes with a shiny “upgrade now” button.

This doesn’t mean the top tier is never right. Sometimes it really is. If you’re the person hitting usage limits every day on AI, if the extra capacity or features solves a real bottleneck, then paying for it is a good decision, not an indulgent one. The point isn’t to default to the cheapest option either. It’s to make the decision on purpose, based on what you actually need, rather than on what feels like the safer or more impressive choice.

Where to start

If this has you wondering what’s sitting in your own tech stack unexamined, start small. Pick your three most expensive tools and ask whether you’re using enough of what you’re paying for to justify it.

If you haven’t looked at your tool stack since spring, this week is a good week. Not to cancel everything. Just to see what you’re actually paying for.