Where will the saved hours go?
Most talk about AI stops at how many hours it could save. I think the more important question comes next: what are those hours for? If nobody decides, the time tends to vanish, and you've paid for a saving you can't point to.
This isn't just my opinion
In the UK government's Copilot trial, users reported saving 26 minutes a day, but it wasn't possible to identify how that time was spent. And in the government's 2026 business survey, 64% of businesses using AI had seen no immediate financial effect.
Decide before you start, not after
Once the hours appear, they fill up on their own. One US study of a single firm found people took on more and worked longer, often without being asked. That's only one company, but it's a useful warning: a tool can make the day fuller, not better.
That's why I'd make this decision before anything is switched on, while it's still a choice.
Be honest about what's left
The hours on a sales page aren't the hours you'll get. Take off the time spent checking AI's work and putting mistakes right. What's left usually comes in small pieces across the week, not as a free afternoon.
That changes what you can realistically plan. Twenty minutes here and there might mean a proper follow-up call. It won't mean a new service.
Four places the time can go
- Work that keeps slipping. The customer check-ins and follow-ups everyone means to do. This is usually the best place to start, because the need is already there.
- Work only people can do. More time with customers, or helping a new starter find their feet.
- Learning. Time to build the skills that make someone better at their job. It's easy to cut and hard to get back.
- Breathing room. Fewer late finishes at month-end. That's a real result, even if it never shows up on an invoice.
How you'll know it worked
Pick one thing you want more of before you start, such as same-day callbacks, and see whether it actually goes up. If the hours were saved but nothing improved, they weren't really saved.
