Skip to content
← All articles

AI · November 14, 2026 · 3 min read

AI saves time. Why are costs not falling?

by Federica Grazia Bartolini

Federica Grazia Bartolini’s hands on a laptop keyboard, in warm black and white.
Share

The answer is simpler than it may seem: time saved is not, by itself, a business result. More room in the diary certainly offers some breathing space between meetings. But is that the value the business is looking for? Lower costs depend on improvement across the whole workflow, rather than individual schedules alone.

A UK government survey published on 29 October illustrates the distinction. Some 32% of businesses use at least one AI tool, twice the 16% reported in 2025. Half of those using AI report higher productivity. Yet 64% have seen no effect on operating costs. These findings come from 3,503 UK interviews: self-reported results, rather than directly measured outcomes, and not figures that can simply be transferred to Italy (survey executive summary). The underlying question matters here too.

A few weeks ago, I discussed this with a manager. His team was producing documents and materials in half the time, but the business was not seeing the benefit: for one reason or another, delivery still came close to the deadline. In that case, the problem was an endless approval process, alongside a tendency to put things off.

The bottleneck moves

Take a communication team that uses AI to prepare a draft in one hour rather than two. If the text then goes through the same four levels of approval, it still takes three days to reach its recipient. The internal customer notices no difference. AI has accelerated one stage; the outcome still depends on how the work is organised.

That is why I would never simply add up everyone’s “hours saved”. I would look at the complete activity, from request to delivery, including the time spent checking and correcting what the AI produces.

Before measuring, decide what success means

Reducing costs is a legitimate objective, but it need not be the only one. Released time may help a business respond to customers sooner, improve quality, clear a backlog or support a better work-life balance. These are different outcomes with different measures. Confusing them creates expectations no project can meet.

In the Metodo Golden Bridge®, I propose measuring the quality of change, rather than tool use alone. Counting logins tells us whether a platform is being opened; it does not tell us whether work is improving.

Three questions before claiming a saving

  1. In which activity, from beginning to delivery, has total time actually fallen?
  2. Where has the released capacity gone: lower costs, quality, speed or additional work?
  3. Do people notice a difference in their working day, or are they simply receiving more requests?

Released time is also a promise to employees

When a business promises “more space for valuable work”, people need to experience it. If every recovered minute is immediately filled with another request, the promise breaks down and trust in the change weakens. Concrete examples of progress are more useful than adoption percentages alone.

Deciding where the time released by AI goes is a leadership choice. No usage counter can make it on our behalf.

For businesses exploring where time saved with AI can become a meaningful result, this is how we work at FGB Communications.

English version published 14 November 2026. Original Italian article published 5 November 2026; the survey cited above was published 29 October 2026.

Frequently asked questions

Does AI reduce business costs?+

It can, but not automatically. Time savings turn into lower costs when the organisation changes the processes around the tool.

How is AI productivity measured?+

Across the complete workflow, including time spent checking and correcting outputs, and the benefit for the person receiving the work.

Where can a business start?+

With a familiar, repetitive activity and an outcome defined before the experiment begins.

Enjoyed this? Share it →

Find me on

FGB Briefing · AI, Business, Communication, People

Federica Grazia Bartolini’s briefing on AI, governance, reputation and corporate communications. Original summaries, practical implications and links to sources.

Discover FGB Briefing
Check your email and confirm your subscription using the link you receive.

Choose the language in which you would like to receive each edition.