Admin and operations
The freed hour is a decision, not a result
Workers report saving around eleven hours a week and losing over six to managing the tools. What happens to the difference is a management choice.
Written by
Founder desk, Agentic Global
Updated
2 August 2026
Reading time
4 minutes
- Setting
- Cross-cutting
- Freed to
- do the creative, judgement-heavy work — but only if someone protects the time
The uncomfortable finding
Glean's Work AI Institute surveyed 6,000 digital workers across the US, UK and Australia and found AI saves them roughly eleven hours a week — over a quarter of the working week. The same workers spend 6.4 hours a week on what the report calls botsitting: prompting, checking, correcting, re-running.
Only 13% said AI significantly improved company performance. Eighty-seven per cent were using it.
What that means in practice
Time freed by an agent does not become high-value time on its own. Absent a decision, it fills with more of the same work at higher volume, plus the new work of supervising a machine.
So make the decision explicitly. Name what the recovered hours are for before the pilot starts, write it down, and check after ninety days whether it happened. This is the whole thesis of this site, and we would rather state it as a deliberate choice than pretend it is automatic.
What to watch for
- Budget the supervision time. It is real, it is recurring, and it is rarely in the business case.
- Measure the work you wanted people to do, not only the work you removed.
- If the recovered time has no owner, it will be absorbed within a quarter.
More in admin and operations