An anonymous real example
A professional-services team explored cutting the time needed for a fixed-fee job. One person spotted the catch: if they finished sooner and there was no other billable work waiting, their utilisation rate would fall. The business could benefit from a more efficient service while the individual appeared to be performing less well. This was uncovered while testing the idea, before treating the proposed time saving as a result.
The assumption worth revisiting.
Utilisation can be a useful measure in an hours-based model. But when the same outcome can be delivered with less effort, it may tell a different story from contribution, quality or capacity created. The assumption to unlearn is that the measures which helped you run yesterday’s business will reward tomorrow’s improvement. Asking people to release time without deciding how that time counts leaves them to resolve the contradiction. Useful improvements can stall, time savings can stay hidden, and people may fill the space with low-value work to protect a familiar number. The investment in AI then changes tools more than it changes performance.
Does your scorecard reward the behaviour your strategy needs?
Give the capacity a purpose.
Give Operations, Finance and People time to agree how released capacity is recognised and used. Make room for better service, learning, risk work and growth experiments before the old measure absorbs the time again.
How this fits the capacity allocation →
Use AI to challenge your first interpretation
Explore this question without assuming our first interpretation is right: Does your scorecard reward the behaviour your strategy needs? Offer competing explanations, identify assumptions and suggest evidence we could obtain from the people doing the work.
Treat the answers as possibilities to test with people and evidence.
Go deeper into the evidence.
NBER; author manuscript on arXiv · Randomised field experiment; working paper
The six-month experiment covered 7,137 knowledge workers at 66 firms. In its second half, the 80% of treated workers who used the tool spent about two fewer hours on email weekly and reduced work outside regular hours. Researchers did not detect broader changes in the quantity or mix of tasks from individual tool provision.
These are specific tool and work-pattern findings, not proof of two days released per week. Some authors were employed by Microsoft, the tool maker. Use the revised estimates rather than the earlier 3.6-hour estimate for frequent users.
podcast
Agentic AI and the future of global business services
McKinsey · McKinsey Talks Operations
A useful prompt to reconsider the boundary of shared services: which decisions and activities could move closer to internal operating knowledge?
Listen or watch for: Would moving more work inside improve the complete process, or simply move a bottleneck?
About this resource
Consultancy examples and forecasts. The episode does not establish that insourcing is always cheaper or that predicted headcount reductions will occur.
The Cobra Effect: test the incentive before scaling it
The familiar cobra-bounty story is a parable about a measure that rewards the wrong behaviour. Its historical details are not established. The useful question is what people may do to succeed against a target, and what happens to the wider outcome.
Use some released time to explore unintended responses before committing the whole organisation.
Watch LSE: good intentions, perverse incentives ↗