Imagine a customer service team after a successful AI rollout. Routine queries disappear from the queue. Customers can check deliveries and reset passwords without waiting. The people who remain on the phones spend their day handling disputed charges, unusual problems and customers who are already frustrated.
The dashboard shows fewer calls requiring human attention. It may even show more cases resolved across the business. Yet the human job has become harder: the straightforward conversations that once sat between difficult ones have largely disappeared.
This is an overlooked question in AI deployment. When technology removes the easy parts of a job, what kind of working day does it leave behind? A company can reduce the number of tasks people perform while increasing the demands placed on their attention and judgement.
A shorter queue can contain a heavier workload
Automation can make work considerably better. Repeatedly entering the same information or searching for a standard answer is often frustrating, and removing those tasks can give people more interesting responsibilities. The difficulty begins when managers assume that all released time can immediately be filled with demanding work.
An OECD study published in 2023, drawing on nearly 100 case studies in manufacturing and finance across eight countries, found improvements in job quality alongside reports of increased work intensity. The evidence predates today’s widespread use of generative AI and should not be treated as a verdict on current tools. It does, however, show why less routine work and a better working experience are separate outcomes.
Consider the team in the opening example. Resolving a straightforward delivery query and negotiating an exception to a refund policy may each occupy one slot in a workload report. They place different demands on the employee, particularly when several difficult conversations arrive consecutively.
The same issue can arise elsewhere. A finance specialist whose routine checks are automated may spend more time investigating discrepancies. A manager supported by AI may lose much of the administrative work while retaining every sensitive conversation and unresolved decision.
Redesign the day before changing the target
Before increasing targets, leaders should examine the work that remains. A practical starting point is to review a sample of working days with the people doing the job. Which activities require sustained concentration, negotiation or decisions with significant consequences? How often do those activities occur without a change of pace?
This should be a discussion about job design, rather than another system for monitoring individual effort. Employees need room to explain why two apparently similar cases differ. Otherwise, management risks introducing a new complexity score that is as misleading as the old task count.
The findings should change scheduling. Teams handling difficult customer cases could rotate some time into improving guidance, analysing recurring problems or supporting colleagues. Those activities contribute to the business while varying the demands of the day.
Managers also need to preserve time between difficult cases when the work requires it. Recording a decision, asking a colleague for advice and preparing for the next conversation are legitimate parts of delivery. If scheduling treats every minute outside an active interaction as unused capacity, employees will be pushed to compress those steps.
There is also a compelling business reason to protect this thinking time. If employees spend every available moment responding to exceptions, nobody has space to ask why the same exceptions keep returning. Giving the team time to identify a confusing policy or a recurring handover problem can reduce future demand. That improvement work belongs in the capacity plan from the beginning, with someone responsible for making the necessary changes.
Providing authority matters too. Someone handling an exception needs a clear route to resolving it. Asking an employee to calm an angry customer while denying them access to the relevant information or an authorised decision maker creates avoidable work.
Give people a say in the work they inherit
More demanding work is not automatically more meaningful. People differ in the challenges they enjoy, the skills they want to develop and the support they need. An employee who values careful analysis may welcome deeper investigations without wanting a day filled with confrontation.
Development conversations should therefore accompany automation decisions. Leaders can ask which remaining responsibilities employees want to grow into and where they need practice or support. This gives people some influence over their future role, instead of treating them as the destination for everything the system cannot finish.
Evaluation must also change. If routine cases disappear, longer average handling times may reflect a harder mix of cases rather than deteriorating performance. Managers should examine the quality of resolutions, repeated problems and employees’ accounts of the work before drawing conclusions from speed alone.
The ambition should be a working day in which people can use their abilities well and sustain that contribution. AI creates an opportunity to remove needless effort. Whether the resulting job offers room to think, learn and act responsibly depends on the choices leaders make about the work that remains.
