AI IS SAVING YOUR EMPLOYEES' TIME. WHAT ARE YOU DOING WITH IT?

By Dr Fabio Oliveira, 26 October 2026

For the past three years, I have heard the same question from managers experimenting with artificial intelligence. How much time can AI save us? But I think we are asking the wrong question.

The question leaders should now be asking is: What are we going to do with the time AI gives back? As we enter a different phase of AI adoption, this question becomes more important than ever. The experimentation phase is giving way to something more difficult.

Firms now need to convert individual productivity gains into organisational performance. The evidence suggests that this transition is proving difficult. Recent McKinsey research (AI is changing work. Now it has to change the organisation. 5 October 2026 ) found that 80 per cent of respondents report improvements in individual productivity from AI, yet only 37 per cent report a positive impact on EBIT. Just 6 per cent qualify as AI high performers. The message is difficult to ignore. Making employees faster does not automatically turn a firm into a better financial performer.

This week, EY-Parthenon reported that 50 per cent of UK CEOs consider AI-enabled tools among their three main contributors to productivity gains over the past year. That sounds encouraging. But only 14 per cent have comprehensive, real-time visibility of AI costs, while 32 per cent report significant gaps in their ability to assess return on investment. More importantly, 71 per cent believe AI will have a greater impact on roles, skills and the organisation of work than on the number of people employed.

I think this is where the conversation about AI and the future of work needs to move. We have spent considerable time discussing which jobs AI might replace. The more immediate management problem is different: AI is changing the composition of work inside existing jobs.

McKinsey estimates that more than 70 per cent of US workers could need to reinvent how they work as automation changes the tasks they perform. Around a quarter could see more than 30 per cent of their task time reshuffled or moved to new activities. If that happens, installing an AI tool is not enough. Someone needs to decide how the work itself should change.

Consider a salesperson. An AI agent can research a prospective customer, summarise CRM records, identify relevant market information and prepare a briefing before a meeting. Imagine this saves the salesperson one hour. What happens to that hour? If nothing changes, the firm has created an invisible hour of capacity. Perhaps the salesperson answers more emails. Perhaps another internal meeting fills the space. Perhaps management simply increases the number of accounts the salesperson is expected to manage.

None of these outcomes necessarily creates value. Instead, the sales leader could deliberately redirect that hour towards activities where human capability matters, such as speaking with customers, developing relationships, coaching colleagues, understanding a complex account or negotiating a deal.

The same question applies to marketing. If generative AI allows a marketing team to produce content five times faster, should it produce five times more content? Or should marketers spend more time understanding customers, testing ideas, analysing behaviour and developing propositions that competitors cannot reproduce by using the same AI tools?

For CEOs, the issue becomes even more important. AI can summarise reports, synthesise information, prepare scenarios and monitor performance. The objective is judgment, conversations, decisions and leadership.

This is why I believe the productivity question is not “How much time did AI save? but "What did we do with the time AI gave back?"

There is another problem. Many firms are introducing AI into processes designed for a world in which humans performed almost every task. We add Copilot. We introduce ChatGPT. We deploy an AI agent. We train employees to write prompts. Then we wait for transformation. But transformation does not occur simply because a new technology sits atop an old process.

The World Economic Forum recently described AI as reorganising work rather than simply replacing jobs. KPMG makes a similar argument: firms risk falling into an efficiency trap when they use AI to execute existing processes faster without reconsidering how the work should be organised. For leaders, this means moving from AI adoption to work redesign.

I would start with three questions:

1. What work should AI do? Look for repetitive, administrative, and information-intensive activities where AI can genuinely free up human capacity.

2. What should I do with the saved hours? Be explicit. If AI releases five hours per person each week, decide where those five hours should go. Customer relationships? Innovation? Coaching? Prospecting? Strategic thinking?

3. What needs to change around work? Targets, incentives, roles, workflows and performance measures may all need to change.

There is little value in giving employees AI while continuing to manage them based on assumptions made before AI existed.

I do not think the next stage of AI adoption will be defined by who has access to the most AI tools. Those tools are becoming widely available. The difference will come from how firms combine human and AI capabilities and how deliberately leaders redesign work around them.

EY’s latest CEO research offers a warning here. Nearly half of UK CEOs surveyed believe their firms are not developing workforce skills quickly enough to keep pace with AI innovation. Teaching people how to use AI is part of the answer. But it is not enough. This is what I will be writing next for "The Critical Loop". See you back in a few weeks' time.