Today's Digest is a 4 minute-read

AI is making work faster. Whether it is making work better is much less certain.

This week I’ve been looking at what happens after the obvious efficiency gain: whether people are left with more interesting jobs or simply more checking, what AI does to the mental load of work and what changes when we bring it into the team rather than use it alone.

Together the three stories point to an issue I think leaders need to spend much more time on: we are getting better at identifying what AI can do but we now need to pay much closer attention to the work we are creating around it.

1. AI promised to remove the boring work. What if we are redesigning jobs badly?

I have used GenAI almost every day since 2022. It helps me identify where I could do with more evidence to support my argument, spots repetition and – the one I am most grateful for – does some of the repurposing (social media) work I would have once needed to carve out at least an hour a day for. Yet I have also become increasingly frustrated with it. Why? Because I spend a growing amount of time checking content that has been produced so quickly.

This made me wonder about one of the big promises surrounding AI at work. If machines take away the routine tasks are people really being left with more interesting work?

The finding: The people I spoke to for my latest feature suggest the answer depends heavily on what organisations do with the work that remains.

Microsoft researchers studying 319 knowledge workers found that generative AI changed where thinking happened, with more effort moving towards verifying information, integrating AI responses and overseeing the task. Business psychologist Andrew Whyatt-Sames sees this very practically inside his highly AI-enabled business. Time is no longer the constraint it once was. “You run out of energy,” he told me. His working day can involve what he calls “applied critical thinking all day, every day”.

Executive coach Emma Steenson found a different version of the same tension in her MSc research with practising lawyers. Everyone said AI saved time. What happened to that time varied. Some simply completed more work. Others had enough control to think more deeply. Her description of the first outcome is direct: “You’ve just compressed the boring jobs. And now you’re doing more of it.”

What’s really going on: There are at least three things I think leaders should watch.

• Follow the saved hour. Joanna Griffin, head of people and safety at American Golf, has used AI to cut much of the administration around employee relations cases. But this capacity has not simply appeared as spare time. HR can engage with people earlier but easier reporting can create more work elsewhere in the system.
• Be careful what you remove. Steenson’s senior lawyers and Griffin’s HR practitioners learnt their judgement partly through years of doing mundane work. Operations and technology consultant Matthew Wolff Simon sees the same risk with junior software developers who can now get highly convincing answers before they know enough to recognise when those answers are wrong.
• Question the idea of the ‘human in the loop’. I raised this with several of the people I interviewed because I am increasingly uneasy with the phrase. It starts with the AI system and then asks where a person should be inserted into it. All of them thought the framing was worth exploring. Griffin put the alternative particularly well: “A human isn’t in the loop. The human owns the whole process.” In her case-management system AI supports the work but HR practitioners still make the decisions, conduct investigations and handle the difficult conversations. This may prove much more important as organisations redesign jobs around AI. More on human in the loop to come!

The leadership question: When AI makes a task faster what job are you creating around it?

Look at the work people are doing after automation. Is their time being used for better judgement, learning and more valuable human interaction or have you simply replaced one kind of routine work with checking, supervision and a higher volume of output?

👉 Read my new feature on why removing boring tasks does not automatically create better jobs AI promises to remove the boring work. It may be creating more of it. 6-minute read

👉 Read more: If you are interested in the wider gap between how work is designed and what people experience explore The People Space’s Work Reality Gap.

2. AI may be reducing workload. What is it doing to thoughtload?

If AI can make some jobs less satisfying there is another question we can ask alongside it: what happens to the mental effort involved in the work that remains?

The finding: Organisational psychologist Dr Liane Davey has a useful term for this: ‘thoughtload’. She describes it as the cognitive and emotional burden people carry alongside their workload and argues that AI can reduce one while increasing the other.

Her example of an environmental engineer is a good one. AI can turn soil sample data into a remediation plan in one second when the task previously took around 80 hours. Yet much workplace AI use is less straightforward. People can now generate research, analysis and documents extremely quickly but somebody still has to decide whether that output is useful, check it and work out what to do with it.

More output can therefore create more work further along the chain.

Why it matters: The nature of the effort is changing too. AI may remove relatively straightforward processing while leaving employees with more prompting, checking and judgement. Davey warns that people can end up responsible for several threads of AI-generated work at once and still carry accountability when something goes wrong.

She gives the example of a leader who trusted AI with financial reports and later discovered an error had contributed to an emergency cash-flow problem. The technology had done the work quickly but the human consequences remained with him.

What’s really going on: This challenges a very narrow definition of AI productivity. Counting the hours saved on the original task tells you very little about what happens across the whole workflow.

Managers need to look at where the work goes next. Has AI genuinely removed effort or transferred it to the person checking the output? Are employees producing material because it helps someone make a better decision or because producing more has suddenly become easy?

There is also a development question here. Davey argues that managers should be deliberate about which parts of thinking can safely be automated and where people still need to practise judgement themselves. An employee who no longer does the underlying work may eventually find it harder to assess whether the machine has done it well.

The leadership question: When you assess the impact of AI on a role, look beyond time saved. What new checking, decision-making or accountability has appeared elsewhere and who is carrying it?

👉 Read Liane Davey’s full argument on why AI will not fix productivity until managers start managing thoughtload. 5-minute read

3. AI has changed how we work alone. Now comes the team test

Liane’s argument about thoughtload raises an important question about how we use AI as individuals. But much of the work that shapes organisations happens with other people, in meetings, project reviews and decisions that are debated rather than produced alone.

The finding: Gabriele Rosani and Elisa Farri of Capgemini Invent’s Management Lab have been looking at what happens when generative AI becomes part of that collective work.

Around half of leaders and managers are already using generative AI individually, according to figures they presented during a Harvard Business Review webinar last week. Fewer than 10% are actively integrating it into group settings.

Their experiments with more than 300 managers suggest there may be value in changing this. Around two-thirds reported higher-quality outcomes when they worked with AI collectively and a similar proportion felt some AI risks had been reduced through what Gabriele calls ‘collective judgement’.

Why it matters: One reason is simple: using AI with other people can slow us down.

Working alone it is easy to accept a plausible answer and move on. Show several people the same answer and they are more likely to interpret it differently, question it or challenge the assumptions behind it.

Gabriele and Elisa have found some of the strongest results come when AI is asked to challenge the team rather than solve the problem for them. In one innovation workshop AI acted as a ‘sparring partner’, questioning the group’s idea and then waiting while people discussed their response. In another case participants used AI before a quarterly review to pressure-test recommendations so the meeting could begin with some of the obvious weaknesses already exposed.

What’s really going on: This points to a different kind of AI capability. Organisations have spent a lot of time helping people learn how to prompt a chatbot. Team use requires something else: knowing when AI adds something to the conversation and when it starts getting in the way.

Elisa warns against meetings where one person types while everyone else watches the answer appear. “People will switch off. They will just watch a screen.”

Meanwhile Gabriele’s test is a good one: “Are we as a team owning the conversation or is AI leading it?”

The leadership question: Try bringing AI into 10 minutes of an existing team discussion and use it to ask questions rather than provide answers. Then watch what happens. Does it sharpen the debate and encourage people to think harder or does the room start deferring to the machine?

👉 Read the full story on how Rosani and Farri think teams can use generative AI without handing over the conversation. What happens when the whole team starts using AI? 4-minute read

Find your creative space!

Siân

Editorial director, The People Space​

Know someone who’d find this useful? Feel free to forward it.

PS: If these questions about AI, work design and human judgement are coming up in your organisation I run practical workshops for HR and leadership teams on what AI means for their people and how to move from experimentation to responsible use at work. We look at the work itself, where human judgement needs to stay strong and what leaders need to change around the technology.

If you’re planning an AI session, leadership offsite or team discussion take a look at my workshops here.