Working Smarter with AI, not just faster

There is no question that AI can speed up tasks. It is fast. It gets you from A to B quickly, as long as you know where A and B are. Sometimes it helps with drafting a report. Sometimes it takes the fiddly part out of an admin task.

Say AI saves you thirty minutes a day writing emails. Where are you now spending that thirty minutes?

Are you putting it into more billable hours? Did it give you, or a member of your team, more time to think? Did they get an extra thirty minutes to go for a walk and take some of the stress out of a stressful job? Or was it spent using AI to create a second and third draft of something that was possibly fine at the first?

It is a hard question to answer, because the time is saved in small pieces, at every desk, and spent the same way, so it rarely gets looked at as a whole.

So how do we know AI is helping us get the best out of our people, as well as getting through the work faster? What is AI really giving our teams? Let's have a look at the numbers.

1Where the thirty minutes go

Economists at the Federal Reserve Bank of St. Louis surveyed US workers in November 2024 and found people using generative AI saved an average of 5.4% of their working hours. On a 40-hour week, that comes to about 2.2 hours. The same economists noted those gains may not show up in productivity statistics straight away.

Where that time shows up depends on how a firm measures productivity. A firm that counts hours will see the saved time vanish into the count, which raises its own questions for the billable hour. A firm that counts output has a different question to ask: what was the extra output, and was it the work the firm most needed done?

A separate study in Denmark asked where saved time goes. Anders Humlum and Emilie Vestergaard surveyed around 25,000 workers across 11 occupations, accountants and legal professionals among them, and compared the answers with official records of pay and hours worked. Their paper found no measurable change in either.

Asked what they had done with the time AI saved them, 85% said they had moved at least some of it to other work tasks, and about 30% said they had spent more time on the same task they had saved time on. Fewer than 10% said they had taken extra breaks or time for themselves.

In 1955, the historian C. Northcote Parkinson opened an essay in The Economist with this line: "It is a commonplace observation that work expands so as to fill the time available for its completion." He was writing about government departments that kept growing while their workload shrank.

2The load on teams

People in law and accounting firms work hard in demanding jobs, and they were stretched well before AI arrived. The 2025 Victorian Lawyer Census found 53% of Victorian lawyers at risk of burnout, including 25% at very high risk, with workload among the strongest drivers. Across all kinds of office work, Microsoft's research found 48% of workers say their work feels chaotic and fragmented.

The hope was that AI would lighten that load, and so far the research leans the other way. A survey of 2,500 executives, employees and freelancers across the US, UK, Australia and Canada, run for the freelance platform Upwork, found 77% of employees using AI said it had added to their workload, and 47% had no idea how to achieve the productivity gains their employers expected. Of the executives surveyed, 96% expected AI to lift productivity.

Researchers at the University of California, Berkeley followed around 200 employees at a US technology company for eight months. Writing in the Harvard Business Review, they found people using AI "worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so." Those extra hours sit awkwardly with Australia's right to disconnect. AI output that looks finished and has to be redone has a name now: workslop.

Australians have won time back before. On 21 April 1856, stonemasons in Melbourne walked off the job and won an eight-hour day, and a 40-hour, five-day week took effect in January 1948, as the National Museum of Australia records. Value, time and productivity: what are we working for? tells the full story of how Australians won those hours.

So we have the most capable technology we have ever had, and at the same time a lot of people are reporting chaotic and fragmented working days. Why is this happening?

3Why it feels this way

Part of it may be that so many people are learning at once. What we hear in firms often sounds like this:

• The tools have only been around for a few years, so there has been little time to get good at them.

• Many people feel they are expected to use AI, and not all of them have been shown how.

• Some worry about what AI means for their jobs, and the headlines lately have raised bigger questions still.

• Guidance on what is allowed can be thin, or missing altogether.

Much like the early months of COVID, it can add up to decision fatigue. Every time a chat window opens there are small decisions to make: which tool, what can go in it, and whether the answer can be trusted. Add switching between tools and tasks, and a working day can fill with small choices before the main work has started.

The research on decision fatigue is mixed, and some of the famous studies have not held up. One that is often cited in medicine comes from a team led by Jeffrey Linder, published in JAMA Internal Medicine in 2014. GPs became more likely to prescribe antibiotics for respiratory infections as their clinic sessions went on. The pattern was consistent with decision fatigue, although the authors noted other factors may also have contributed. The Berkeley researchers warned that AI-intensified work can lead to weakened decision-making, alongside cognitive fatigue and burnout.

So where to from here?

4Working smarter

Maybe we should be looking at how AI can help us work smarter, not just faster.

Once you know saved time tends to fill itself, you can decide what it is for before it disappears. It could go to thinking a hard problem through properly, or it could go to rest.

Here are four ways a firm could use an extra thirty minutes a day to build a better place to work.

1. Bring burnout down. Burnout is costly as well as hard on people, in law and in accounting. Gallup found employees who frequently experience burnout are 63% more likely to take a sick day and 2.6 times as likely to be actively looking for a new job. Handing back some of the time AI saves is a direct way to take pressure off.

2. Make time to think about the bigger picture. The hardest matter on the desk, the client strategy that never gets a clear hour, or how the firm itself could work better. Gloria Mark's research at the University of California, Irvine found that once people are pulled away from a piece of work, it takes around 25 minutes to get back to it, so thinking time works best as a protected block.

3. Build in creative breaks. Stanford researchers Marily Oppezzo and Daniel Schwartz found people's creative output rose by an average of 60% while they were walking. A review of 22 studies found short breaks lifted energy and reduced fatigue, and that longer breaks did more for performance than short ones.

4. Ask your staff what they want. Professional services firms are competing for good people. A study of 1,891 lawyers in Victoria, Western Australia and New South Wales, reported by Lawyers Weekly, found 28.6% were considering leaving their current employer within the next year. CA ANZ has reported fill rates below 67% for six key accounting and finance roles. Asking your team how they would use the time, and acting on what they say, helps build the kind of workplace people choose to stay in.

We look at these in more detail, along with the four-day week, in What to do with the time AI gives back.

If the thirty minutes goes straight into more billable work, we are asking people for more decisions at the end of a full day, and it is hard to see those decisions getting better. Are we working faster with AI, and no smarter?

5Talk to your team

Three questions to ask your team:

1. Are you saving time? Ask plainly, and ask whether people feel AI is reducing the hours they spend on their work. A short anonymous question can get more candid answers than a meeting.

2. If not, what is in the way? The reasons are usually practical: AI pointed at the wrong tasks, no shared prompts, or no time set aside to learn. Our posts on the prompt library and onboarding are a place to start.

3. If so, what is the time for? Put a few management decisions in place about how saved time gets used, so people end up making better decisions, rather than twice as many. It can be as simple as agreeing which of the options above the firm will offer, and checking in three months whether it is working.

6Where we come in

Our investigation measures how long your firm's work takes today, so you can see whether AI is giving time back and where that time is going. The findings belong to your firm whether or not we go any further together.

If you would like to know what AI is really giving your people, reach out to our team. The first conversation works out whether your firm is the right fit for the AI Uplift Program, and whether we are the right people to help.

Talking to your team