Beyond productivity: What kind of working world are we building with AI?

There is a lot of focus right now on the productivity opportunity of AI, and for good reason. Research is showing meaningful gains in some types of work, with people completing tasks faster and, in some cases, producing higher quality results. I use AI every day and believe strongly in what it can enable. But productivity is only one part of what we are changing. The decisions organisations make about AI today will influence what work looks and feels like in the future. What we automate, what we continue to value, how much work we expect people to absorb, how they learn and where we still make space for human judgment will gradually become the working world the next generation inherits.

I have young children, so perhaps I think about this more literally than some. When they eventually enter the workforce, I want them to find a world where technology has made work better, not simply faster. I want AI to have removed unnecessary work while creating more space for creativity, judgment, relationships, purpose and the parts of being human that technology cannot replace. I also want them to work for successful businesses. These things do not need to be in conflict.

We are not just deciding how to use AI at work. We are designing the working world the next generation will inherit.

What happens to the time AI gives us?

If AI reduces a three hour task to thirty minutes, the productivity gain is easy to see. The harder question is what happens to the remaining two and a half hours. Some of it should create commercial value. Businesses need to grow, compete, serve customers and make money. But there are many ways to create value from capacity. It can mean more output, but it can also mean better decisions, deeper customer relationships, more innovation, stronger capabilities or simply enough breathing room for people to do good work without operating at their limit. Without making that choice consciously, we may simply increase the pace. Three reports become ten, ten customer interactions become thirty, and six important decisions become eighteen.

There is evidence that this matters for business too. Large scale research from Gallup has consistently found relationships between employee engagement and outcomes including productivity, customer loyalty and profitability. Better working conditions and strong performance do not have to sit on opposite sides of the equation.

AI can give us time back. Whether that becomes more output, better work or simply more work is still a human decision.

Faster work can also become denser work

AI does not only remove tasks, it changes the work that remains. Someone who once researched, analysed and wrote something from beginning to end may now start with an almost complete output from AI. Less effort goes into producing the first version, while more goes into understanding what the AI has done, checking context, finding errors, identifying what is missing and deciding whether the result can be trusted. I think of this as cognitive density: the amount of judgment, verification and accountability packed into a period of work. Research into AI assisted knowledge work is already pointing towards this shift. Critical thinking increasingly moves towards verification and integration. In one large experiment involving consultants, AI improved speed and quality when people worked on tasks within the model's capabilities. But when the task sat outside those capabilities, people using AI were less likely to reach the correct answer than people working without it.

This matters because AI can generate work much faster than humans can necessarily evaluate it. If we use that capability simply to increase volume, we may create working environments where people spend their days moving rapidly between AI generated outputs, carrying responsibility for decisions they had increasingly little involvement in creating. That might be productive on paper. I am not sure it is the future of knowledge work we should be aiming for.

A better future of work is not one where humans simply become the quality control layer for an ever growing volume of machine generated work.

We should be careful about what we optimise away

The same question applies to customers. Some human interactions are unnecessary friction and should absolutely disappear. But others contain something customers genuinely value: expertise, reassurance, taste, judgment, care or the feeling that another person has understood them. Research comparing chatbot and human interactions has found lower satisfaction in some situations, particularly where empathy and personal attention matter. The point is not to protect human involvement for its own sake. It is to understand what is valuable before we remove it. This principle extends beyond customer service. As AI becomes capable of doing more, organisations will constantly face decisions about what to automate. Efficiency is easy to measure, which makes it tempting to optimise for. Human value is often harder to put into a spreadsheet. Harder to measure does not mean less valuable.

The challenge is not deciding what AI can replace. It is understanding what is worth keeping human.

What happens when the first rung disappears?

I am particularly interested in what AI means for people entering the workforce. Many of the tasks AI can perform well are the same tasks junior employees have traditionally used to learn their professions. Analysts build models, junior lawyers conduct research, consultants prepare first drafts and developers write routine code. Viewed purely as tasks, much of this work looks highly automatable. Viewed as part of someone's development, it serves another purpose. It teaches them how the work works. If those tasks disappear, we may need fewer junior employees today while inadvertently weakening the pipeline of experienced people we need tomorrow. Someone who has never built the first model may find it harder to challenge an AI generated one later. I do not think the answer is to preserve low value work. But we need to distinguish between work we no longer need people to perform and experiences people still need in order to learn. This is particularly important if we think beyond the next annual planning cycle. The children growing up today may enter professions where many of the traditional first steps no longer exist. We need to start thinking about what replaces them.

If AI removes the first rung of the career ladder, we need to build another way for the next generation to climb.

People need to believe in the future they are helping to create

There is also an emotional reality to AI adoption that I think is underestimated. Employees are being asked to document their expertise, identify what can be automated, test AI systems and help redesign their own work while simultaneously hearing that AI may reduce demand for some of that work. I think of this as the coffin maker problem. We are asking people to help build a more efficient system while some are quietly wondering whether the finished system will still need them. That anxiety is not irrational. Pew found that 52 percent of US workers were worried about the future use of AI in the workplace. The World Economic Forum found that while 77 percent of surveyed employers planned to upskill workers in response to AI, 41 percent also expected to reduce their workforce where AI could replicate existing work.

We should be able to talk honestly about both possibilities. AI can make people enormously more capable, remove work nobody enjoys and create entirely new opportunities. It will also change or remove some roles. Treating every hesitation as resistance misses the point. People are much more likely to participate meaningfully in change when they understand the future they are helping create and can imagine themselves having a place within it.

People will help build the future of work more willingly when they can see a future for themselves within it.

We still get to choose

None of this is an argument for slowing down AI or using it less. I want us to use it well. AI gives us an extraordinary opportunity to rethink work. It can remove administrative burden, expand access to expertise, help people create things they previously could not, make businesses more productive and give us time back. The mistake would be assuming that whatever is most efficient is automatically the best destination. We can build profitable, ambitious and highly productive organisations while also creating better jobs, protecting the parts of work people value, developing the next generation and leaving enough space in people's lives for something other than output. In fact, I think that should be the ambition. The future of work is not something AI is doing to us. Every workflow we redesign, role we change, interaction we automate and productivity target we set is a small decision about what that future becomes.

AI will change the future of work. What that future is actually like is still ours to decide.

When I think about my children eventually entering that world, that feels much more important than simply asking how many hours AI can save. I want them to inherit a working world that is productive and prosperous, but also one that leaves room for curiosity, creativity, relationships, judgment, purpose and a life outside work. We have an extraordinary technology in our hands. The question is what we choose to build with it.