The future of work

I noticed a post from Andreas Horn this week announcing that he was leaving IBM after eight years helping shape its AI strategy. It wasn't a resignation post. It was a reflection on what he had learned, and his observations neatly capture where we are in the AI revolution.

His biggest lesson was that almost everyone is focusing on the wrong thing.

Most organisations obsess over buying AI technology, yet that turns out to be the easiest part. Success is perhaps 10% the technology itself, 20% the underlying data and platforms, and 70% the people, culture and willingness to change. Unfortunately, most companies invert those numbers and spend almost all of their energy on the technology.

The same is true of data. AI doesn't magically fix broken organisations. It amplifies them. If your data is fragmented, inconsistent and mistrusted, artificial intelligence simply spreads that confusion faster. If your organisation is bureaucratic, agentic AI doesn't remove the bureaucracy. It gets trapped inside it. Increasingly, the bottleneck is no longer the technology. The bottleneck is the business.

Perhaps the most important lesson is that nobody really knows what AI is capable of yet. There is no playbook. The organisations moving furthest and fastest are not those producing glossy AI strategies. They are the ones running hundreds of small experiments, failing cheaply, learning quickly and constantly adapting. Curiosity has become more valuable than certainty.

That resonates because Andreas didn't spend much time talking about technology at all. Instead, he reflected on the people he had worked with, the privilege of building technologies that mattered, the freedom to experiment and the constant opportunity to learn. He also acknowledged something many professionals quietly feel but rarely admit: eventually the next chapter becomes more exciting than the current one.

That observation says a great deal about where work itself is heading.

For most of the industrial age, careers were measured by longevity. Spending thirty years with one employer was considered success. During the digital revolution, careers became measured by progression and promotion. Today, in the AI era, careers are increasingly measured by learning.

Learning has become the only sustainable competitive advantage.

Artificial intelligence is evolving so rapidly that expertise now has a remarkably short shelf life. The banker who dismisses agentic AI risks becoming irrelevant; the software engineer who stopped learning three years ago is already behind; the consultant relying on yesterday's frameworks increasingly discovers that the client has already asked ChatGPT for the first draft before the meeting has even begun.

Some people find that threatening whereas I find it liberating as the AI economy rewards curiosity far more than certainty. It favours people prepared to experiment over those determined to defend yesterday's expertise. The most valuable professionals will not have the longest CVs or the highest qualifications. They will be the people who learn fastest, adapt quickest and remain permanently curious.

That represents a fundamental shift in the nature of work.

For decades we hired people because they possessed knowledge. Increasingly we will hire people because they know how to acquire knowledge. Remembering information has become less valuable because AI already does that better than any human. The real value lies in asking better questions, framing problems clearly, connecting ideas across disciplines, exercising judgement and accepting responsibility when there is no obvious answer.

This is why I think we have misunderstood the impact of AI on employment. We keep asking which jobs will disappear whereas the far more interesting question is which human capabilities become more valuable.

Every industrial revolution has amplified human capability. Steam amplified muscle; electricity amplified manufacturing; computers amplified calculation; artificial intelligence amplifies cognition.

That changes everything.

The software engineer becomes the conductor of thousands of coding agents; the lawyer supervises autonomous legal research; the banker orchestrates intelligent financial agents that continuously monitor risk, execute payments, detect fraud and advise customers; the doctor increasingly interprets AI-generated diagnoses rather than producing every diagnosis alone.

The job survives but the work has changed.

This also changes how companies are built. The next generation of global businesses may not employ tens of thousands of people. They may employ a few hundred exceptionally capable individuals, each directing fleets of AI agents with productivity levels that would previously have required entire departments.

Which brings us to the real shift.

The future of work is not about replacing people with artificial intelligence. It is about amplifying people through artificial intelligence or, put even more simply:

The future of work is no longer people doing jobs. It is people becoming exponentially more capable through artificial intelligence.