12 Weeks Later: What AI Is Really Changing

Twelve weeks ago I started this series with a simple objective. Not to predict the future. Not to chase the latest AI headline. But to understand what AI is actually changing inside organisations.

Each week I explored a different question. Sometimes the evidence confirmed what I thought. Sometimes it challenged it.

Looking back, something surprised me. The first twelve weeks are not really about artificial intelligence. They are about competitive advantage. More specifically, they are about how AI is changing what creates competitive advantage.

Every major technology shift follows a familiar pattern. It makes something that was once scarce become abundant. Printing made books abundant. The internet made information abundant. Cloud computing made computing infrastructure abundant. AI is making many forms of knowledge work abundant.

The important question has never been what becomes abundant. It is what becomes valuable afterwards. That is the pattern that emerged from the first twelve weeks.

Competence became easier to produce. Judgement did not.

Lesson one: AI is making competent work cheaper

The series began by asking whether AI was inspiring people to do more, and what people were actually replacing with it. As the weeks progressed, a clearer picture emerged.

AI is remarkably good at helping people produce competent work. Drafts. Research. Summaries. Analysis. Software. Plans.

Harvard Business School and Boston Consulting Group demonstrated this in their landmark experiment involving 758 consultants using GPT-4. On tasks that fell within the model's capability frontier, participants completed 12.2% more work, completed it 25.1% faster, and produced significantly higher-quality output. Lower-performing consultants improved 43%. Higher performers improved 17%.

Competent work became dramatically easier to produce.

Then the researchers deliberately introduced a task that sat outside the model's capability frontier. On that task, consultants using AI were 19 percentage points less likely to reach the correct answer than consultants working without AI.

Same people. Same tools. Same day. AI improved routine competence while degrading complex judgement within the same experiment. That is why the paper is called Navigating the Jagged Technological Frontier.

Competence became easier to produce. Judgement did not.

Lesson two: AI exposes organisations

One theme kept appearing throughout the series. AI rarely creates organisational strengths or weaknesses. It reveals them.

Week three argued that AI removes hiding places. Week four suggested that the real AI skill gap is judgement rather than prompting. Week seven explored why weak leadership becomes more visible before workers are replaced. Week eight argued that AI does not replace thinking. It exposes it. Week twelve reached the same conclusion from another direction: AI does not suddenly make trust valuable. It makes reputation dramatically easier to verify.

Looking back, these were not separate observations. They were all describing the same phenomenon. AI is dramatically reducing the cost of verification. It is becoming easier to see how organisations actually think. How they actually lead. How they actually execute. And how customers actually experience them.

There is an important balance to this. AI is also reducing the cost of fabrication. Fake reviews. Synthetic testimonials. Generated content. Manufactured evidence.

The difference is that while any single signal has become easier to fake, consistency has not. A decade of customer experiences, successful implementations, independent references and reliable delivery remain extraordinarily difficult to manufacture.

That is why reputation becomes more valuable, not despite AI, but because AI increasingly compares evidence rather than simply accepting claims.

Lesson three: AI changes competition before it changes employment

Much of the public conversation still revolves around jobs. Will AI replace workers? Which roles disappear? Those are important questions.

But over the past year, working with banks, fintechs and payment providers across African markets, I have noticed something different. Executive conversations rarely begin with AI. They begin with delivery. Integration. Compliance. Operational resilience. Commercial outcomes.

AI has entered those conversations. It has not replaced the fundamentals.

That observation mirrors the broader pattern emerging across the series. AI lowers the cost of capability. Markets raise the standard for competition.

That naturally led to week eleven. Who actually captures the value created by technological change?

The closest historical evidence comes from William Nordhaus' landmark research. Across the American non-farm business economy between 1948 and 2001, producers captured around 2.2% of the total social returns created by technological innovation. The remainder flowed largely to customers and the broader economy.

History does not prove AI will behave identically. But it reminds us of something important. Productivity does not automatically become profit. Competition determines who captures much of the value.

That changes the strategic question leaders should ask. Not "How much can AI save us?" But "How much of those savings will we actually keep?"

Lesson four: competitive advantage keeps moving

Perhaps the biggest insight from the first twelve weeks is this. Every time AI makes one capability more common, another capability becomes more valuable.

As AI makes something abundant, something else becomes more valuable:

• Competent work becomes abundant → judgement becomes more valuable.
• Productivity becomes abundant → execution becomes more valuable.
• Software becomes abundant → differentiation becomes more valuable.
• Information becomes abundant → trust becomes more valuable.

This became the framework that emerged from the series. Technology changes what is abundant. Markets determine what becomes valuable. Competitive advantage follows.

The organisations that win are rarely those that compete hardest in the commoditised layer. They are the ones that move towards whatever remains difficult to copy. That is where advantage accumulates.

The framework that emerged

When I began writing this series, I thought I was writing about AI. Looking back, I think I have actually been writing about four enduring business capabilities.

• Thinking. AI accelerates answers. It does not decide which questions matter.
• Execution. AI creates potential. Organisations create outcomes.
• Competition. AI commoditises capability. Strategy determines where value is retained.
• Trust. AI does not create reputation. It reveals it.

Those four capabilities existed before AI. They simply become more valuable as AI becomes more capable.

What surprised me most

Some of the strongest conclusions in this series were not the ones I expected.

I thought AI would primarily be a productivity story. It turned out to be a competition story.

I thought the biggest challenge would be learning new tools. It increasingly looks like building better organisations.

I thought trust would become more important. The evidence suggested something more interesting. Trust has always mattered. AI simply makes it dramatically easier to verify.

None of those conclusions came from opinion. They emerged from following the evidence. That has been the most rewarding part of writing this series.

Where the series goes next

The first twelve weeks explored what AI is changing. The second half of the series will explore what leaders should do about it.

We'll move beyond productivity. Beyond prompting. Beyond the technology itself. We'll explore decision-making, organisational design, leadership, governance, culture and strategy.

Because I increasingly believe the organisations that succeed will not simply be those with access to the best AI. Almost everyone will have access to capable AI. The winners will build organisations that think more clearly, execute more consistently, compete more intelligently and earn trust more reliably.

The bottom line

When I wrote week one, I thought this series would largely be about artificial intelligence. Twelve weeks later, I think it is about something much bigger. It is about how every major technological shift reshapes competitive advantage.

One sentence captures the first half of the journey. AI makes capability increasingly abundant. Markets determine what becomes valuable. Leadership decides whether organisations capture that value.

So I'd like to leave you with one final question. If AI made your greatest strength available to every one of your competitors tomorrow morning, what would still make your organisation exceptional?

---

References

Dell'Acqua, F., McFowland III, E., Mollick, E., Lifshitz, H., Kellogg, K. C., Rajendran, S., Krayer, L., Candelon, F. & Lakhani, K. R. (2026). "Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality." Organization Science, 37(2), 403–423. doi.org/10.1287/orsc.2025.21838

Nordhaus, W. D. (2004). "Schumpeterian Profits in the American Economy: Theory and Measurement." NBER Working Paper No. 10433.