Finally, AI strategy is getting more real

Tech has got to the point where its easier to make change than to talk about it

January 15, 2026

TLDR; In the last 6 months strategy speak has been replaced by real actions and practical advice. It’s time to just do things.

Last July I wrote a follow-up* article for IMD magazine on AI strategy. Due to the vagaries of publishing schedules, it only went live last week—a six-month lag that, in the world of AI, feels like a decade.

A delay like that is a useful stress test. Does my thinking hold up as the AI world moves on?

Happily, I still agree with myself, although the lofty title—“Is your strategy built for a world that no longer exists?”—definitely represents a time where real change felt more like ambition than reality.

I’ll save you reading the entire article (it’s long), but the framework can be distilled into two ways we see AI being implemented:

Bottom-Up: The team is given tools to create their own efficiencies. The problem is this doesn't address structural issues or answer fundamental questions—like whether we are optimizing a process that shouldn’t really exist anymore in the first place.

Top-Down: Management designs completely new processes based around AI. This is highly disruptive, risky, and can destroy as much value as it creates.

The sensible middle ground is to tackle the problem top-down, one challenge at a time, while enabling bottom-up benefits along the way. This creates momentum and builds real capability.

The HOW is less important than the WHAT

However, the key argument I’m making is that we often get the WHAT wrong because we are focused on optimizing what we do today. We need to be thinking about the different set of future opportunities that open up when we can apply unlimited intelligence to our goals.

The over used analogy I provide is that we are optimizing the shelves in Blockbuster when Netflix is already here.

The terrifying truth for leaders today is that AI is not a Netflix-level disruption for just one industry. It targets the universal substrates of knowledge work – how we think, communicate, and organize. It is Netflix-level disruption for every company, everywhere, all at once.

Today, I can make this argument much shorter and more practical, because in the intervening six months the technology has evolved. The conversation has shifted from chat tools to agents.

Simply put:

  • ChatTools are the bottom-up answer. In skilled hands, they supercharge efficiency, but widespread use isn’t driving growth.

  • Agents are the top-down answer. With skilled investment, processes can be completely reinvented. Work that used to cost 250 salaries can now be done for $2,000 in agent compute costs.

As 2026 begins, we are seeing a big rush to "agentify" everything. In other words, what I was talking about now has a name, and this is where I can now translate my article into practical advice. I wrote:

“This requires considering your organization across four core components, replacing the linear questions of yesterday with the exponential-type questions of the future.”

Let’s translate the consultant-speak in the article for a few practical agent building guidelines:

Just look at what goes in and what comes out

Forget the steps in between. Treat the work as a black box, then find a smart partner to brainstorm these basic questions:

  • Is there additional information we can use as input?

  • In a perfect world what outputs would make us more valuable?

  • How can we use AI to transformation input into output?

  • How can humans assist with or validate that transformation?

Reframe the role of the team

Going forward, your team's success doesn't depend on doing the "work." It depends on:

  • Being networkers, listeners, and gatherers who feed the AI as much context as possible.

  • Acting as expert judges who validate and challenge the AI’s output.

  • Actively avoiding actual work. Their job is to find new ways for AI to handle the heavy lifting.

Think of data as memory

We get bogged down in talk about "data integration." Instead, focus on two types of memory.

  • The first is the past: your team's hard-earned experience is the blueprint for the agent's initial design and rules.

  • The second is the future: as the agent performs tasks, it builds its own history. The goal isn't just to finish a task, but to retain that experience so the agent can learn from its own actions and improve the process tomorrow.

Finally, think about your impact and ambition level. AI enables us to scale the rarest commodity in the universe and the one thing that has bottlenecked humans until today: intelligence. That thought alone deserves ambitious thinking.

*And here is the article I was following up: Making the impossible possible

This newsletter is my current analysis of a rapidly evolving landscape. I welcome pushback, different perspectives, or data that challenges these theses. The best thinking on this topic will come from a collective effort. 

As AI gathers pace, we are all navigating a fascinating new chapter, and the full story is still being written. I look forward to exploring it with you.

And.. if you are still reading this and still want more you can read my archive of previous posts.