Do You Have a Portfolio Full of Horse Stables?

TL;DR

• AI is a tool: It is not a strategy, a product, or a replacement for thinking. Treat it like any other tool, and the hype gets much easier to navigate.

• Perform and transform: Disruption asks you to protect what pays the bills today while making bets on what will pay them tomorrow. Clarity about which is which is the strategic work.

• Find your horse stables: The most dangerous part of your portfolio is the part you are defending that AI is about to reinvent.

• Judgement stays human: AI speeds up the middle. The beginning and the end still belong to people.

• Your transformers aren’t on the org chart: The people who run today’s business are rarely the ones who will build tomorrow’s.

The Profound Revelation

I have had the chance to listen to a few AI talks over the last week, and I want to share my take on the landscape from my perspective as a strategy consultant.

The headline is: AI is a tool.

That is it.

I will give you a minute to get up off the ground from this profound revelation.

The problem is that a measured take doesn’t sell. In fact, half the subscribers to this newsletter have probably already stopped reading because my take isn’t polarizing enough.

Like all tools, AI serves a purpose. It can help you get the result you are looking for, or it can be profoundly unproductive.

What follows is my synthesis of what I learned from the CTO of a global technology firm and the principals of a board advisory firm. Neither conversation was technical or platform-specific. Both were about how to leverage AI, manage it, and protect yourself from it.

Nobody Won the Car Era by Building a Better Stable

One comparison stuck with me: the transition from horse-drawn carriages to cars.

It raised two points worth sitting with. First, if we had used the horse-drawn carriage as the blueprint for the automobile, the design would have been inherently flawed. Second, if you had viewed your investment portfolio purely as leverage on what you already owned, you would have spent the transition maximizing the value of your horse stables.

The underlying message was that disruption requires two lenses: perform and transform.

Perform is what you are doing successfully today, even if it may not pay the bills in the future. Transform is the set of bets that will. The strategic work is getting clear about what inside your portfolio belongs in which bucket.

It should not be surprising that disruptions are moving faster. According to one speaker, cars took about 55 years from inception to full adoption. The internet took about 12. AI is predicted to get there in roughly 8.

That pace matters when you consider the difference in the two speakers’ approaches. The more conservative governance view was to avoid, contain, and worry about liability. The more nuanced view was to apply appropriate judgement to different situations, recognizing that containment is extremely difficult and competitiveness is paramount.

Now is not the time to be an ostrich. If your business is not using AI in some capacity, you are probably missing opportunities. The hard part is knowing how to use it to maximize return and minimize risk.

Mapping Your Portfolio

One of the speakers had a great 2x2. Needless to say, it was my favourite. I love a good 2x2.

When presenting the model, the speaker did not label the axes; it was just four boxes of intent. So I took it a step further.

The vertical axis is strategic intent.

At the top is transformative or offensive work: innovation, new value creation, expansion. At the bottom is core maintenance or defensive work: protecting existing revenue, operational efficiency, and margin.

The horizontal axis is the extent of AI’s impact.

On the left, AI is incremental: it optimizes existing workflows or has lower leverage. On the right, AI is radical: it fundamentally alters the value proposition or business model.

Everything below the midline is the perform bucket. It is not sexy, but it is the blocking and tackling of running your business and what people trust you to do. You cannot get distracted from delivering it.

Above the line is where you are making bets. Top left is incremental bets to modernize or reposition what you already do. Top right is where you might make a big bet to invent something new or think completely differently about part of your business.

Pay very close attention to the bottom right. That is work you are defending and optimizing in an area AI is about to change radically. That quadrant is where the horse stables live. It feels safe because it is familiar and profitable today. It is also the most likely place to be blindsided.

For students of the strategy game, this thinking is not new. Geoffrey Moore described something similar in Living on the Fault Line with his Core and Context model. Core is the work that differentiates you, the reason customers choose you over anyone else. Context is everything else you have to do well to run the business, but it doesn't set you apart. Moore argued that companies waste their best people and resources on context when they should be investing them in core.

AI fits that model neatly. For most companies, AI is not the product. It is a tool that helps you innovate faster, whether by reducing friction and cost in your context work or by accelerating new and existing products in your core.

If You Don’t Have a Strategy, AI Will Not Become Your Strategy

The frontier firm I listened to still runs a 36-month business strategy cycle. They are making bets that AI is accelerating, but the foundation of how they run the business has not changed.

What has changed is how they fund innovation. Because AI is speeding up development and iteration, value shows up faster. So they now fund and review bets quarterly instead of annually. That seems smart.

They are also wrestling with a tension that is driving a bigger strategic conversation. As delivery speeds up, internal and external stakeholders expect costs to come down. But the judgement required, the liability, and the risk profile have not changed.

So the question becomes: if we can make more things faster. Still, clients’ expectations around liability are the same; how do we dramatically increase our judgement capacity without increasing our risk?

What was refreshing is that the companies at the front of this transition have not figured it out. If you are struggling, you are not alone.

Judgement. The Good Kind.

I mentioned judgement above, and I cannot stress this enough. Every speaker reinforced it: AI does not replace thinking. It just speeds up the middle.

An easy way to think about it is the 20/60/20 rule.

The first 20 percent of any idea comes from intelligent, contextually aware humans. AI and agents can accelerate the middle 60 percent of iterating, formulating, and challenging that idea. Humans must do the final 20 percent of refinement, judgment, and endorsement. If you offload any part of the 40 percent, you get slop.

If you or anyone on your team is offloading the beginning or the end to AI, that should concern you. The skills of prompting, iterating, and developing with AI need to be deliberately built, not assumed. What AI churns out is still not ready for primetime without a healthy dose of human intervention.

Do You Need to Know “How”?

Is it important to know how the AI produced something?

This debate conjures very specific memories of Arnold, the T-800, and Skynet. It is also one I am still chewing on, prompted by two unrelated podcasts.

The first was about AI solving several math problems long considered unsolvable. Humans could not solve them. The models did, and the mathematicians do not fully understand how. One mathematician’s lament surprised me. His point was not that not knowing is bad. It was that people go into mathematics to pursue the journey, not the answer, and AI had removed that joy.

The second was the story of a woman diagnosed with breast cancer who asked her doctor about her prognosis and was told, “I am not sure.” She survived and set out to collect data. That work became one of the most powerful breast cancer prediction models in existence, able to flag cancer before there are any medically diagnosable symptoms. Nobody fully knows what the model is detecting. But it is saving lives.

When asked whether it worried her that she does not know why it works, her answer was essentially: why should medicine be limited to what humans can understand?

I am honestly undecided about where I land. And I suspect most leaders will face their own version of this question sooner than they think.

Your Transformers Aren’t on the Org Chart

One final point stuck with me: the rule of thirds.

Even in this advanced frontier firm, the company is still split roughly into thirds. One group is well ahead on the frontier, one group is a fast follower, and one group is a laggard resistant to change, just like every other major disruption.

Two things stood out.

First, the company has realized that the people currently doing the work are not the ones who can transform it. They need to find the people in the lead group who are eager to transform, and those people aren't on an org chart.

Second, the transition matters too much to leave anyone behind, so they need to create incentives that bring the laggards along. Reassuringly, they have not figured out how to do that either. Once again, if you are struggling, you are not alone.

So What?

• AI is a tool, not a strategy: Without a clear strategy, AI will only help you do the wrong things faster.

• Find your horse stables: The defensive work sitting in the path of radical AI disruption is the most dangerous part of your portfolio, precisely because it feels safe.

• Shorten the bet cycle: Keep your long-term strategy, but fund and review AI-accelerated bets more frequently so you learn at the speed the technology allows.

• Protect the first and last 20 percent: Humans own the initiation and the endorsement. AI accelerates the middle.

• Your transformers aren’t on the org chart: The people who run today’s business are rarely the ones who will build tomorrow’s. Find the people eager to transform, wherever they sit.

➡️ Find Your Horse Stables

With your leadership team, map your major products, services, and internal functions onto the 2x2.

Then spend the rest of the conversation on three questions:

What is sitting in the bottom right?

Where should we place our bets?

Who can deliver them?

AI will not tell you which stables to close or who should build the car. That is still a leadership job.

Remember, nobody won the car era by building a better stable. Be bold and smart.