Why Is AI Making Me Think About Business Strategy?

A few days ago, I wrote down a phrase: “A data-driven product repair company.”

Normally, that would have been enough to make me start building something. I would have opened Figma, made a homepage, invented a visual language, registered a domain and spent several days turning the idea into an object. The object would gradually become more convincing. By the end, I might even have believed I had a business.

This time, I started asking different questions.

How large is the market? What budgets would this company compete for? Is there a trend arriving that makes the idea more urgent? What would the first service be? What kind of company would buy it? Could the service eventually produce proprietary data or software? Why would this become more valuable over time?

I have been making things on the internet for most of my adult life, yet I have rarely thought this seriously about business strategy. I understood that businesses needed customers and revenue, obviously. I could write proposals, price projects and explain the value of design. Still, my thinking usually began with the thing itself. The product, website, film, platform or interaction came first. The business was something I expected to discover around it.

AI is changing that sequence.

Building used to consume so much attention that it could postpone almost every other question. An idea had to survive weeks or months of production before the market had a chance to reject it. During that time, the effort itself created conviction. Every new screen, feature and technical solution made the project feel more real. It became difficult to separate the value of the idea from the amount of work already invested in it.

There was also a certain safety in being a maker. When the outcome was disappointing, there were always practical explanations. The design needed more refinement. The technology was incomplete. The launch came too early. The marketing never really happened. I could remain inside the craft and imagine that the missing piece was another round of making.

AI removes much of that shelter.

I can now turn an idea into a functioning prototype before I have developed an emotional attachment to it. I can ask an agent to create the interface, architecture, database and first version of the product while I continue thinking. The gap between imagining something and seeing it work has become so small that execution no longer provides the same sense of progress.

The prototype appears, and then there is an awkward silence.

Who needs this?

Why would they change their current behaviour?

How would they hear about it?

What existing budget would pay for it?

What advantage would remain after someone copied it?

These questions used to arrive after the work. Now they arrive almost immediately, because the work is already sitting in front of me.

This may be one of the less discussed effects of AI on people who make things. We talk constantly about productivity, job displacement and the growing number of people who can code. We spend less time talking about what happens psychologically when execution stops being the main obstacle.

For years, the ability to execute was close enough to a strategy. A designer who could also code had an advantage. A small team that could ship quickly had an advantage. A founder who understood video, branding, product and technology could combine those skills into something few others could produce. The difficulty of making the object created a form of protection around it.

That protection is becoming thinner.

A polished interface no longer proves that a serious company exists behind it. A working product no longer suggests that the founders possess unusual technical capabilities. A long feature list may simply mean that nobody was forced to choose. AI can generate convincing evidence of progress without answering whether the progress is moving anywhere useful.

For someone like me, this creates a strange reversal. I have more power to build than I have ever had, and this power is making me more hesitant about what deserves to be built.

The hesitation feels healthy.

I recently started examining the idea of HAAM as a data-driven product repair company. The initial phrase was intuitive. Many digital products are confusing, fragile, inaccessible or difficult to maintain. AI is allowing companies to create software faster, which will probably produce even more products that technically function while failing as coherent systems.

A few years ago, I might have turned that observation directly into a website. This time, I compared the idea with adjacent markets. Software testing, application modernization, product analytics, user experience consulting, accessibility, conversion optimization and observability already represent large pools of spending. Product repair could potentially connect parts of those markets around a clearer outcome.

That exercise changed the idea. It stopped being a description of what I enjoy doing and began becoming a theory about where money, urgency and technological change might intersect.

I had never naturally thought in terms such as market entry, category creation, serviceable market, accumulated advantage or strategic wedge. Some of this language used to feel like the territory of management consultants and people who had spent too much time around venture capital. My own identity was built around noticing problems, making unusual connections and producing the thing.

AI has made strategic thinking more accessible because I can remain in conversation with an idea for much longer. I can ask how a company like this would fail. I can ask which neighbouring markets are growing. I can ask what the strongest counterargument is. I can test ten framings without arranging ten meetings. I can move between product, technology, pricing, distribution and positioning without pretending I already understand all of them.

This does not mean the resulting strategy is correct. AI is extraordinarily good at giving weak ideas a persuasive structure. Almost any product can be placed inside a growing market, connected to a major trend and described as the missing layer between several established categories. A few prompts can make a consultancy sound like a platform and a feature sound like an industry.

The fluency is dangerous because strategy can become another form of production. Instead of generating screens, I can generate business models, market maps and immaculate narratives. The output feels smarter, but the basic avoidance may remain unchanged. I am still making something inside the computer rather than exposing the idea to customers.

There is nevertheless an important difference. I am beginning to recognise which questions belong at the beginning.

An idea now has to compete for my own commitment before it gets the benefit of my execution. I want to understand what is changing in the world, what that change makes newly scarce, who experiences the scarcity as an expensive problem and what position I could occupy before the answer becomes obvious to everyone else.

This is a much more demanding way to think. It requires choosing among possibilities at exactly the moment when AI is producing more possibilities than I could ever pursue. The abundance of execution creates a shortage of conviction.

Perhaps this is what strategy becomes in the AI era. The central skill is no longer generating the largest number of plausible futures. Machines are already very good at that. The skill is deciding which future deserves contact with reality.

For makers, AI offers an intoxicating promise: you can finally build almost anything you imagine.

It also reveals the uncomfortable part that came before the building.

You have to decide what matters.

Connected reading

• The AI Was Easy. Understanding the Company Was Hard.
• The Product Is No Longer the Product
• Every Interface Contains a Business Model
• The Product Needs a Harness
• What Kind of Worker Is AI Allowed to Become?

AI authorship note

This essay began with Kris Haamer noticing that AI had pushed him to think seriously about business strategy for the first time while developing HAAM as a data-driven product repair company. AI helped pressure-test the market, adjacent categories, timing, positioning and possible business model, then supported the drafting process. Kris rejected the generic productivity angle, insisted on a personal maker's perspective, shaped the argument around execution no longer providing shelter from strategy and made the final publication decision.