Why Most Businesses Will Get Their AI Investment Thesis Wrong

There is a question I keep hearing in PE advisory conversations: "Is this company's AI strategy real, or is it a slide deck?" More often than I would like, the answer is the latter.

Most businesses approaching AI investment are making a version of the same mistake. They are evaluating AI as a technology procurement decision when it is actually an organizational redesign decision. The investment thesis focuses on which platform, which vendor, which use case. It should focus on whether the company has the talent, the structure, and the patience to absorb what AI demands.

The three mistakes I see repeatedly.

Mistake 1: Treating AI as a department. The most common pattern is a newly created "AI team" of three to five people, often reporting through technology or innovation. This team builds a pilot. The pilot works. And then nothing happens, because no one restructured the commercial, operations, or revenue teams to actually integrate AI into daily decisions. AI is not a department. It is a capability that must be distributed across the organization.

Mistake 2: Measuring AI like a software project. Boards want ROI in two quarters. AI does not work that way. The most successful deployments I have been part of took 18 to 24 months before the compounding effects became visible. Early wins are possible, but they are not the measure of whether the thesis is right. The measure is whether the organization is making fundamentally better decisions, and that takes time to observe.

Mistake 3: Ignoring the middle layer. Most AI investment conversations focus on two extremes: the customer-facing experience (chatbots, personalization) or the back-end infrastructure (data lakes, model training). The opportunity being overlooked is the operational middle layer: the thousands of repetitive, judgment-light decisions that consume disproportionate human time. Pricing. Forecasting. Supply Chain. Service Recovery. Inventory allocation. Campaign optimization. This is where AI delivers the fastest, most measurable impact, and it is where the investment thesis should start.

What a good AI investment thesis looks like. A credible AI investment thesis for a hotel group answers four questions:

1. Talent: Does the leadership team include people who have built and scaled AI products, or are they learning on the job?
2. Structure: Is AI embedded in commercial and operational decision-making, or is it a parallel track?
3. Patience: Does the board understand the 18-to-24-month timeline for compounding returns, and are they capitalized accordingly?
4. Starting point: Is the first deployment targeting the operational middle layer where ROI is fastest, or is it chasing headline-grabbing use cases?

If the answer to any of these is unclear, the thesis needs more work.

The cost of getting it wrong.

The businesses that get their AI thesis wrong will not fail spectacularly. They will fail slowly. A pilot here, a write-off there, a gradual widening of the gap between them and competitors who invested with discipline. By the time the board realizes the strategy was never real, the window has narrowed.

The ones that get it right will compound their advantage. Better decisions, faster. Lower cost to serve. Higher guest satisfaction. And a talent flywheel that attracts the people who want to build, not just maintain.

The AI investment thesis is not about technology. It is about whether a company is willing to change how it works. That is the question worth asking.