Turning talk into results: where AI creates value, how to measure it, and where to start.

There’s been plenty of talk about artificial intelligence over the past two years. But as an investor active in the field for almost a decade, the real question for me is this: “Where is AI actually creating measurable value today?” That question is my compass. It helps me separate noise from substance, and marketing hype from real results.

Three key observations

• AI is a multiplier. It accelerates what already exists. When well planned, it amplifies human expertise; when not, it just drives you faster toward failure. Before automating, you need to clarify the problem to be solved, assess the quality of your data, and define the process you want to support. A project succeeds when AI delivers a measurable operational advantage and is built on a solid plan. For example, if your business model is already losing money, poorly applied AI will just make you lose it faster.
• Discipline beats enthusiasm. The best teams I have supported set simple, easy-to-read indicators right from the start: cost per sale, processing time, error rate, customer satisfaction. They move forward gradually, trim what doesn’t deliver, and keep a record of the risks.
  AI must live in a dashboard, not on a marketing poster. The difference is clear when management links AI to measurable goals and concrete decisions; conversely, when it is treated like a magic wand, enthusiasm quickly fades.

• The main risk is not technical, it is human: resistance to change. Projects do not derail because of an algorithm, but because habits are disrupted without proper support. In this context, people may sometimes, consciously or unconsciously, try to prove that it is not worth the effort.
  To succeed, you have to anticipate this tension, clarify roles, provide training, explain who saves time, who is freed from repetitive tasks, who keeps the final decision, and how biases and customer data confidentiality are handled.

Why talk about it now? Because every company can now gain significant benefits from AI, and automotive dealerships are no exception. The advantage no longer comes from a “technological miracle,” but from the ability to generate real-world results that serve a specific department.

This column opens a series. In the next one, I will move from general to practical by answering a simple question: how can what I have just outlined be applied in the automotive sector? We will look at three areas where the impact is immediate: decision support in F&I (speed, fairness, compliance); customer experience in dealerships (prospecting, qualification, follow-up, personalization); and operational efficiency (inventory, demand forecasting, after-sales service). I will share what works, what fails, the pitfalls to avoid, and straightforward benchmarks for estimating return on investment. My goal is not to celebrate a trend, but to provide practical, hands-on criteria.

See you in the next issue, where we’ll take AI out of the clouds and put it to work in the dealership.