Where to start with AI when everything feels urgent

Every leadership team we talk to feels the same pressure: do something with AI, and do it now. The risk isn't moving too slowly. It's moving fast in the wrong direction — and spending a year proving AI "doesn't work here" when the real problem was where you started.

The good news is that picking a strong first move is a skill, not a gamble. Here's the approach we use with clients to find a first project that ships, earns trust, and creates momentum for the next one.

Start where the pain is expensive and the data already exists

The best first AI project sits at the intersection of three things: a problem that costs real money, a decision that gets made over and over, and data you already collect. When all three line up, you can show value in weeks instead of quarters.

• Expensive and repeated: the same judgment call made hundreds of times a week is where small improvements compound fast.
• Data already exists: if the project needs a year of data cleanup before it starts, it's a foundation project, not a first win.
• A clear owner: someone whose job gets measurably easier — and who will champion it internally.

The goal of your first AI project isn't to be impressive. It's to be undeniable.

Avoid the two traps

Most first projects fail in one of two ways. The moonshot tries to transform the whole business at once and never ships. The science experiment ships something technically clever that no one actually uses. Both burn the credibility you need for everything that comes after.

The test we apply

Before committing to a first project, we ask one question: if this works, will a real person change what they do on Monday morning? If the answer is fuzzy, the project is too abstract. If it's obvious, you've found your starting point.

Plan the second move before you finish the first

A first win is only worth it if it opens a door. As you scope project one, sketch where it leads — the foundation it strengthens, the next decision it could improve, the team it could expand to. That's how a single project becomes a strategy instead of a one-off.