Why three
Ask any executive team for AI ideas and you will get forty in an afternoon. Some will be good. The problem is that most companies can only change a few things well at once. Each serious initiative needs a business owner with time, a team that knows the process, data that someone has to clean, and change that someone has to lead. Those are scarce.
Three is not a magic number. It is roughly what an executive team can sponsor, review and protect at the same time. Fewer than that and you risk betting everything on one outcome. More than that and attention spreads thin, which is how good ideas turn into stalled pilots.
Start from the numbers, not the tools
Good priorities begin with the results the business is already trying to move. Before anyone discusses models or vendors, write down the three to five outcomes that matter most in your plan for the next two years. Revenue growth in a specific segment. Margin in a service line. Cycle time on a core process. Retention of your most valuable clients.
Then ask each business leader to bring the ideas that would move their outcome. This keeps the list anchored. An idea that cannot be tied to one of those outcomes goes on a separate page. It may still be interesting. It is not a priority.
Four questions for every idea
Once the list is tied to outcomes, test each idea against four questions. Keep the scoring simple, high, medium or low. The goal is a shared conversation in the room, not false precision in a spreadsheet.
Value
If it works, how much does it move the outcome it is tied to, and how soon?
Readiness
Do we have the data, the process clarity and the people to do this now?
Ownership
Is there a business leader who wants this and will be measured on it?
Learning
Will doing this teach us something we need for what comes next?
The fourth question is the one teams skip. Your first priorities are also how the company learns to work with AI. An initiative that builds a reusable data asset, or proves a new way of working that other teams can copy, is worth more than its own business case suggests.
Readiness deserves honesty too. An idea with high value and no usable data is not a first priority. It may be a second-year priority, with a first-year project to get the data in shape.
Balance the three
Look at the shortlist as a portfolio rather than a ranking. A strong first set usually has some mix of the following.
- One initiative aimed at growth: something clients will notice, such as faster quotes, better service or a new offer.
- One aimed at efficiency: a costly, repetitive process where the savings are visible and the risk is manageable.
- One that builds a foundation, often data, sometimes a capability, so the next round moves faster.
The balance protects you. Efficiency projects pay back sooner and buy patience. Growth projects keep the board and the front line interested. Foundation work is what makes the second year easier than the first.
The hardest part of choosing three priorities is saying no to the fourth.
Stop the rest, out loud
Choosing three only works if the other ideas actually stop, or at least wait. That means telling the sponsors of the ideas you did not pick, explaining why, and saying when the list will be looked at again. Without that, the unfunded ideas carry on quietly with borrowed time and borrowed people, and the three priorities never get the attention they need.
Then set a date, six months out, to review the list. Some priorities will be on track, some will need to change, and a few ideas from the second page will have become ready. That review, done honestly, is how three priorities become a program.
