Eight weeks of conversations about planning season, and the same problem kept arriving wearing different faces. A board chair who cannot see what half a million in AI spend bought. An owner using AI daily who cannot explain how the pieces fit. A marketing team sitting on an attrition list nobody has printed. An analyst who keeps being told it is too late to bring data into the decision.

Most institutions are not bad at buying software. They are bad at articulating what they need before the demo starts, which is a different problem with a different fix. Here are four questions you should ask every vendor before signing any contract.

I keep hearing a version of this: we are not seeing the return on the data.

Then you look, and the attrition model exists. It runs. It produces a ranked list of members likely to leave, sorted by how much they have on deposit. And the list sits in a folder.

Nobody is refusing to use it. There is no political fight. It is simply that no one owns the step between the list existing and somebody picking up a phone.

Two things are happening to your data budget at once, and most plans treat them as one thing.

Some of it is getting cheap. Dashboards and reporting. Basic segmentation. Off-the-shelf models. Standard KPI reporting. Everyone will have these next year at close to nothing, which is exactly why spending more there buys you parity rather than advantage.

The rest is getting more valuable. Data nobody else has. People who know your business. Governance and accountability. Turning a model into a decision somebody actually makes.

Change management costs roughly $3 for every $1 spent on AI, and most organizations budget for neither. Here is what community FIs cannot afford to get wrong.

Brute force marketing scales faster than ever with AI. The question worth asking is whether multiplied volume compounds the right approach, or just makes the wrong one noisier and faster. This is the best way to avoid spending millions on faulty marketing.

Disciplined execution rarely shows results right away. Spain's World Cup run teaches mid-market operators what sustained pressure looks like before the scoreboard catches up.

Business problems speak before you spend. A real diagnostic surfaces what the data is actually pointing to, and the answer is rarely the tool you came in expecting. Consultants that are willing to listen and wait for the patterns to emerge are more likely to address the right problem.