The “not yet” column: how to decide what AI should wait on.

Saying no to a good idea is among the most valuable things a readiness assessment can do. Here is how to tell what should wait, and what would make it ready.

Every AI roadmap holds two kinds of ideas: the ones the business can absorb now, and the ones that look better on a slide than they will in production. Most plans treat them the same. We sort them into three groups instead: ready now, needs groundwork, and not yet.

“Not yet” is not “no”

A “no” closes the conversation. A “not yet” is a decision with a condition attached. It names what is missing and what would change the answer. That is what makes it useful: it turns a disappointing idea into a defined piece of work, and it keeps good ideas from dying in a backlog.

Four reasons an idea belongs there

• The data is not trustworthy yetThe idea depends on information that is incomplete, inconsistent, or out of reach of the people who would use it.
• The process underneath is unstableIf the workflow changes every month, or no one can describe it, there is nothing solid to improve.
• Nobody owns it after launchWithout a named owner and a measure, even a good result fades once the project team moves on.
• There is no plan for when it is wrongNo review path and no accountability means a single visible mistake can end the program.

Groundwork or not yet?

The difference is the size of the gap. An idea needs groundwork when the gap is narrow and can be closed as part of the work, such as cleaning up a single data source. It is “not yet” when the gap is broad, spans several foundations at once, or sits outside the team’s control. A useful rule of thumb: if closing the gap would take longer than the idea stays relevant, it should wait.

Make the list visible

For every “not yet,” write down three things: what would make it ready, who would do that work, and when you will look at it again. Done well, the parking lot becomes a roadmap. Review it each quarter, because foundations improve and some ideas will move up.

The cost of skipping this step

Building on a foundation that is not ready does not fail quietly. A team spends months proving something its data, processes, or rules could never support, and the visible miss makes the next project harder to fund. Waiting on the right things is how the right things eventually succeed.

Try it this week

Take your list of AI ideas and place each one in one of the three groups. Beside every “not yet,” write one sentence on what is missing. If you cannot say what is missing, you have found something worth finding out. The Fathom Index builds this same map with evidence behind every placement.