The start of a new fiscal quarter is when a lot of plans get rewritten, and AI is usually one of the casualties, either added back in with fresh enthusiasm or quietly dropped because January’s version never survived contact with the calendar.
A roadmap that only exists as a slide from a planning offsite is not a roadmap. It is a snapshot of a mood.
Short answer: a usable AI roadmap has three horizons, not one long list. What ships this quarter, what gets scoped for next quarter, and what stays parked until a dependency clears. Each item needs an owner and a decision date, or it is not on the roadmap, it is on a wishlist.
Why quarterly resets are the right moment
Budgets reset, priorities get reargued, and teams have just enough attention freed up to look at something beyond the immediate fire. That combination does not happen in the middle of a quarter, which is why a roadmap built then rarely survives.
If a January roadmap has already drifted, the fiscal quarter reset is a legitimate second chance, not a failure to admit to.
The three horizons
This quarter
One or two projects, already scoped, with an owner and a decision gate before the quarter ends. Anything more than two is usually a sign the list has not been prioritized yet, only collected.
Next quarter
Candidates worth scoping now so they are ready to move the moment this quarter’s project clears its gate. Scoping here means a rough cost and value estimate, not a full build plan.
Parked
Real ideas with a real blocker, a data problem, a tooling gap, a dependency on another team’s project. Naming the blocker explicitly is what separates a parked item from an abandoned one.
What goes wrong with most roadmaps
| Common mistake | What it produces | The fix |
|---|---|---|
| Everything listed as this quarter | Nothing actually finishes | Cap active projects at two |
| No owner per item | Items drift without anyone noticing | Require a name before it counts as scoped |
| No decision date | Projects run indefinitely | Set a gate date when the item is added |
| Built once a year | Stale by the second quarter | Revisit at each fiscal reset |
What belongs on the roadmap versus what does not
A roadmap is not a running list of every AI idea anyone has mentioned. It is the short list the business is actually committing time and money to in the near term. Ideas without a plausible owner do not belong on it yet, no matter how promising they sound in a meeting.
How this differs from January planning
A January roadmap is often built from a blank page, heavy on ambition and light on evidence. A quarterly reset roadmap has the advantage of a completed or in-flight first project to learn from. Use that evidence. What the first project’s fully loaded cost turned out to be should directly shape what gets scoped next.
FAQ
How long should this take to build?
Half a day for the ranking and horizon assignment, assuming the underlying use case audit already exists.
What if nothing shipped last quarter?
Diagnose why before adding new items. A roadmap built on top of an unaddressed execution problem repeats the same failure.
Should every department have its own roadmap?
One company-wide roadmap with department tags works better early on. Separate roadmaps tend to duplicate effort before the organization has built the muscle.
How many items should be in the parked category?
As many as are real, provided each has a named blocker. An empty parked list usually means ideas are being discarded rather than tracked.
Who owns the roadmap document itself?
One person, distinct from the people executing individual items, responsible for keeping horizons current at each reset.
Where to go next: Use this quarter’s reset to rebuild the roadmap around the three horizons. If January’s version already drifted, this is the natural point to try again.
