January produces more AI intent than any other month, and by March most of it has evaporated into a slide deck nobody opens again.
The intent was not the problem. The absence of a sequence was.
Short answer: a first quarter that actually produces something follows a fixed shape. Inventory and ranking in the first two weeks, mapping and rules through the end of the month, a build phase that stays in draft-and-approve, and a live phase measured against a baseline captured before anything started. Ninety days, one project, one decision at the end.
Why January intent usually fails by March
The pattern is consistent across companies of very different sizes. Leadership commits to AI as a priority in January. Someone is asked to look into it. A few tools get trialed. By March, the trials have quietly stopped, the priority has been reasserted in a different meeting, and nothing has moved.
The failure is rarely enthusiasm. It is sequencing. Intent without a week-by-week shape drifts toward whatever is most urgent that week, which is never the AI project.
Days 1 to 15. Inventory and ranking
List eight to fifteen candidate workflows from across the business, each with a rough cost in hours, delay, or lost opportunity. Score them on value, feasibility, risk, and adoption readiness, and name an owner for the top candidate before the month ends.
This phase produces the single most valuable artifact of the quarter: a ranked list with a name attached to the first item. Everything after this is execution.
Days 16 to 30. Mapping and rules
Map the chosen workflow with the people who actually perform it, including the exceptions they handle by instinct and have never written down. Define what the system will do, what stays with a person, and what the success metric will be.
Capture the baseline number now, from records rather than memory. This is the step most programs skip, and its absence is why so many teams cannot answer whether anything worked by June.
Days 31 to 60. Build in draft-and-approve
The system drafts. A human approves everything. No autonomy in this phase, regardless of how well early testing goes.
This is not caution for its own sake. It is how you find the exceptions the mapping session missed, without a customer or a regulator finding them first.
Days 61 to 90. Live operation and the gate
Run it against real volume with the approval model in place. Review the log weekly. At day ninety, compare against the baseline and make one decision: expand, adjust, or stop.
All three are acceptable outcomes. The only failure is reaching day ninety without a clear answer, because that means the measurement was never set up properly in January.
What derails this timeline
| Derailment | Where it shows up | The fix |
|---|---|---|
| No named owner | Weeks 1 to 15 | Do not proceed past ranking without a name |
| Scope creep mid-build | Weeks 31 to 60 | Freeze scope after mapping ends |
| Autonomy granted early | Weeks 31 to 60 | Hold draft-and-approve for the full quarter |
| No baseline captured | Weeks 61 to 90 | Reconstruct from CRM or accounting timestamps if missed |
Why ninety days and not a shorter or longer window
Shorter windows rarely surface the exceptions that determine whether a workflow actually works at real volume. Longer windows lose momentum, because the organization’s attention moves to whatever the next quarter’s priority becomes.
Ninety days is long enough to hit the messy cases and short enough that the same leadership team who approved it in January is still paying attention when the gate arrives in April.
FAQ
What if the first ninety days produces a stop decision?
That is a legitimate and useful outcome. You learned what the workflow actually costs and why the approach did not fit, which is cheaper than a year of an unexamined subscription.
Can we run more than one workflow in the first quarter?
Resist it. Attention is the scarce resource in the first ninety days, not budget. One project done properly teaches the organization the method for the next one.
Who should own the first project?
Someone with allocated hours and standing in the department affected, not necessarily the most senior person in the room.
What if leadership wants to see results before day ninety?
Share the log and early observations honestly, but hold the formal decision for the gate. Deciding early defeats the purpose of measuring at all.
Does this apply to a small business the same way?
Yes, with a smaller candidate list. The sequence matters more than the scale.
Where to go next: Start the inventory this week. Eight to fifteen candidates with real numbers is a few days of asking, and it is the input the rest of the quarter depends on.
