Most AI rollouts in wineries fail on the calendar before they fail on anything else. The day a leader writes "kick-off Monday" on her notepad determines whether the rollout finishes or stalls, and most leaders pick that Monday on the basis of when budget cleared rather than when the team will be able to take a new habit on board. The cost of that choice does not show up until month two, by which point it is too late to move it.
The shape of the problem is easy to draw on a notepad. A winery owner gets her budget cleared in June, decides to be running before crush, writes "kick-off Monday" on the second week of July, and works out the eight-week play from there. Month two of that arc — the week where the staffer is supposed to be crossing the habit threshold and the leader is supposed to be running a consolidation retro — lands at the back end of August, which in most California cellars is the week the lab tests come back and the harvest call starts being a question of days rather than weeks. The schedule looks reasonable on the notepad in June. It does not survive the August it produces.
This is the closing post in a twenty-post series that ran up to a webinar on employee AI engagement. Of all the failure modes I have written about in the run, this is the one I would put first if I could only warn a winery leader about one thing. The 60-day play works. The timing is what makes it work or wastes it.
The eight-week math, against the harvest calendar
The shape of the play comes from Microsoft's Work Trend Index work on AI habit formation, which I unpacked earlier in the 3x-a-week-for-8-weeks rule. The pattern is structural: a staffer who hits three uses a week and holds it for seven to eight weeks crosses the threshold where the tool stops being effortful and starts being automatic. Below that line, the habit does not form. The play I run with consulting clients is built backwards from that window.
That window has a shape. The first two or three weeks are easy because the tool is new and the leader is paying attention. Week four is the dip — novelty fades, the staffer has hit a few outputs that did not land, and her old workflow is still well-paved. Weeks five and six are where the second user joins and where the staffer who started the play needs a small amount of new input to stay on. Weeks seven and eight are where the habit either locks in or quietly dissolves.
Now drop that eight-week arc onto a winery calendar that starts the kick-off in the second week of July. The dip lands in late July. The second-user moment lands in early August. The habit-lock weeks land in late August and early September, which in most North Coast AVAs is the run-up to crush. By the time the staffer is supposed to be crossing the threshold, every other thing in her week has gotten heavier, and the new habit is the first thing she drops. She does not say she is dropping it. She just does not open the tool that Wednesday, and then the next Wednesday, and the calendar takes care of the rest.
What "AI brain fry" looks like in month two
The Microsoft threshold is the upper bound on the timeline. The lower bound is a different number, and it is the one I think a lot of winery leaders have not read yet.
CNBC reported in April on a phenomenon their sources were calling AI brain fry — heavy AI users in the second and third month of a daily-use rotation reporting cognitive fatigue and decision overwhelm. The piece documents what happens when employees who have built the use into their week start doing it on top of jobs that were already at capacity. The novelty wore off; the load did not.
The fix the piece describes is the deliberate insertion of pause and consolidation rituals around day forty-five — a short retro on what is working and a clean-out of prompts that are not, scheduled into the existing rhythm rather than bolted on as a new meeting. The teams that did not bolt that in tended to fade. The teams that did bolt it in cleared the dip.
Drop that finding onto the harvest calendar and the math gets sharper. The day-forty-five consolidation moment in a late-June rollout lands on the first or second weekend of August, which is the week most California cellars are juggling lab work and pre-crush prep. A leader who scheduled that retro is going to skip it because the calendar is shouting at her. The staffer who needed it skips with her. By the time anyone notices the rollout has stopped breathing, harvest is on and there is no oxygen to give it.
Why the seasonal cognitive load is the load-bearing constraint
A winery year is not flat. The cognitive load on the team is not flat either. Spring is shoulder-busy with bottling and getting the tasting room ready for the season. Summer is the visitor surge, which is volume-busy on the tasting room and the DTC team but lighter on cellar and production. Harvest is the year's peak load on the cellar, the lab, the production team, the GM, and anyone the cellar pulls in. Winter is post-harvest cleanup, financial close, planning for the year ahead, and the longest stretches of low cognitive load on the most senior people.
The 60-day play I run is small enough to fit into many of those stretches. It is not small enough to fit into harvest. A staffer in the middle of crush does not have forty-five minutes of focused new-habit time three days a week, because that time is already spoken for by the thing the winery exists to do. The play does not lose to AI brain fry during harvest; it loses to the harvest itself, which is the right thing for the harvest to win.
The leaders who get this wrong are usually leaders who have run the year-ahead plan in months and quarters their whole career and who plan the AI rollout the same way they plan a marketing calendar. The marketing calendar can survive a missed week in August. The habit-formation calendar cannot. You either get the eight weeks on the table or you get a Tuesday in mid-October where the staffer remembers that there was a folder, and that is the last time anyone opens it.
When to run the play, and when not to
There are two windows in the winery year that the eight-week play fits cleanly into.
The first is the off-season window from early January through late March. Kick-off in the second week of January gets the day-forty-five consolidation moment around the start of February and the habit-lock weeks running into the first week of March. The whole arc finishes before bottling presses hard and well before the visitor surge. This is the window I steer most consulting clients toward, because the cognitive load on the staffer the play centers on is at its annual low and because the second-user moment in week five lands when the team has bandwidth to add one more person to the rotation.
The second is the shoulder-season window from early November through the third week of December, which only works if the team has come up for air after harvest. Some cellars are still finishing primary fermentations through October and into early November, in which case this window does not exist for them and the January start is the only real option. For wineries whose harvest is genuinely closed by early November — most of California's North Coast in most years, much of Oregon, much of Washington — the shoulder window has the same calendar properties as the off-season one. Day forty-five lands in early December. The habit locks in before the holiday close-down. The staffer comes back in January with a tool that is already in her week.
There are two windows the play does not fit. June through August is the obvious one. The less obvious one is the four-to-six-week run leading up to harvest in any given AVA, because the staffer who is the right champion is usually the staffer whose week is filling up fastest in those weeks. The early flush of activity is not the visible part of harvest, but it is when the GM's attention starts to splinter, and a play that needs the GM's attention every week is going to lose what it needs at week three.
If you are reading this in late May and the next clear window for your winery starts in November, the right move is to use the summer for the work that should have happened before the play starts. Pick the role you want to start with — the three winery roles where AI lands fastest post is the short version of that calculus. Build the starter folder before you need it. Have the conversations about what the hated tasks are. Watch how the staffer's week shapes up across the summer so you know what you are walking into in November.
The play moves fast when it runs. The preparation can be done in slow weeks across the months that are not the play.
The pause-and-consolidate ritual the calendar usually skips
There is one piece of the play I would name specifically, because it is the piece most winery leaders skip even when the rest of the timing is right.
The day-forty-five consolidation moment is a fifteen-to-thirty minute retro between the staffer running the play and the leader who set it up. Two people, in person, with the staffer's prompt history pulled up on a laptop, scheduled into a slot that already exists in the staffer's week rather than added as a new line item. Two questions: what has she been opening the tool for and what has she stopped opening it for because it did not land. The leader's job in the retro is to listen, write the answers down, and decide one small change to make to the starter folder before the next Monday.
That retro is the moment where the play either stays alive through the dip or starts to fade. Leaders who do it find that the staffer comes out of it more confident than she went in, because the conversation legitimizes the fact that some of her early use was clunky and that some of it has started to pay off. Leaders who skip it usually find at week eight that the staffer politely says the tool was helpful and then they both move on. The CNBC piece's pause-and-consolidate finding is what this retro operationalizes. It is not optional. It is the cheapest thing in the eight weeks and it is the easiest thing to drop on a busy week.
If a winery owner cannot guarantee that the leader running the rollout will have a clear afternoon for that conversation at week six, the rollout is timed wrong. The fix is the calendar.
The webinar is built around this exact problem
The webinar that this twenty-post series has been running up to is built around the question this post raises. The hour is a live demo of the starter folder build for one winery role, run with the kind of artifacts a staffer would use on a Monday. The second half is the 60-day play itself, walked week by week against a winery calendar — including the day-forty-five retro, the second-user moment, and the read-and-respond on cognitive load that the CNBC piece pointed at. The end of the hour is a Q&A; bring the role you want to start with and the rough shape of your year-ahead calendar, and we can talk about whether November or January is the right month.
This is one of the calls I am most looking forward to running this year. The play that holds up in a small winery is now well enough documented that the failure modes do not have to be discovered by running into them. Timing — the thing this post is about — is the failure mode that costs the most when it lands and is the cheapest to fix in advance.
If you have read this series and you have been waiting to put the rollout on the calendar, the webinar is the place to walk out with the calendar drawn. Check the webinar schedule for the next date and to register. I have kept the call small enough that there is room for working questions, so if your winery is in the spot where the budget exists and the months are still up for grabs, this is the hour to bring those months into the open.
The 95% number that opened this series is what happens when leaders run the play in the wrong month. The webinar is for getting it on the right one. I will see you there.
Part of a 20-post series on employee AI adoption for wineries — see the full series under AI Adoption.