Home AI use runs 33 points ahead of work use in Accenture's consumer survey

The two work numbers are close enough together to divide. Twenty-nine percent weekly against 49 percent who have used AI at work at all means roughly three in five workplace users are weekly users, and the remaining 20 points, about two in five, get through most weeks without touching it ^[19]. At that ratio the scarce input inside a company is recurring occasion rather than licence coverage, and spending that widens reach mostly buys more of those 20 points.

The board-deck version of this survey is two lines: workplace use nearly doubled in a year, personal use is near-universal, keep funding the programme. What the deck omits is the base. A near-doubling that lands on 29 percent implies something close to 15 percent twelve months earlier ^[21], which is real movement off a small number and says nothing about where the curve settles. The 33-point gap between personal and workplace use is the part likely to still be there next quarter ^[18].

The cohort argument is where this gets operational. Sixty-eight percent of Gen-X told Accenture they expect to increase their personal AI use over the next twelve months ^[10], a stated expectation collected in November about a year still ahead, rather than a logged behaviour. Accenture's reading is that adoption historically starts with 18- to 24-year-olds and spreads to adjacent groups once that group reaches critical mass ^[16]. If that pattern holds here, the people who write next year's internal standards will be acquiring the habit at the same time as they are codifying it.

Habit and payment come apart most visibly in the cohort with the most habit: 30 percent of Gen-Z report using AI services daily ^[6], while 15 percent of Gen-Z personal users pay for access ^[7]. Daily use does not convert into a card on file at anything like the same rate, which is why vendor revenue leans on the employer contract, and why the internal enablement pitch is getting made so energetically.

The numbers arrive bolted to advice that serves the firm that gathered them: segment AI offerings around the proof each cohort needs rather than ship them as a single feature list ^[17]. That is worth flagging, and the way to weigh it is separating the top-line splits from the inferences drawn on top of them. The published write-up gives generational percentages without subsample sizes, and without birth-year boundaries beyond noting that the Boomer range begins at 62 ^[22]. It also does not say which tools people used at home versus at work, or whether any of the personal use was employer-sanctioned ^[23], so the shadow-AI reading of this data is available but not evidenced.

The trade-off for the current quarter can be named plainly. Money spent on coverage shows up in the reach figure, which already sits at 49 percent ^[3]; money spent on occasion means picking a handful of recurring tasks and making the tool the default route through them, which shows up in the weekly figure. Next quarter's renewal conversation will be argued on the weekly number, and the choice made now determines whether there is anything to say about it.