Invest1 publisher3 min readPublished
Weekly AI tax research nearly doubled to 60% among more than 1,000 tax professionals
FloQast puts 85% of accounting teams on AI as a strategic priority and 10% on extensive use. The 60% weekly figure comes from a different survey of a different population. Subtracting one from the other tells you nothing.
The Investor · Invest desk
What happened
- CPA.com and Blue J surveyed more than 1,000 tax professionals and found 60% using AI-powered tax research at least weekly, up from 33% the prior year.
- RSM's 2026 Middle Market AI Survey found 85% of respondents saying executive leadership is more enthusiastic about AI than employees are.
- CPA Practice Advisor reported that a third of lawyers, accountants and compliance professionals were using AI their organizations had not approved.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- constraint Readiness at 17% limits how much of a rising AI budget turns into production workflow this year, and buying more seats leaves that share where it is.
- contradiction Because the 10% and the 60% measure different populations at different thresholds, the same page supports both the flattering read (one workflow cracked it) and the harsh read (nine firms in ten are stuck).
- decision With a third of professionals already working around the absence of rules, the near-term choice for firm leaders is publishing a permitted-tool and permitted-data list, an administrative expense.
- exposure An 85% leadership-employee enthusiasm gap puts the risk of a failed rollout in staff behaviour, beyond the reach of anything a vendor sells.
The 60% and the 10% come from different firms. FloQast asked accounting teams and got 85% calling AI a strategic priority against 10% using it extensively, a spread of 75 points [1][1]. CPA.com and Blue J asked more than 1,000 tax professionals and got 60% doing AI-powered tax research at least weekly, up from 33% the year before [3]. Weekly use and extensive use are different thresholds, and the two questionnaires went to different groups of people [5].
The change is 27 points in a year, an 82% rise off the 33% base [2]. CPA Practice Advisor's reading is that tax research moved because the tool solved a specific problem and professionals understood how it fit their work [6]. That reading rests on two data points twelve months apart.
For anyone selling software into accounting, start with 17%, the share of teams FloQast found ready to put rising AI investment to work [2]. Set that beside the 85% who call AI strategic and there are 68 points between intent and capacity [3]. The research reports readiness as a share of teams and does not attach a dollar figure to the spend.
Individual use is running ahead of firm deployment. CPA Practice Advisor reported that a third of lawyers, accountants and compliance professionals were using AI their organizations had not approved, with the rate rising among those who thought their organization was moving too slowly [5]. That third is more than three times FloQast's 10% extensive-use share [4], across different samples again. The remedies the piece prescribes are administrative: name the approved tools, define which client and firm data can enter them, specify when outputs require source checking or second-person review, and create an escalation path for uncertain cases [7]. It also asks leaders to commit that AI-created capacity will first go to growth, client service, higher-value work and redeployment before it becomes a headcount target [9].
There are two readings of the 10%, and the sources carry both. One says the block is trust and permission, and RSM's 2026 Middle Market AI Survey supports it: 85% of respondents said executive leadership is more enthusiastic about AI than employees are [4]. The other says the block is plumbing, and the middle-market research on pilots hitting data quality, integration, security and workforce readiness barriers supports that [8]. The first is fixed cheaply and fast, by writing rules. The second is fixed by paying integrators.
I think readiness will track realized software revenue: 17% of teams saying they can absorb investment is a ceiling on how much gets deployed [2]. The test runs the other way in one year: if extensive use climbs while that 17% stays flat, the constraint was permission, and the cheap governance work was enough.
What to watch
- Whether FloQast's extensive-use figure moves off 10% next year while the 17% readiness share stays flat.
- Whether the CPA.com and Blue J weekly tax research figure holds above 60% in a third reading or flattens.
- Whether firms publish permitted-tool and permitted-data lists, and whether the one-third unapproved-use rate falls after they do.