New benchmark finds AI-mature teams spend 34% of their time on manual work, compared with 63% among organisations at the earliest stage
Finance teams that have embedded AI into everyday accounting workflows are closing their books two days faster and spending half the time on manual work compared to those at the earliest stage of adoption, according to new research from FloQast.
The State of Accounting AI 2026 report found that the most AI-mature finance operations complete the monthly close in an average of 6.7 days, compared with 8.7 days for the least mature. Manual work accounts for 34% of finance teams’ time at the most advanced organisations, versus 63% at those in the earliest stage.
Despite this performance gap, most finance teams have yet to translate AI ambitions into widespread adoption. Although 85% of respondents said their leadership considers AI a strategic priority, only one in 10 reported extensive use across accounting workflows. Four in five (82%) described adoption as limited or partial.
The independent survey of accounting and finance professionals in the UK and US identified five levels of AI maturity. While a small group of organisations is seeing measurable operational improvements by making AI part of everyday accounting, most are in the middle or early stages of AI implementation.
The finance teams pulling ahead are going beyond using more advanced tools and building the controls, governance and internal capabilities needed to introduce AI across repeatable, measurable accounting processes.
That approach is missing within most organisations, with 51% of finance teams saying their financial controls are either informal or inconsistently applied across the close process. While 55% of finance leaders identified account reconciliation as the biggest opportunity for AI, only 5% reported a high level of automation.
In comparison, the most advanced organisations have moved beyond isolated experiments, with 69% actively executing an AI roadmap and 95% already embedding AI into the month-end close.
Investment in AI is accelerating among finance teams, however it’s increasing faster than the ability for teams to develop the capabilities needed to use it. While 92% of finance decision-makers expect AI investment to rise over the next two years, only 17% believe teams are ready to put that investment to work.
Hugh O’Neill, Principal, Accountant in Residence at FloQast, said: “For the past few years, the conversation has centred on whether finance teams should adopt AI, but that question is now largely settled. The real issue is that most organisations still haven’t figured out how to use it effectively.
“A small group have already changed the way they work and are seeing clear gains in speed, efficiency and capacity. Too many organisations, however, are simply layering AI onto existing processes instead of redesigning those processes around what AI can do. That limits the impact they’re able to achieve.
“That gap is no longer theoretical. It’s already showing up in how quickly teams can close, how much manual work they’re doing, and ultimately how much value finance teams can create.”
“AI isn’t something you buy your way into — it’s something you build toward,” said Jonathan Mears, VP of Finance and Accounting, Liquid AI. “We made thinking about AI our default way of working, and with FloQast, that mindset snowballed into massive time savings with thoughtfully documented processes and reliable change management. The result is a faster, more consistent close, far less manual prep, and a team that spends its time reviewing and analyzing instead of preparing. It elevated our accountants; it didn’t replace them.”
The report recommends that organisations in the earlier stages of adoption begin by documenting and standardising their workflows, strengthening financial controls and establishing clear governance. Rather than searching for a single tool to transform the finance function, teams should initially focus on well-defined, repetitive processes where automation can deliver measurable value.
For more advanced organisations, the opportunity lies in redesigning additional workflows around AI and ensuring that every output remains traceable, defensible and subject to appropriate human oversight.
