What AI Adoption Actually Does to a Firm's Turnaround Time

"AI makes accounting firms faster" is the kind of claim that shows up constantly in software marketing and rarely gets examined honestly. It's worth separating what genuinely speeds up a firm's turnaround time from what's a smaller, incremental gain dressed up as transformational.

Fact: the stages of client work that see the clearest turnaround improvement from AI adoption are the ones that were previously slow specifically because they were manual and repetitive — data entry from receipts and invoices, first-pass categorisation, draft reconciliation, generating an initial set of management accounts from clean data. These stages can genuinely compress from hours to minutes once AI handles the mechanical extraction and matching, because the bottleneck really was manual processing speed.

Where the turnaround gain is smaller than expected

Stages that were already bottlenecked by something other than manual processing speed see much smaller gains. Waiting on a client to send missing documents doesn't get faster because a firm adopted AI. A judgement call requiring genuine consideration — how to treat an ambiguous transaction, how to advise on a specific decision — doesn't compress much either, because the bottleneck there was thinking time, not typing time, and AI assistance speeds up drafting a first pass at that judgement more than it speeds up the judgement itself.

Analysis: the realistic aggregate effect is that overall turnaround time improves, but unevenly — dramatically on the mechanical stages, modestly on everything else — which means the proportion of total turnaround time spent on genuine judgement and client communication actually goes up, even as absolute turnaround time goes down. A firm that measures success purely by how much faster the whole process is will see a real number, but it understates what's actually changed: the composition of where time goes has shifted, not just the total.

Opinion: firms that get the most durable benefit from this aren't the ones chasing the biggest headline turnaround-time reduction — they're the ones that notice the shift in time composition and deliberately reinvest the reclaimed hours into the judgement and communication work that client relationships actually depend on, rather than simply taking on more volume at the same margin per client. Faster mechanical processing is only valuable if what fills the freed-up time is actually worth more than what it replaced.

Prediction, held loosely: turnaround-time improvements from AI adoption are likely to keep being the headline metric firms advertise, because it's the easiest one to quantify, even though the more consequential change is qualitative — what a firm's people actually spend their time doing shifting from mechanical processing toward review, judgement and client-facing work. The firms that talk about the second thing, not just the first, are probably the ones that have actually thought it through rather than just adopted a tool.