Why Faster Work Can Mean Busier Accountants

Working papers are finished faster. Why does the day still feel so full?

Preparing meeting minutes and working paper drafts is faster than it used to be. Formatting reports takes less effort too. Yet I find it difficult to say that the working day has become correspondingly easier. In my experience, AI’s convenience and a demanding workload coexist.

Time saved does not automatically become spare capacity for the individual. If the team becomes smaller or more work is assigned, each task may finish sooner while the day’s workload grows. Understanding that difference requires looking at how teams allocate work alongside the speed of the tool.

1. Drafting time and completion time

AI makes it easier for new associates to start working papers and organise responses to review comments. They still need to request information from clients and establish whether what arrives is sufficient for the review.

If a file omits a required period or supporting evidence for testing a transaction is missing, someone must follow up. Without the necessary material, a draft working paper does not make it possible to finish the work.

In-charges can also spend less time writing and organising facts. At the same time, they need to understand what team members have checked and what they have yet to grasp. Discussing issues with the client and allocating work through to the deadline remain part of the role.

Project managers and engagement leaders benefit from help with spotting report errors and drafting analyses of accounting issues. Documents intended for external readers may still need revision to fit the company’s circumstances. I have also had to rewrite AI-generated PowerPoint text because its language was too grand.

Considering these tasks together reveals what remains between producing a document and completing the work. Staffing decisions need to account for that work too.

2. Fee pressure and graduate recruitment

Fees and costs also shape a firm’s search for efficiency. If audit or other service fees cannot be increased sufficiently, or have to be reduced, the firm needs to find ways to complete the work within the available fee. Hours and team composition become part of that review.

These comparisons use firm-wide recruitment figures from each firm’s 2024 and 2025 materials. They do not measure audit-only hiring, a country-wide trend or the size of individual audit teams. Engagement demand, business structure and other factors can affect recruitment, so these figures do not establish that AI adoption or audit fee pressure caused the declines.

Separately from those recruitment figures, I see moves in practice towards running teams with fewer people in anticipation of AI efficiencies. Better output from new associates can sit alongside pressure on the in-charge to finish the work with fewer team members. Expectations of higher productivity, combined with fee and cost pressures, can strengthen the incentive to reduce staffing.

If time savings are reflected first in staffing plans, individuals may end up with more work. That helps explain how a team can be described as more efficient while its members feel busier.

3. Team efficiency and individual workload

The distinction becomes clearer when we separate total team hours from the hours worked by one person.

Suppose a task takes 20 hours in total, divided equally between two people with similar skills. AI reduces the time required to 14 hours, after which the task is assigned to one person.

Total team hours fall by 30%. But the remaining person works 14 hours, 40% more than the previous ten. AI can reduce the time required for a task while the individual’s workload changes according to how it is staffed.

Actual staffing plans also need to consider capability. A well-organised working paper does not establish that a new associate can independently handle more work. If the quality of the document exceeds the preparer’s understanding, supervision and explanation from the in-charge are still needed.

4. Staffing and time for learning

Reflecting AI’s benefits in a staffing plan starts with identifying which tasks take less time. Even if contract information can be organised faster, someone must still obtain the same client records and examine the same transactions if those requirements have not changed.

An in-charge’s workload is also difficult to assess from the volume of documents they write alone. More clients or assignments to manage may mean more time discussing issues and answering the team’s questions. Looking at actual responsibilities and hours by level helps explain why an individual has become busier.

The training discussed in the previous article belongs in the work plan too. For an associate to explain the evidence behind a judgement, the in-charge needs time to listen and give feedback. If learning depends on whatever time remains after the deadline, the busiest assignments are likely to offer the fewest opportunities.

Reducing graduate recruitment also raises a question about developing the in-charges and reviewers needed in a few years. This year’s staff costs and hours alone are an incomplete basis for evaluating that decision. Fewer opportunities to learn the work can affect the development of people who will later make judgements and supervise others.

Comparing work before and after AI adoption should include drafting time, additional client responsibilities, and time spent on review and training. It should also identify where team-wide savings have coincided with increased work at a particular level.

That information provides a basis for deciding whether saved time should go towards additional assignments, deeper review or teaching new associates. How a tool’s efficiency translates into practice is ultimately a decision about running the team.

What principles should guide a team working with AI as it starts, reviews and hands over an assignment?

The final article brings these discussions together as principles for everyday practice.