Thought Leadership

Every firm has a version of the same Monday morning. A senior calls in sick three days into fieldwork. A client still hasn’t sent in their late records. A tax return runs over and eats the hours booked for the next one. Someone has been double booked since last Tuesday and nobody noticed.

None of this is dramatic on its own. The problem is volume and pace, on a schedule with no slack. In the AICPA’s 2026 CPA Firm Top Issues Survey of 629 firms, staff workload management ranked in the top four current issues for every firm-size group above 10 employees. When capacity is that tight, ordinary disruptions become business risks.

Most firms call what follows "rework", as if it were one task. It is at least three. Treating them as one undermines the effort and importance of resource planners, and drains chargeable time from billable seniors.

Job one: finding the issues

Before anyone can fix anything, someone has to work out what is wrong: scrolling a calendar or spreadsheet, line by line, for red flags; the person now over capacity, the work that no longer fits, the hours parked against a placeholder because nobody was free when it was planned.

Because this work is done only to the depth time allows, the loudest or most urgent problems get solved, but the list of urgent problems never dies down. The overbooked associate nobody spotted is not merely an untidy cell in a spreadsheet. Left unresolved, the conflict can contribute to write-offs, missed deadlines and sustained pressure on staff.

Job two: deciding what to fix first

Not every conflict matters equally; deadlines occurring sooner, jobs with higher margins, or the most strategic clients are all different ways of slicing importance.

The profession already thinks this way about its client base. Writing in the Journal of Accountancy in January 2025, Amanda Aguillard and Anita Dennis advised firms to rank every client on satisfaction, responsiveness, profitability and referral value, and give each a capacity score based on the difficulty of the engagement against what it pays. Their question is the right one for a broken schedule too: if capacity is limited, is this engagement the best use of the firm’s resources compared with another?

Yet most schedules show every conflict as the same red cell, and the fees exposed and the client grade live in a different system, or a different person’s head. Under pressure this is the job most often skipped, and issues are resolved in the order they appear.

Job three: fixing it without breaking something else

Move the work, find a substitute, or shift the dates. But every change has a second-order effect. The replacement may be available and wrong for the file. The overloading might get kicked into next week rather than truly resolved. The junior and the senior who need to work together might not have any overlapping time for weeks. The reschedule that solves one problem creates another three weeks out.

With a lack of time and control, reworking simply creates more rework.

Where AI and automation help

The Thomson Reuters Institute’s 2026 Future of Professionals report, drawn from more than 1,800 professionals in 62 countries, found 74% now use AI tools several times a week. Scheduling disruption is a useful test of where that pays off, because the three jobs respond to technology differently.

• Finding is pattern-matching over data the firm already holds. Software should do that looking continuously, so nobody has to scroll for it.
• Effective prioritization needs the fees, margins and the client’s standing attached to each issue. Technology can surface all three; the firm sets the policy on what matters most.
• Fixing requires technology to suggest solutions that consider the before-and-after impacts on all affected resources and clients. The decision to commit stays with a human.

Two new Beeye products were built to address this:

Beeye Resolve delivers issues on live engagements in one place, on one screen, each with the hours and fees at risk and a recommended substitute already matched on skills, grade and availability, with the before-and-after impact on both people. A manager reviews each fix or confirms the whole list at once, every change is logged, and affected staff are notified. The looking and the impact are automated; the decision is not. Because every firm is different, over time the system learns which choices the firm prefers and recommends accordingly.

Beeye AI connectors bring all your scheduling data into the assistant your firm already uses: Claude, Microsoft Copilot or ChatGPT. Ask "Sophie is off Thursday, who could take over the review?" and the answer comes back grounded in your context. Agents can run recurring actions for you, and dashboards can be built with plain-language prompts to help you address strategic questions such as client prioritization and future hiring and training needs.

The AICPA’s Professional Ethics Division made the general point in the Journal of Accountancy this August: professional judgement cannot be delegated to AI, and the habits to watch are automation bias, overconfidence and anchoring on the tool’s first answer. Hand the finding and the prioritizing to machines, keep the confirming for people, and you get the efficiency without losing control.

Three questions for your own firm

• When the plan changes, how long before anyone knows what is in conflict, and who is spotting it, and when?
• How do you decide which issues need addressing first?
• When a change is made, is the second-order impact visible before you commit?

In firms where a manager or senior owns the schedule, every hour spent hunting for conflicts, guessing what to fix first and undoing last week’s fix is an hour not billed to a client. In firms with a dedicated resource manager, the cost is different but no smaller: a role built to judge where the firm’s people create the most value spends its week on clerical searching instead, and the judgement the firm hired for never gets used.

Either way, resource planning is where a firm’s most expensive hours are decided, which makes it one of the few levers that returns capacity without hiring anyone. Treating it as admin is a grave mistake.