The Workpaper Review Problem Accounting Firms Stop Noticing

Review is where every upstream problem in an engagement becomes a downstream problem for the firm to solve.

The client delivered documents late, delivered the wrong ones, or delivered them in three different formats across two email threads and a portal message. The testing environment didn't talk to the workpaper tool, so someone re-keyed information that already existed somewhere else. A version drifted between handoffs. None of this registered as a crisis when it happened. But by the time the workpapers reached the reviewer, the reviewer wasn't reviewing, they were catching.

That's the quiet version of how review gets expensive, and most firms never see it on a timesheet.

The headcount constraint for accounting firm audits

Manual workpaper review is the last stop in an engagement before sign-off. That makes it the place where every inefficiency from earlier in the process shows up, compressed into whatever time is left before the deadline.

While review itself isn’t hard, it scales linearly with volume: more engagements simply mean more reviewer hours, while every other part of the firm is being asked to do more with less. That combination creates a structural constraint or a point at which adding clients doesn't add margin. The review function can't absorb the load without adding headcount, and the headcount isn't there to add.

Firms can see the ceiling in its symptoms, like stretched senior reviewers, compressed timelines, comments that come back faster than they should, or partner involvement in work that was supposed to be resolved at the manager level. But the ceiling is the cause. Everything else is a consequence of hitting it.

Where the audit review fails

Review fails the same way the engagements leading up to it fail, through a dozen small problems that each seem manageable and together become expensive.

The most common failure points:

• Tie-out rework. Preparers and reviewers are working from different versions of the same document. A correction gets made, reviewed, and then re-reviewed when the underlying file changes because nobody flagged that the source had been updated. The workpaper reflects what the evidence said last Tuesday, not what it says now.
• Disconnected review comments. Feedback lives in many different places: email, tracked changes on a Word document, a marked-up PDF that was emailed to the wrong person, or a verbal conversation that happened without a notetaker. The reviewer has no single place to see what's been addressed and what hasn't. Neither does the preparer.
• Documentation quality that varies by preparer. Some preparers document to the standard the reviewer expects, others don’t. In that case, the reviewer has to compensate: adding context, reformatting evidence, writing the explanation that should have been there when the document arrived. Some of this traces back further than the review stage, to how the original PBC request management was handled and what expectations were set with the preparer at the outset. That's review time spent on documentation, not on judgment.
• Late-stage discovery. An issue that would have taken twenty minutes to resolve in week two of fieldwork takes three hours to resolve in the final review push. By then the engagement timeline is compressed, the client has moved on mentally, and getting a corrected document requires re-engaging someone who considers the engagement closed.

None of these failures look like a breakdown from the outside. The engagement closes. The deliverable goes out. But the hours spent on each breakdown were reviewer hours, the most expensive and most constrained resource in the firm, spent on problems that originated somewhere earlier in the process.

The accounting talent shortage is draining your margin

Firms have traditionally absorbed review inefficiency through headcount. When review took longer than expected, the answer was to have the senior reviewer work later, pull a manager in, or staff the next engagement more heavily on the review side. That solution is increasingly unavailable.

The accounting profession has been running a practitioner deficit for most of this decade. The number of first-time CPA exam candidates fell from 48,004 in 2016 to just 28,082 in 2024, a drop of more than 40 percent, according to AICPA's own Trends Report data, even as demand for assurance services has held steady or grown. In AICPA's most recent Top Issues survey, finding qualified staff ranked as the single biggest concern for firms of nearly every size, ahead of every other challenge firms reported. The firms that absorbed review burden with headcount are now doing it with the same people managing more engagements, fewer of whom are moving up the pipeline to replace the ones who left.

What this means in practice is that the reviewer who used to carry four concurrent engagements is now carrying six. The comments that used to come back in two days come back in four. The partner who used to do a light final review is doing a full one because the manager didn't have the bandwidth to get it where it needed to be. None of this shows up as a crisis. It shows up as a margin that quietly disappears.

Rushed review is where audit quality risk is highest

The biggest risk in a compressed review cycle isn’t inefficiency but the quality risk that comes with it.

Manual review under time pressure fails selectively not uniformly. Perhaps reviewers skip to the areas they already understand, spend less time on sections where the risk is ambiguous, and rely on pattern recognition rather than fresh evaluation. These are reasonable human responses to an unreasonable situation, but they create blind spots. Exceptions get missed, documentation gaps pass through that shouldn't, and the issues that turn into peer review findings or QC problems trace back to those same blind spots.

Rework often traces back to a review that was faster than the engagement required, on a timeline that didn't leave room for anything else. That asymmetry only gets worse when the cost of client delays earlier in the engagement has already eaten into the time review was supposed to have in the first place.

The reputational risk here is real and asymmetric. A peer review finding, a revised opinion, or an engagement reopened after close all cost more than a thorough review ever would. The price shows up in partner time, client relationships, and firm reputation, not in the hours it would have taken to catch the problem the first time. The math on rushed review always looks better before the problem surfaces than after.

Where AI and automation help in accounting audits

Automation reducez the volume of work that reaches review without passing a basic completeness and consistency check, which is a different, and more honest, description of what these tools actually do.

The specific use cases where AI-powered document review is delivering improvement in audit workflows today include:

What the tool does	What it catches before review	What the reviewer still does
Automatic document and support type classification and routing	Mislabeled files, missing items, wrong formats	Evaluate whether the error is caught before it gets to review
Mathematical accuracy checks	Math errors in totals, subtotals, and cross-footed figures	Determine if the anomaly is an error or an explainable exception
Cross-reference checking across workpapers	Mismatched figures between the financials, footnotes, and prior-period report	Assess materiality and decide on adjustment
Status tracking across request items	Open items that haven't received a response	Prioritize follow-up and decide when to escalate

We call this is audit automation software. What it does is shift the review function from catching problems that shouldn't have reached review to evaluating problems that actually require human judgment. That's a meaningful change in what review hours are spent on, even when the total hours don't change dramatically at first.

The firms getting the most out of these tools aren't the ones that implemented them hoping to reduce headcount. They're the ones that implemented these tools to stop wasting reviewer time on problems that could have been caught earlier. The result: review gets both faster and better, because reviewers are finally spending their time on the part of the job that actually requires judgment. For a closer look at where these tools fit into a modern engagement workflow, see our guide to AI tools for accounting firms.

The constraints and costs are real

Each of the three previous posts in this series described a different place where engagement margin disappears. The first traced it to friction in the client experience that drives quiet attrition. The second traced it to the Client Readiness Gap, which transfers the burden of client disorganization onto the engagement team. The third traced it to the hidden cost of stitching five disconnected tools together across a single engagement.

Manual review is where all three of those causes land. It's the final catch point, which means it's also the place where everything that slipped through upstream becomes most expensive to fix.

The firms that are starting to scale past the review constraint aren't adding headcount or cutting clients. They're strategically changing what review touches in the first place, by using tools to fix what comes in and standardizing what the reviewer sees. And they use audit automation software to handle the consistency work, so judgment gets reserved for what actually requires it.

This is the fourth post in our series on the operational gaps quietly driving inefficiency in audit and advisory engagements. Catch up on the framing post, the Client Readiness Gap deep dive, and the look at disconnected engagement workflows if you missed them.

The next post in this series puts numbers on the rework cycle: what a single round of late-stage corrections costs in reviewer hours, delay to close, and partner time, and what the aggregate looks like across an engagement book.