Clear the workpaper review queue by making sure only complete files reach a reviewer. Most review notes in busy season are about missing references, unsupported figures and unanswered prior notes. AI can check for those before a file is submitted, list what is missing and draft the reply to each note, while managers keep the actual review.
#Key takeaways
- The review queue backs up because reviewers spend their time on preventable notes, not judgement calls.
- A pre-review check catches missing cross-references, unsupported numbers and open prior notes before submission.
- AI drafts and checks; managers and partners still review and sign off every workpaper.
- Start with the workpaper types that generate the most notes, not the whole file.
- Measure notes per workpaper and how often a file goes back, before and after.
#Why does the review queue stall in busy season?
Associates finish testing late and submit in batches. The manager opens a workpaper and finds a total that does not tie to the lead schedule, a tickmark with no legend and a prior-year note nobody answered. The note goes back, the associate is already on another client, and the file waits again. The manager's week fills with the same notes on different files.
The pressure sits mostly in the middle of the firm. A Distinct survey of tax and audit professionals found that about three in four senior accountants at public firms called busy season stressful, against 22% of associates, and nearly 80% of respondents worked more than 51 hours a week (CPA Practice Advisor on a Distinct survey, august 2025).
#What can a pre-review check catch?
- Figures that do not tie to the trial balance, lead schedule or prior workpaper they reference.
- Tickmarks used without a legend, and conclusions left blank.
- Support referenced in the workpaper but not attached to the file.
- Review notes from the prior year or a previous round that have no reply.
- Sign-off fields missing for preparer or date.
Most of these are mechanical, which is why AI helps here. The check runs on the file, lists the issues with a link to each, and the associate fixes them before pressing submit. The reviewer then sees a cleaner file and spends their notes on judgement.
#How do you roll this out without slowing the team down?
- Pull last season's review notes and group them by type; pick the three most common.
- Agree with managers what "ready for review" means for those three, in writing.
- Run the check on a closed engagement and compare what it finds with the notes reviewers actually wrote.
- Turn it on for one team, with associates running it before submission, not managers after.
- Track notes per workpaper and send-backs per file, and drop checks that only add noise.
#When do you not need AI for this?
If your audit software already has completion checks and the team uses them, switch those on fully first. A firm-wide checklist and a clear "ready for review" standard fix a lot on their own. AI is worth adding when checks need to read the content of a workpaper, a scanned confirmation or an Arabic bank letter, not just whether a field is filled.
The reviewer stays in charge either way. PCAOB staff found that firms investing in generative AI expected it to augment, not replace, people in the audit (PCAOB staff, july 2024).
Also be honest about capacity. If the queue is long because the team is short of reviewers, a check reduces the noise but does not create a manager. That is a staffing decision.
#Where does Stennir fit?
Stennir runs an AI consulting audit on your engagement workflow to see whether the review queue is the right first build, or whether PBC handling or extraction would pay back sooner. Then AI implementation builds the pre-review check with your managers deciding which checks run and every result logged.
AI training runs short role-specific sessions for associates and seniors on their own files, with usage and send-back rates measured before and after. Read how a similar approval queue works in AI for approval workflows.
We are
Stennir builds checks that run before review, so managers and partners spend their review on judgement.
We aren't
Stennir does not review or sign off workpapers, and it does not replace the reviewer.
#What should an audit manager do next?
Count last season's review notes by type on two engagements. If more than half are mechanical, a pre-review check is worth testing. See more AI for finance teams, then take the free readiness assessment or book a 30-minute call.
#Common questions
Does it work inside our audit software?
It depends on what your software exposes. Many firms export workpapers or work from files in a document system the check can read. Stennir looks at that during the audit and says plainly whether an integration is practical, or whether a simpler checklist inside your existing software would do the job.
Will associates just rely on the check instead of thinking?
The check covers mechanical gaps only: ties, references, attachments, sign-offs. It says nothing about whether the conclusion is right. Training makes that line explicit, and reviewers keep writing judgement notes. If send-backs on judgement points rise, that is a training signal, and it shows up in the numbers.
What do you need from us?
Last season's review notes, a few closed engagement files, your "ready for review" standard if one exists, and one manager who can decide which checks matter. Stennir works on your own files, so the checks match how your firm actually documents an audit.
Where is client data processed?
In the hosting region your firm and your clients' contracts require, agreed in the scope before any file is processed. Your firm owns the accounts, the data and the off switch, every check is logged, and no client data is used to train a public model.
