How we helped an accounting firm add $1.5M in recurring profit, reported 8 weeks after completion.

They wanted to grow without asking for a longer busy season.

Their tax group was moving clients onto value pricing. The partners wanted more tax revenue and partner income, but they didn't want growth to mean a longer busy season for their team.

The new pricing model made the effort behind each return harder to ignore. Their partners could see the fee, but they couldn't see how much work it took to finish the return. Without that comparison, they couldn't tell whether a value-priced client was actually more profitable than an hourly one.

Their COO brought us in. More preparers, outside help, or a better AI tool looked like ways to increase capacity. But before recommending any of them, we needed to find out why the work already in progress wasn't getting finished.

40% of measured returns came back from first review.

We mapped the path from client documents through preparation, review, and delivery, then walked it with their tax leader and managers. During one peak week, their team logged its work every 30 minutes. That let us follow where returns waited, what information was missing, and which tasks had to happen again.

The study found that 40% of measured returns came back from first review. Most came back because of missing documents or inconsistent workpapers. Some returns had entered preparation before the client's documents were complete, so preparers had to stop and wait for the rest.

As a result, managers spent ~33% of their peak week on second passes. Review took 45 to 60 minutes per return, and the wait for a manager assignment stretched to 5 to 7 days. Adding preparation capacity would still send work into that same review queue, which is why we focused on what made returns come back.

We put those findings into a scorecard showing the work behind each return. Their COO used it with leadership, giving the busy-season capacity problem a measured cause and giving us a clear place to intervene.

We changed what reached the reviewer.

Because incomplete documents were holding up preparation, we built a readiness check that held unready returns before work began. We also standardized the preparation package and review checklist so managers received the same set of information in a consistent form.

But complete documents alone wouldn't catch the repeat issues in prepared returns. So we built an AI review assistant that checked the work against patterns in their firm's own prior review notes before a manager opened it. Their managers and signing CPA kept responsibility for professional review.

We also recorded effort by return type because more capacity wouldn't answer the partners' pricing question on its own. With that record, their pricing champion could put fees and effort side by side when making pricing decisions.

We moved the checks for missing documents and repeat preparation issues earlier, so managers could review better-prepared returns instead of sending the same work back.

A pilot check failed, so we changed the system and tested again.

The AI assistant had to help more returns clear first review without lowering quality. We agreed on that bar with their team and tested the assistant against the previous season's returns.

A pilot check failed, so we changed the system and ran the tests again. It took ~3 complete rounds of testing and changes before new kinds of failures stopped appearing and the pilot cleared the agreed bar. That gave us a tested system to take into their next peak season.

The new process had to fit the way their team worked.

We put the process into production for the next peak season and trained their preparers and reviewers. Where the readiness check took over, we retired the old email path for chasing documents so staff had one way to move that work forward.

Keeping the process useful would also take staff time. And so we proposed credit against billable targets for approved process-improvement work. The firm adopted that credit, allowing the team to count this work toward its targets. We watched daily use as they put the new process to work.

Less work came back, so more work could finish.

The follow-up 8 weeks after engagement completion showed how the workflow had changed from the diagnostic baseline. Among measured returns, the share sent back from first review fell from 40% to 11%. That meant 89% cleared the first pass.

Managers recovered time they'd spent on second passes, which had taken ~33% of their peak week. With less work returning to the same reviewers, each review took 20 to 30 minutes, down from 45 to 60 minutes before. Returns reached a manager in 1 to 2 days instead of 5 to 7 days.

The readiness checks also kept incomplete work from getting stuck partway through preparation: fewer than 5% of measured returns stalled there. Together, these changes helped their measured workflow finish 20% more returns each week. The team met the agreed quality bar, with 1 to 2 minor review notes per measured return in the follow-up.

Their leadership confirmed an observed $1.5M recurring profit increase. The partners had wanted growth to produce more income, and the result gave them evidence that the improvement reached profit as well as output. Their pricing champion also used the effort record in pricing decisions, so the work behind a fee no longer had to stay hidden.

Their team kept improving the process after handover.

The review assistant needed to change as their work changed, so we handed over documentation and adjustable checks. We trained their internal owners to change the tool's settings and add return types.

Those owners went on to adjust the AI review tool independently at least quarterly. Our handover gave them the skills to keep improving the system we built, while their firm kept responsibility for pricing and professional tax judgment.

Inside the readiness checks and review assistant.

See how we built and evaluated the tax-review workflow, including the source records, pilot quality bar, measurement definitions, and handover.

They came for more capacity. We removed the work consuming it.

More preparers wouldn't have explained why so many returns came back to the same managers, or whether the fees justified the effort. We measured both problems and built the checks and records their team needed to address them.

As a result, fewer returns went back, more finished each week, and their leaders could see the effort behind their pricing decisions. We gave their internal owners the skills to keep improving that process, and they did. The capacity they wanted came from removing work their team shouldn't have had to repeat.