The account comes from our engagement record and the founder's follow-up with the firm's chair. The financial results were confirmed through that conversation. Client workpapers, security reports, and the contract aren't published here.
The approved position was the unit the check used
A past deal can show what a partner accepted in one situation, but that alone doesn't make it the firm's standard. We drew candidate positions from previous mark-ups and closed deals, reviewed them with partners, and obtained approval before using them in the check.
The approved-position store linked each position to supporting material. We used the same approved positions to configure existing drafting tools and to build the additional check. That gave lawyers and tools a common basis for assessing drafts.
A flag needed a source a lawyer could inspect
The check ran before partner review through an add-in in the drafting workflow. It compared a draft with approved positions, retrieved supporting passages, and linked flags to the relevant prior material. Lawyers could inspect those sources before deciding what to change.
The implementation used the firm's cloud environment and existing document permissions to govern access to source material. The client retained responsibility for approving positions and legal work. The check could decline an assessment when its approved positions or supporting material were insufficient.
The application recorded checks and outcomes so the team could distinguish three questions: whether the check worked on the test set, whether eligible documents went through it, and whether the department's review work changed.
Technical acceptance and rollout acceptance were different tests
We tested against closed deals with a separate development set and a held-out set reserved for acceptance. Each test used source deals that predated the deal being assessed. Partners supplied the reference judgments.
| Technical measure | Acceptance requirement | Reported acceptance result |
|---|---|---|
| Recall | Flag at least 85% of the divergences from approved positions. | 89% |
| False flags | No more than 20% of flags are incorrect. | 16% |
| Whole-document abstention | Decline to assess no more than 10% of documents. | 7% |
These measures have different denominators. Recall concerns known divergences, false flags concern the flags produced, and abstention concerns documents. The 89% figure isn't a general accuracy score.
Passing that test showed how the check performed on the acceptance set. But rollout also depended on each group's approved positions and time to adopt the process, so we measured eligible documents routed through the check separately in production.
The rollout missed its agreed coverage requirement
Coverage meant eligible primary deal documents routed through the check before partner review. The agreed requirement was at least 70% during Launch week. The initial reading was 41%, and the reading after four free remedy weeks was 52%.
Two groups couldn't route all their eligible work against an approved standard because their position sets were incomplete. Missing credit-agreement position types also caused the check to decline documents, while approval and adoption work competed with billable targets. We'd set a shared coverage objective without securing those inputs or scheduling the compensation decision that could address the incentive conflict.
We added and tested the missing positions, helped a senior partner demonstrate the route, and worked with the firm to assign document volumes through it. Those changes raised coverage, but the final 52% still missed 70%, which is why we refunded the $25,000 Launch allocation under the engagement's historical terms. Training and handover proceeded during the remedy period. The later coverage result of 74% doesn't change the fact that we missed the original commitment.
The operating results measured review work
| Measure | Baseline | Later department result |
|---|---|---|
| First-pass acceptance of primary deal documents | 18% | 59% |
| Median active partner review per primary deal document | 75 minutes | 36 minutes |
| Associate revision work after review | 5.2 hours per week | 2.9 hours per week |
| Coverage of eligible documents | No baseline stated | 74% |
First-pass acceptance meant no partner override on a standard firm position. It didn't mean a document needed no edits of any kind. Review time measured active partner work, not elapsed time to deliver a completed deal.
The baseline and early follow-up used a one-week work study with the same participant group and instrument. The record also reports later department results. The account doesn't establish that every condition stayed unchanged throughout the later period, so the before/after results aren't presented as a controlled causal experiment.
The 75-to-36-minute change is a 52% reduction. The hours released are capacity. The financial results below retain their own measures and attribution.
The financial results answer different questions
| Financial measure | Reported result | Meaning |
|---|---|---|
| Annual tool costs before the check's operating costs | About $210,000 reduction | The firm reduced licensed seats and ended a pilot after comparing tools against the approved standard. |
| Annual tool costs after the check's operating costs | $120,000 reduction | This is net annual tool-cost savings. It doesn't deduct the one-time engagement fee. |
| Earlier full-year net income margin | 36% → 37.1% | The CFO attributed about 40% of this improvement to the drafting and review change. Rates and work mix also contributed. |
| Later net income margin | 38% run rate | The firm's rate of performance at the later follow-up. |
| Profit per equity partner | $111,000 increase, from $679,000 to $790,000 | The chair-confirmed result closed about 80% of the gap to the $820,000 goal. |
The CFO's 40% attribution applies to the earlier 1.1 percentage-point improvement. It isn't applied to the later run-rate change. Gross and net tool-cost savings aren't added together, and the firm's total profit improvement isn't attributed solely to this engagement.
Handover included a change the client shipped
We supplied the position-review process, evaluation procedure, operating instructions, and training for the client team. The knowledge team added and published a new position set through the evaluation process before handover without our help. It later extended the check to another practice group.
The firm could therefore keep improving the approved standard while retaining partner control over the underlying decisions. The measured outcome included better review work and lower tool costs, alongside a rollout commitment we missed and refunded.
We helped this law firm increase profit per equity partner by $111,000 by finding and removing a bottleneck in their drafting workflow.
If you’re a law firm leader and want to grow your firm without making partners carry more work, we’ll identify the biggest bottleneck standing in your way, build the plan to remove it, and implement it in eight weeks* or less - guaranteed or we continue to work for free and you don’t pay.
Find my bottleneck* Eight-week delivery, or twelve weeks when substantial data foundations are required, as agreed before work begins. If a covered stage fails its agreed benchmarks, we provide up to four additional weeks of remedy work at no charge, then waive or refund that stage’s fee if it still fails. Successfully delivered stages remain payable. Agreed client-caused delays pause the clock. Delivery guarantee terms.