What 60% Faster Contract Review Actually Looks Like in Practice

Will Drewes

Founder, Fern Strategy · 3 min read

The numbers are real. The governance question is whether your firm is ready for them.

Most AI-in-legal coverage leads with capability demos. This post leads with two firms that shipped something to production and measured what happened.

The sample is small. Two cases, not twenty. But both are specific enough to be useful.

Lander & Rogers: 60%-plus time reduction in due diligence

Lander & Rogers, an Australian law firm, integrated AI-assisted contract review into their due diligence workflow. The outcome: more than 60% reduction in contract review time.

The system extracted clauses, flagged deviations from standard positions, and surfaced high-risk terms before a lawyer touched the document. Lawyers reviewed the flags, not the raw contracts. Human-in-the-loop validation stayed in place throughout.

That last part matters more than the headline number. A 60% time reduction that produces unreviewed AI output is a liability. A 60% reduction where lawyers are reviewing structured, flagged summaries is a different workflow entirely. Lander & Rogers kept the human judgment layer. The speed came from eliminating the mechanical read-through, not from removing attorney sign-off.

A UK M&A firm: 64% associate time reduction, three additional deals per year

A UK law firm running M&A due diligence on a £45M acquisition (250 contracts) ran AI across the full document set. Processing time: 18 hours. Manual baseline for the same work: 180 associate hours. That is a 64% reduction in associate time on that engagement.

The downstream effect: the same team completed three additional deals per year, and the firm calculated an annual capacity gain equivalent to four mid-level associates, roughly £680K in recovered capacity.

The deal timeline compressed from a six-week estimate to four weeks completed.

This firm is not named in the public case study, so treat the specific figures as directional rather than auditable. The Lander & Rogers outcome is the cleaner citation. But the M&A numbers illustrate something the time-reduction figure alone does not: capacity gains compound. Faster review per deal means more deals with the same headcount, which changes how you model the ROI entirely.

What this means for firm leadership

Three things worth noting before you scope a build or a buy.

First, the workflow change is the hard part. Both examples work because the AI output feeds a structured review process, not because the model is particularly impressive. Clause extraction and deviation flagging are only useful if your attorneys know what to do with the flags. That requires a defined playbook, not just a deployed tool.

Second, audit trails are not optional in this context. Contract review in M&A and due diligence sits inside regulated transaction processes. If your AI-assisted review produces a summary that a partner relies on, you need a record of what the model extracted, what it flagged, and what a human reviewed and approved. A black-box output with no log is a problem the first time a deal goes sideways and opposing counsel asks how the review was conducted.

Third, the build-vs-buy question is really a governance question. Off-the-shelf contract review tools can deliver the speed numbers above. The gap is usually in the audit layer: whether outputs are logged, whether human review steps are enforced in the workflow, and whether the system can produce a defensible record of what happened. Before you commit to a platform, map those requirements explicitly. The capability is largely commoditized. The governance architecture is where most firms are underinvested.

If you are scoping AI for contract review and want to pressure-test the governance layer before you build anything, that is exactly what a structured two-week audit is designed to surface.