How a Mumbai Litigation Firm Reduced Contract Review Time by 60% with AI
· 6 min read · Case Study
A 12-lawyer firm was drowning in contract review work. Here's how they used AI-powered clause extraction and risk flagging to reclaim their time.
When Priya Mehta joined Kapoor & Associates as a senior partner in early 2025, she inherited a problem that had been festering for years. The 12-lawyer firm, based out of Lower Parel in Mumbai, handled a mix of commercial litigation and corporate advisory work. Their bread and butter was contract review - vendor agreements, service-level agreements, employment contracts, and the occasional M&A due diligence package.
The problem was volume. On any given week, the firm was reviewing between 40 and 60 contracts for their roster of mid-market clients. Each contract required a lawyer to read through every clause, flag potential risks, cross-reference with the client's standard terms, and produce a summary memo. A single vendor agreement could take 2-3 hours. A complex SLA might take an entire day.
"We were spending 70% of our billable hours on work that felt mechanical," Priya recalls. "Our associates were burning out, and we were turning away new clients because we simply didn't have the bandwidth."
The firm had tried hiring more associates, but the economics didn't work. They'd also experimented with outsourcing to contract review agencies, but the quality was inconsistent and the turnaround times were unpredictable. What they needed was a way to handle the initial pass - the clause extraction, the risk identification, the comparison against standard terms - without requiring a lawyer's full attention for every page.
That's when they started evaluating Lysa.
The implementation was straightforward. The firm uploaded their standard clause library - around 200 clauses covering indemnity, limitation of liability, termination, IP assignment, and confidentiality. They configured risk thresholds for each clause type: what constituted a deviation from their clients' preferred positions, what was a red flag, and what was acceptable.
Within the first week, the results were striking. Lysa processed a batch of 15 vendor agreements for one of their largest clients, a logistics company with operations across six states. The AI extracted every relevant clause, compared each against the client's standard terms, and produced a risk-flagged summary for each contract.
"What would have taken our team three full days was done in about four hours," says Rajesh Iyer, the associate who ran the pilot. "And it wasn't just faster - it was more thorough. The system caught three liability clauses that we'd missed in our manual review of similar contracts the previous quarter."
Those three missed clauses were significant. One was a uncapped indemnity provision buried in a sub-contractor agreement. Another was a change-of-control clause that would have allowed a vendor to terminate services if the client underwent a restructuring. The third was a data processing clause that conflicted with the client's privacy policy.
After the pilot, the firm rolled out Lysa across all their contract review work. The numbers after three months told the story clearly. Average review time per contract dropped from 2.5 hours to 55 minutes - a 63% reduction. The firm was able to take on 8 new retainer clients without adding headcount. Associate satisfaction scores improved measurably, as lawyers spent more time on strategic analysis and client advisory work rather than mechanical clause-by-clause reading.
The financial impact was equally compelling. The firm estimated they saved approximately 120 billable hours per month on contract review alone. At their blended rate, that translated to capacity worth ₹18 lakhs per month that could be redirected to higher-value work.
"The misconception about AI in legal is that it replaces lawyers," Priya notes. "That's not what happened here. Our lawyers are still reviewing every contract. They're still making the judgment calls. But now they're starting from a position of clarity rather than a blank page. The AI does the extraction and flagging; the lawyer does the thinking."
For firms considering a similar approach, Priya offers one piece of advice: start with your highest-volume, most standardized work. "Don't try to use AI for your most complex, bespoke matters first. Start where the patterns are clear and the volume is high. That's where the ROI is immediate and obvious."
Kapoor & Associates has since expanded their use of Lysa to include legal research and matter management. But it was contract review - that unglamorous, high-volume workhorse of legal practice - that proved the value of AI in their firm.