How In-House Counsel at a Fintech Startup Uses AI to Review 200 Vendor Contracts/Quarter
· 7 min read · Case Study
A 2-person legal team at a Series B fintech replaced ₹15,000-per-contract outsourced reviews with AI-powered first-pass analysis, saving ₹25 lakhs annually while onboarding 50+ vendors per quarter.
PayGrid Technologies is a Series B fintech startup based in Bangalore, operating in the payments infrastructure space. They process transactions for over 400 merchants, integrate with 12 banking partners, and rely on a sprawling ecosystem of technology vendors, cloud service providers, data partners, and compliance service providers to keep their platform running. Every quarter, they onboard between 50 and 70 new vendors - each bringing their own contract.
The company's legal team consists of exactly two people: Ananya Reddy, General Counsel, and Farhan Sheikh, Legal Associate. Between them, they handle everything from regulatory filings with RBI to employment contracts to board governance. But the single largest consumer of their time, quarter after quarter, was vendor contract review.
"Every vendor has their own paper," Ananya explains. "Their own terms of service, their own data processing agreements, their own SLAs. And in fintech, you cannot just sign whatever lands on your desk. We handle sensitive financial data. We are regulated by RBI. A single problematic clause in a vendor agreement - say, an inadequate data breach notification provision or a unilateral termination right - could create regulatory exposure for the entire company."
Before adopting AI tools, PayGrid's approach to vendor contracts was a combination of internal review and outsourcing. Ananya would personally review contracts from strategic vendors - banking partners, core technology providers, and anyone handling sensitive data. For the remaining contracts - the long tail of smaller vendors, SaaS tools, and service providers - the company outsourced review to a law firm that charged ₹15,000 per contract.
The economics were painful. At 50 to 70 new vendors per quarter, with roughly 40 contracts outsourced for review, the firm was spending ₹6 lakhs per quarter - ₹24 lakhs per year - on external contract review alone. And the turnaround time was frustrating: the law firm typically took 5 to 7 business days per contract, which meant vendor onboarding was perpetually bottlenecked by legal review.
"Our product team would close a deal with a vendor on Monday, and the vendor would be ready to integrate by Wednesday," Farhan recalls. "But legal review would not be complete until the following week. We were the bottleneck, and everyone knew it."
The turning point came when Ananya mapped out exactly what the outsourced review actually involved. The law firm was not providing strategic advice on most of these contracts. They were performing a mechanical comparison: checking the vendor's terms against PayGrid's standard requirements (data protection provisions, liability caps, termination rights, IP ownership, audit rights) and flagging deviations. It was important work, but it was pattern-matching work - exactly the kind of task that AI excels at.
PayGrid implemented Lysa for contract review in October 2025. The setup involved configuring the system with PayGrid's standard contract requirements - a checklist of 35 provisions that every vendor contract needed to contain or address. These covered data protection (aligned with the Digital Personal Data Protection Act and RBI's cybersecurity framework), liability and indemnity, intellectual property, service levels, termination and exit, and audit rights.
The new workflow operates in three tiers. When a new vendor contract arrives, it goes through AI-powered first-pass review. The system reads the entire contract, identifies all relevant clauses, and compares each against PayGrid's standard requirements. It produces a deviation report: which requirements are met, which are partially met, which are missing entirely, and which contain provisions that conflict with PayGrid's standards.
Tier one contracts - those where the AI finds no significant deviations from standard requirements - are approved by Farhan after a 15-minute review of the AI's analysis. These represent approximately 30% of incoming contracts and are typically standard SaaS agreements from established vendors.
Tier two contracts - those with minor deviations that fall within pre-approved negotiation parameters - are handled by Farhan with AI-suggested redline language. The system not only identifies the deviation but suggests alternative language drawn from PayGrid's clause library. These represent about 50% of contracts and typically require 30 to 45 minutes of Farhan's time.
Tier three contracts - those with significant deviations, unusual provisions, or strategic importance - are escalated to Ananya for full review. These represent about 20% of contracts and receive the same level of attention they always did. The difference is that Ananya now starts with the AI's analysis rather than a blank reading, saving her approximately 40% of the time she previously spent on these reviews.
The results after two quarters were transformative. Average turnaround time for vendor contract review dropped from 5 to 7 business days to 1 to 2 business days. The outsourced review cost of ₹24 lakhs per year was eliminated entirely - replaced by the cost of the AI tool, which was a fraction of that amount. Net savings exceeded ₹25 lakhs annually.
But the benefits went beyond cost and speed. The quality of review actually improved. The AI consistently checked all 35 standard requirements against every contract - something that even diligent human reviewers occasionally missed when working through a batch. In the first quarter of using the system, the AI caught 14 contracts with inadequate data breach notification provisions (requiring notification within 72 hours rather than PayGrid's required 24 hours) and 8 contracts with unlimited liability exposure that the previous outsourced review process had approved without flagging.
"The AI does not have bad days," Ananya observes. "It does not rush through the last contract of the day because it is tired. It applies the same 35-point checklist with the same rigor to every single contract. That consistency is something we never achieved with outsourced review."
The vendor onboarding bottleneck dissolved. Product teams could now close vendor deals knowing that legal review would be complete within 48 hours rather than two weeks. This had a measurable impact on PayGrid's ability to move quickly - a critical competitive advantage in the fast-moving fintech space.
For Farhan, the change in his daily work was profound. "I went from being a contract-reading machine to being a legal analyst," he says. "I still review every contract, but I am reviewing the AI's analysis and making judgment calls rather than reading 40 pages of boilerplate. I have time now for the work that actually requires a lawyer's brain - negotiating complex terms, advising on regulatory questions, and supporting the business on strategic decisions."
Ananya's advice for other startup legal teams drowning in contract volume is straightforward: "Define your standards clearly - what you need in every contract and what you will not accept. Then let AI enforce those standards consistently across your entire vendor portfolio. Your time as in-house counsel is too valuable to spend on mechanical comparison work. Save it for the contracts and decisions that actually need human judgment."
PayGrid's legal team still consists of two people. But with AI handling the first-pass review of 200 contracts per quarter, those two people now operate with the throughput of a team three times their size - and with greater consistency than the outsourced alternative they replaced.