← Back to all case studies

NyayaNidhi — Scaling Legal-Tech Operations Across India

One court to five states — without quality falling apart.

Operations Agile AI/ML Process Design Stakeholder Mgmt Team Building Sprint Planning SOP Development
DurationMarch 2025 – February 2026 (11 months)
RoleAssociate Project Manager (Founder's Office)
TeamStarted solo with 2 co-founders → grew to team of 9
IndustryLegal-Tech / AI
ToolsNotion, Jira, Power BI, LeadSquared, Google Workspace

Starting point: one court, zero process

NyayaNidhi was a 3-person AI legal-tech startup operating in one High Court in Bangalore. There were no standard processes, no SOPs, no sprint structure. The AI translation engine existed but had a 90% accuracy rate — not good enough for legal documents where a single error can change a court ruling. Revenue was $6K/month. I joined as employee #3.

$6K
monthly revenue
1 court, 0 SOPs, 90% accuracy

What it actually took

The first thing I did was build a hub-and-spoke operating model. The hub handled quality control, training, and AI oversight. Each spoke was a state-level delivery team — translators, proofreaders, and a local coordinator who understood regional court formats and legal terminology.

I hired and trained 5 proofreaders, built onboarding SOPs, created error taxonomies for legal translations, and set up a QC process that sampled 10% of pages (100% for new translators). This pushed AI-assisted accuracy from 90% to 98%.

Simultaneously, I was doing court visits — physically going to High Courts in Maharashtra, Gujarat, and other states to pitch the translation and drafting services to lawyers and judges. Each state had different formats, different legal conventions, different stakeholder expectations.

I redesigned the AI architecture from a single-agent system to a multi-agent pipeline — separating translation, proofreading, formatting, and QC into distinct stages. This enabled 90% automation while maintaining the accuracy standards.

I launched a new drafting service across 10 district courts with sub-3-hour turnaround. Built sprint boards in Notion, ran weekly ops reviews, set up KPI dashboards in Power BI, and managed the DC/AC government pilot that scaled from 30 to 100 cases/month.

I also supported the $2M seed round from 3one4 Capital — building the operational traction slides, maintaining MIS analysis, and demonstrating scalable unit economics to investors.

And through all of this, I was building the BFSI pipeline — identifying 20+ potential enterprise targets for legal document processing in banking and financial services.


What we built

Revenue
$6K/month
$24K/month
Coverage
1 court
5 states
AI Accuracy
90%
98%
Govt Tender
0
₹24L
Team Size
3 people
Team of 9
Seed Round
$0
$2M
"Give me chaos and I'll leave a working system."