Order.law

At a glance

Order.law is a legal research and case-management product for Indian lawyers. I started the first version in October 2024 with my co-founder Ankit Chaudhary, who owns the data and AI intelligence layer. I owned the product, design, research, front-end, API integration, marketing, SEO, and go-to-market packaging.

The strategic bet was simple: do not sell access to public legal data. Give away original judgments for free, use that layer to earn trust and distribution, then charge for the proprietary intelligence and workflow layer that law firms actually value.

• Product: AI legal research and case management for Indian lawyers
• Team: Two-person founding team
• My role: Product, design, research, front-end, SEO, marketing, packaging
• Business model: Free public judgments, paid AI usage and document storage
• Early validation: First paying law firm by May 2025; public judgment layer launched later as a distribution channel
• Distribution: 1.13M+ impressions, 24.6K clicks, and 28K users

This case study is less about isolated screens and more about the product decisions underneath them: market positioning, workflow frequency, pricing, packaging, and the mistake that changed the roadmap.

The market gap

Indian legal research is dominated by legacy tools such as SCC Online and Manupatra. The core contradiction is that lawyers often pay for access to judgments and legal material that are, in principle, public record.

But price was not the deepest problem. The deeper product problem was the shape of the tools. They worked well when a lawyer already knew the citation or case name. They worked poorly when the lawyer had a legal question and needed to discover the authority that answered it.

That means the most important mapping lived outside the software:

• from question to doctrine
• from doctrine to case law
• from case law to current legal status
• from a new matter to the right research path

The product assumed the lawyer already knew where to look. Younger lawyers, generalist lawyers, and lawyers entering adjacent practice areas had the question but not always the memorized citation. Experienced lawyers were also constrained, because the tools rewarded staying inside areas where their recall was already deep.

The opportunity was not a better search box. It was a product that could hold more of the legal context that existing tools left inside the lawyer’s memory.

The strategic bet

The first major decision was that Order.law would not monetize public judgments. We would make original judgment copies free and easy to access.

That decision came from a practical founder constraint. Early demos worked, but founder-led demos do not scale. Every hour spent selling was an hour not spent building. We were also sitting on a large database of legal material that very few people had seen.

The free-judgment layer became a distribution engine:

• it captured search-intent traffic from people already looking for legal material
• it built trust before asking for money
• it let the product demonstrate usefulness without me in the room
• it created a low-friction path from public legal information to paid workflow value

The first paying law firm came through direct founder-led selling by May 2025. The public free-judgment website launched later in October 2025 and became a scalable acquisition layer, reaching 1.13M+ impressions.

The important product principle was this: give away the commoditizable layer to win trust and distribution; charge for the proprietary layer that scales with customer value.

Field research

I spent time in lawyers’ offices and visited courts in Pune, PCMC, and Mumbai to understand how legal work actually moves. The product changed because the field reality was different from the first AI-research thesis.

Three observations mattered most.

Research is episodic

Lawyers do not do deep research every day. A senior lawyer may need deep research only a few times across the life of a case. That is a hard retention problem for a standalone research tool. The AI research experience could impress people in a demo, but it was not naturally a daily habit.

Legal work is distributed

Lawyers frequently represent each other across cities. A Mumbai lawyer may tie up with a Pune lawyer or firm for local appearances because much of the court process is paperwork and coordination. The product could not assume that each lawyer belonged to only one self-contained office.

Firms lacked operational visibility

I saw firms with 900+ active cases where the senior partner had no clear system-level view of which associate owned which case or what had happened recently. The state of the practice lived in one person’s memory.

The firms did not describe this as a software problem. They said they were busy. The product diagnosis was different: the firm could not see itself.

The real competitor was not only a legacy legal-research tool. It was WhatsApp, phone calls, shared logins, and memory.

The product shift

The key product move was to stop treating AI research and case management as separate products.

Research is high-value but episodic. Case management is daily, operational, visible, and easier for a firm to adopt. So case management became the wedge into the firm’s daily workflow, and AI research became the high-value moment inside that workflow.

That changed the product shape:

• matters and cases became the organizing object
• lawyers could assign cases to associates
• teams could log progress notes and attach documents
• senior partners could see firm-wide progress
• lawyers could belong to more than one firm
• AI research could use the existing matter context instead of asking the lawyer to re-explain everything

The dashboard was not just a dashboard. It gave the senior partner a way to see the firm without becoming the index for every update.

Key design decisions

Design around matter context

Legal research becomes more useful when it understands the matter it belongs to. The system needed to let a lawyer move from a case to a research question without re-entering the whole context.

Make the firm visible

The product had to show assignments, progress, documents, and status in one place. The job was not only to help one lawyer search faster. It was to help a firm operate with less dependency on memory and phone calls.

Support real legal networks

Because lawyers represent each other across cities and firms, the product needed to support a lawyer belonging to more than one firm. That came directly from field observation, not from a generic SaaS permission model.

Refuse seat-based pricing

Most software in the category priced by seat. I chose not to, because seat pricing punishes the exact behavior the product needs: getting the whole firm into the system.

In a law firm, a senior partner, associate, junior, and interns may all touch the same case. If every extra person increases the bill, the firm is incentivized to keep work outside the product or share one login. I had already seen that behavior in the market.

So seats are free and unlimited. The firm pays for the things that scale with actual value: document storage and AI usage.

The pricing model follows the same product logic as the free-judgment layer:

• free public judgments: win trust and distribution
• free internal seats: win adoption and data capture
• paid AI and storage: monetize the proprietary value layer

Results

• 1.13M+ search impressions for the free-judgment website
• 24.6K clicks from search
• 28K users reached through the public legal-information layer
• First paying law firm by May 2025
• Paid adoption from large firms, including firms managing 900+ active cases
• Consumption-based pricing validated with free seats and paid AI/storage

The result is not only a product interface. It is an acquisition loop, a workflow product, and a pricing model that came from how Indian law firms actually operate.

What I got wrong

I initially believed the AI research capability was the core product. The demos reinforced that belief because lawyers liked the insight and the experience felt impressive.

But demo delight is not the same as workflow demand.

A High Court lawyer explicitly asked us to build case management, and I discounted it because I was too attached to the AI thesis. I eventually reached the same conclusion through fieldwork, but slower than I should have.

The lesson was important:

• Distinguish delight from demand. The AI got applause; case management got adoption.
• Keep conviction falsifiable. A strong thesis helps you ship, but it should not filter out direct evidence from the right customer.

This is the part of the case study I would want a hiring team to notice. The strongest product decision was not being right from the start. It was updating the product when the market showed me the real workflow.

What this proves

Order.law is the clearest proof on this site that I can operate beyond traditional product design boundaries.

It shows that I can:

• read organizations, not only requirements
• turn field observations into product strategy
• connect UX decisions to distribution and pricing
• design around workflow frequency and adoption, not only screen quality
• work across product, design, front-end, research, SEO, and business model decisions
• be honest about the difference between a good idea and a product people will use

For a design leadership role, this is the work I would point to when a company needs someone who can move between strategy, product judgment, hands-on design, and execution.

---

Built with Ankit Chaudhary, who owns data and AI. Product, design, research, front-end, marketing, and SEO by me.