Why we built an AI agent for legal pricing

At LegalGeek this year, one theme dominated: everyone is building AI agents. Over coffee with friends both inside and outside legaltech, the same question kept surfacing - does everything need to be wrapped into an AI agent?

We think the answer is yes, as far as legal pricing is concerned. In fact, it’s essential.

1. Traditional pricing tools are bought, not used

Law firms have spent years buying pricing software that never really caught on with lawyers. The reason isn’t lack of value — it’s usability. Most systems were designed for pricing teams: complex, data-driven, and too rigid for how lawyers actually work.

Lawyers don’t want to fill in forms or reconcile data fields; they want to describe the work and get a credible answer. So they revert to emails, spreadsheets, and instinct — tools that feel fast and natural but leave the firm flying blind on commercial performance.

The result is a persistent adoption gap: the people closest to the client and the deal rarely touch the pricing tools meant to support them.

2. Lawyers now expect conversational, intuitive software

ChatGPT has changed expectations. Lawyers have grown comfortable explaining complex work in plain English and getting useful, structured output. The same pattern is appearing across the new generation of legal AI tools — Harvey, Lega, Luminance — all making powerful systems accessible through natural dialogue.

Our pricing agent brings that experience to the commercial side of law. Instead of clicking through templates, a lawyer can say:

“Give me a quote for a mid-market M&A deal — £50m value, UK target, buyside, bilateral.”

The agent handles the rest.

3. Under the hood, the agent is doing serious work

The experience feels simple, but the system is doing several complex things in sequence:

• Interpreting the matter description (whether from a prompt, an RFP, or a termsheet) and identifying likely scope and structure.
• Finding relevant precedent matters to use as benchmarks for time-spent assumptions (leveraging ayora's data enrichment engine).
• Building a data-driven quote using the ayora quote builder, any firm templates, and precedent insights.

This orchestration - the ability to act across multiple tools on behalf of the user - is what makes it a valued companion, not just a chatbot. It creates the seamless, “magical” experience lawyers associate with great consumer AI, but with the rigour and compliance demanded by the use case (pricing) and the context (enterprise law firms).

4. It also advises, not just executes

Because the agent understands the data behind each quote, it can surface commercial insight in context:

“We’ve found ten similar precedents - you can be more confident fixing this fee.”

“Few precedents here - be careful with assuming too much risk.”

This guidance matters. Pricing isn’t just about maths; it’s about judgement under uncertainty. By showing where the data supports a position, the agent helps lawyers make smarter commercial calls.

5. Adoption is the real ROI

Pricing products only deliver return if lawyers actually use them. Adoption has always been the hardest part - and it’s where AI agents change the game.

By meeting lawyers where they already are - in a conversational interface that feels familiar, fluid, and fast - we finally make pricing technology part of everyday work, not an optional extra.

Law firms don’t need another spreadsheet. They need a capable counterpart that understands their data, their deals, and their commercial objectives, and helps them quote with confidence.

That’s why we built an AI agent for legal pricing.