We Built Evatt AI Because Brilliant Lawyers
Were Wasting Their Careers Being a Search Engine.
Built by a law student in a Munich dorm room. Tested in six jurisdictions. Used by 2,500+ practitioners. Still getting started.
In a Munich Dorm Room. At a Law School Exchange. With a Question That Changed Everything.
Ashley Edgar was a law student on exchange in Germany in 2023. It was late. He was watching a classmate — one of the sharpest legal minds in the room — spend four hours tracking down a case that probably existed, probably in the right jurisdiction, probably said what they thought it said.
He started building that night. Not because he had a product roadmap, but because he had a laptop, access to Python, and a very clear picture of the problem that needed solving.
“I didn’t build Evatt AI to replace lawyers, I built it to make their lives easier. I built it because lawyers I respect were spending their best hours hamstrung by outdated legal search tools. It seemed like such a waste of their intelligence — using tools from 15 years ago.”
The first version barely worked. But it answered a legal question faster than a human — and it cited its sources. Everything since then has been about making that answer more accurate, more jurisdictionally specific, and more trustworthy for the lawyers who stake their reputation on it.
"The question isn’t whether AI enters legal practice. It already has. The question is if your firm is making use of this new advantage."
Why Evatt AI?
Dr. Herbert Vere Evatt was one of Australia’s greatest legal minds — a High Court justice, a foreign minister, and a fierce advocate for the idea that legal expertise should not be the exclusive property of those who can afford to wait for it.
He believed in the democratisation of law before anyone used that word. He believed that justice delayed by access was justice denied.
We named the product after him because we believe the same thing. Every lawyer — whether at Kirkland and Ellis’ New York office or a sole practitioner in Cairns — deserves access to the same quality of legal research. That’s not idealism. It’s a product decision.
From a Munich Dorm Room to 2.5k+ Lawyers Across 6 Countries.
The First Prototype
Built in a dorm room during a law school exchange. First answered a legal question faster than a human. Cited its sources. The seed of everything.
Three Things. Not Ten Things That Sound Good.
The things we actually optimised for — not the ones that look good in a pitch deck.
Accuracy above fluency
Generic AI is optimised to sound convincing. Evatt is optimised to be right. We'd rather flag uncertainty than present confident nonsense. That's why three AI models cross-check every answer before you see it. Being wrong in law has consequences. We don't forget that.
Privilege is a constraint, not a feature
We didn't add attorney-client privilege protection after the product was built. We built the product around it. Zero data retention, isolated environments, and no model training aren't features we offer. They're constraints we designed to. The distinction matters legally and architecturally.
Solo practitioners matter as much as large firms
The lawyer in Perth with one client and one case deserves the same research quality as the partner at a 500-person firm. That's a product decision — it's why we have a free plan, a solo plan that costs less than one billable hour, and why we don't have a 10-seat minimum. Dr. Evatt would have agreed.
Try Evatt AI on Your Next Case.
Free. No credit card. 30 seconds to start. Cancel whenever you want.