Every litigator has lived it, the nightmare of “fact chaos” when discovery dumps, endless transcripts, buried emails, clinical records, and reports scattered everywhere become a spreadsheet grind. Hours (or days) wasted hunting timelines, cross-checking details, and praying you haven’t missed the one fact that swings the case.
Australian legal tech startup Mary Technology just raised A$7 million to kill that problem for good.
Their Fact Management System is built specifically for litigation and disputes. Upload the mess — thousands of docs, PDFs, chats, whatever — and the AI instantly extracts every key fact, tags themes and context, builds searchable chronologies, and creates matter summaries.
Ask in plain English (“What did the defendant admit on 12 March?”), get clear answers with one-click citations straight to the source document and page. Relevance reasoning included. Full audit trail. Duplicate detection. Ongoing updates as new material lands.
The result is verifiable facts instead of hallucination-prone ChatGPT guesses, which is exactly the safeguard courts and clients now demand.
Lawyers report 50-90% time savings on document review. One associate called it “a game-changer for massive clinical record loads”; another said it “eliminated a huge problem we had.”
Traction is real too with reported 10× ARR growth in 2025, 2,000+ lawyers already using it across 100+ Australian firms including Hall & Wilcox, Shine Lawyers, and Maurice Blackburn.
Fresh funding (led by OIF Ventures, with Sydney Angels and Empress Capital) is fuelling the global push. San Francisco office opening now, plus a new self-serve model rolling out in the US so smaller and mid-sized firms can jump in without enterprise sales cycles.
As CEO Daniel Lord-Doyle (co-founder alongside lawyer Rowan McNamee) puts it: “Facts determine whether a legal case is won or lost, yet the way they’re managed hasn’t materially changed in decades.”
Investor Oliver Darwin nailed it: “Fact chaos is a genuine hair-on-fire problem that’s gone unsolved for far too long.”
