The problem
A large client needed visibility into its litigation portfolio: around 1,900 cases spread across courts in several states, with no consolidated database, no risk overview, and no reliable numbers for financial provisioning. Building that picture by hand — locating every case, downloading filings, reading, summarizing, and extracting figures — would take a dedicated team months.
What we built
In a fixed-scope project (TaaS model), we built a pipeline that first mapped and consolidated the entire portfolio into a single database, then went deep on the priority cases: it downloaded the filings directly from the courts — including systems that require authenticated access — and used AI to summarize each document in structured form: parties, rulings, deadlines, amounts, and risk.
The per-document summaries were consolidated into a per-case view (stage, decision history, overall risk) and into a provisioning spreadsheet with dozens of legal-financial columns, item by item. Everything was made available to the team through an internal lookup dashboard.
The outcome
In about 5 days, the full portfolio was mapped and consolidated, and the first deep-analysis phase covered around 120 cases — over 6,000 court documents downloaded and summarized by AI, from 8 different courts. Done by hand, this would have taken an entire team months — with the added advantage of being repeatable: extending coverage to the rest of the portfolio means rerunning the pipeline, not starting over.
The system remains available and under maintenance, ready to expand the analysis to the remaining cases.