Strip the interface away from a legal case-management tool and what remains is a data problem. There are entities (people and organizations), documents, events with dates, deadlines with obligations, and sources that must be cited exactly. The hard part is not storing these. It is keeping every derived view consistent when one record changes.
ilovelawyer, developed by FORHU, is an AI workspace for lawyers built around that problem. This piece reads its public feature set through a data-modeling lens.
The core records
The workspace is organized per case. A case has parties and uploaded documents, in PDF or Word format. Around those, the product maintains:
Case: Parties, documents, status, next date
Timeline: Key dates drawn from the documents
Deadlines: Procedural dates tied to filings and obligations
Research results: Related case law with title, case number, source link, snippet and a relevance note
Analyses: Contradictions, strategy, findings, outlook, mind map
The research records carry the most metadata. Each answer carries the metadata a reader needs to reach the underlying source, and the case law documents are read-only, so the text stays as published.
Derived views and the sync problem
The analyses are derived data. They come from the documents, and they must change when the documents do. ilovelawyer refreshes them on demand, or automatically after documents are uploaded, deleted or archived. When the file changes, contradictions, strategy, key dates, findings, outlook and mind map are re-derived.
This matters for trust. A stale analysis can mislead because it looks authoritative. Tying every view to the current state of the file removes a class of silent errors. After each run, a "What changed" summary lists what moved, which doubles as a lightweight diff for reviewers.