The case file that thinks along.
Instead of a PDF pile and RAG search, Smart Legal Pro builds your case as a persistent, sourced relationship graph — with chronology, evidence matrix, gaps and contradictions. Multilingual, jurisdiction-aware, configurable per case.
A living knowledge wiki of your case
Every document is distilled into facts, relationships and sources — not thrown into a vector index. The result is a traceable, correctable case wiki: you see where each statement comes from and can confirm, edit or discard it.
Who connects to what
Parties, claims, evidence and issues are linked as nodes and sourced edges. Supporting and contradicting relationships are immediately visible.
- The AI proposes edges — you confirm, edit or discard.
- Every edge carries a source (quote + document location).
- Contradictions are flagged, not swallowed.
The sequence, sorted automatically
The agent pulls dated events from the full text and arranges them into a timeline — with a source for each event.
- Events with a source reference (exhibit, page).
- Linked to parties and evidence in the graph.
- 14 MarTermination receivedExhibit C1, p. 2
- 21 MarObjection filedLetter C2
- 4 AprEnd of the 3-week period (§ 4 KSchG)Deadline calculation
What is proven — what is missing
Each claim is mapped to supporting and contradicting evidence. Evidence gaps are surfaced, not glossed over.
- Status per claim: supported, contested, gap.
- Gaps become concrete follow-up suggestions.
- Receipt of the termination on 14 MarAcknowledgment of receipt (Exhibit C1)
- Verbal severance promiseNo evidence — ask the client
- Deadline metOpponent disputes the receipt date
Three mentions, one person
A witness appears as “Max Mustermann” in one document, as “M. Mustermann” in another, and as “Mr. Mustermann” in the transcript. The platform recognizes such variants and suggests merging them — but only you confirm it.
- Three independent signals (typo similarity, name components, containment) surface plausible candidates; shared neighbors in the relationship graph additionally raise their priority — never an automatic merge.
- You actively choose which spelling wins going forward; the discarded mention stays recoverable via undo.
- Two similar names can also be two different people — father and son, a company and its identically named managing director. That is why a human decides, never an automatism.
Agents in this workflow
Builds the relationship graph via map-reduce across the whole case — facts, relationships, case state with gaps and contradictions.
Ingests new documents, detects corrections (changed values) and updates the graph incrementally.
Proactive AI suggestions: name missing documents, next steps and client follow-ups.
Extracts dated events from the full text and links them to the parties.
Maps each claim to supporting and contradicting evidence and finds gaps.
Checks for missing information per case type and jurisdiction — optionally requesting it from the client.
Control & safety
- All AI outputs are suggestions — you confirm, edit or discard.
- Every statement is sourced (quote + reference) and traceable.
- Reversible actions with an action ledger and audit log.
- EU hosting, client-separated (RLS); pseudonymization as a firm setting with a named scope.
Let your case files think along.
The AI lawyer portal — EU-hosted, profession-compliant, client-separated. You keep control, the AI does the groundwork.
