In one sentence
"Extract the contracting parties, the term and the notice period from every document in the case file" — that's all it takes: SmartLegalPro reads each case document individually, merges the results and delivers one single, clean table with a source and reference column.
One sentence is enough
Until now the AI could pull facts from a single document — we introduced the case extract some time ago. What's new is the step across the whole file: in the lawyer chat you name the fields you need, and the AI works through each document in the case file one by one. No copying and pasting things together, no giant prompt with countless attachments — every document gets its own focused reading pass. That is precisely what keeps the results attributable.
With source references instead of gut feeling
Every row of the table names the source document and the passage the value was taken from. That fundamentally changes the nature of the check: instead of reading all the documents yourself again, you check specific samples — look at the row, open the passage in the document, tick it off. The extraction remains what it is: an AI assessment and preparatory work. The professional review stays with you — it simply becomes considerably faster, because every value comes with a traceable source.
Onward processing without a media break
The finished table doesn't get stuck in the chat. You export it as CSV into your spreadsheet, as DOCX into your pleading — or file it directly in the case file with a single click. No reformatting, no retyping, no screenshots of chat histories.
Where it pays off straight away
- Contract portfolio ahead of a transaction: parties, terms and notice periods across all contracts in one overview — as a starting point for the detailed review.
- Deadline inventory across legacy files: which notice and renewal periods are lurking in existing files? One table instead of many individual reviews.
- Claims overview for the statement of claim: amounts, dates and parties from the exhibits captured in a structured form — with the source reference that later goes into the citation.
Why the merging is code, not AI
One detail that matters to us: the AI reads each document individually — but the merging of the individual results into one table is done by deterministic code, not a language model. So there is no further AI step sitting between the documents that could mix up rows, blend values or invent entries. What appears in the table comes from exactly one reading pass over exactly one document — and is traceable back there via the source-reference column. That confines the risk of error to a place where you can check it by sampling.
Who it's for
Multi-document field extraction is part of the Anwaltsportal. It does not replace legal review — it makes sure that your review begins with a finished overview rather than with a stack of files.
