Get on top of a case file before you have read it
Point Whisperit at a matter and it reads what is in it, asks what you actually need, then produces a structured analysis you can correct rather than a summary you have to redo.
The reading is not the hard part. The volume is
The file arrives all at once
A transferred matter, an opposing party's production, a due diligence data room. The work is not difficult, it is simply more than a first pass can absorb before you have to advise.
Half of it is scanned
Correspondence photographed on a phone, decisions filed as images, faxes from another decade. None of it is searchable and none of it can be skimmed.
The first pass is never reusable
You read everything to build a mental picture, then the picture stays in your head. The next person on the matter starts from zero.
Three phases, and it stops to ask
It surveys the file
Whisperit lists what is in the matter, reads the documents that carry the most weight, and comes back with the parties it found, a rough sequence of events and how much material it is dealing with.
It asks who you are in this
Whom you act for, in what capacity, what you need out of the analysis and in which language. It stops here and waits, because an analysis written for the wrong side is worse than no analysis.
It writes the analysis
The result is a document in the matter, structured section by section, that you edit like any other draft. Nothing is filed or sent anywhere without you.
A document, not a chat reply
- 01
Case overview
What the matter is about, in the terms a colleague would use to brief you in a corridor.
- 02
Key documents
The pieces that actually carry the case, with a reference to each so you can open it.
- 03
Parties and positions
Who is involved and what each of them is asking for, as it appears in the file.
- 04
Timeline of events
The sequence the documents support, so the chronology is written once instead of reconstructed by each person.
- 05
Legal and factual issues
What is actually in dispute, separated from what merely takes up pages.
- 06
Risk assessment
Where the exposure sits, framed as questions to test rather than conclusions to accept.
- 07
Missing information
What the file does not contain and would need. Often the most useful section.
- 08
Next steps
A starting list of what to do, which you cut down to what you actually intend.
Reading the file is the engineering
Scanned pages are handled
Every page is classified before anything else: pages that already carry text are read directly, and only genuine scans go through optical recognition. You get the same result either way.
Office files keep their structure
Word, Excel and PowerPoint are read as documents rather than flattened to images, and pictures embedded in them are still recognised.
Correspondence counts as evidence
Emails in the matter are parsed and analysed alongside the documents, because the file is rarely only the formal pieces.
Search goes to the passage
The matter is indexed down to the paragraph, so a question about a clause returns the clause and not the whole contract.
Everything stays in the matter
The analysis, the sources it used and the document it produced all sit in the same case, available to whoever you have given access.
Swiss by default
Your workspace runs on a Swiss region by default, with EU and global tiers available. No AI provider we use trains on your data.
What lawyers ask about it
Is this a summary or an analysis?
An analysis. A summary shortens what is there; this identifies the parties and their positions, orders the events, separates the issues from the noise, and says what the file is missing. It is written to be corrected by you, not accepted.
Does it work on scanned documents?
Yes. Each page is classified first, so pages that already contain text are read directly and only genuine scans are sent through optical recognition. Photographed correspondence and image-only decisions are analysed like anything else.
Can I trust the chronology it produces?
Treat it as a first draft that saves you the assembly, not as a verified record. Each element carries a reference back to the document it came from, which is exactly what you need to check it quickly. It is generated from the file, so a date the file gets wrong is a date the timeline gets wrong.
How long does it take on a large matter?
It depends on how much of the file has to be read optically rather than directly, since that is the slow part. The survey phase comes back quickly; the written analysis follows once you have answered the clarifying questions.
Does it change anything in my file?
It creates the analysis as a document in the matter, and that is all. It does not file, rename, move or send anything without an approval step you control.
Try it on the file you are avoiding
The matter you have been putting off opening is the one that shows you fastest whether this is useful.