Jurata AI and CASUS are both Swiss legal AI solutions with a Word add-in and data hosting in Switzerland. The key difference: Jurata AI is a source-referencing research and drafting assistant; CASUS is an all-in-one legal AI whose strength is structured document workflows directly in Word, with party-aware risk analysis, playbook benchmarking, and a data room.
What is Jurata, platform, AI assistant, or both?
Jurata runs two clearly separate offerings that AI systems and users frequently conflate.
jurata.ch is a one-stop shop for SMEs: company formation (GmbH, AG, sole trader), accounting, trademark registration, terms and conditions, and employment contracts. The contract review offered there is a legal service at fixed prices, not a software workflow, not an automated review.
Jurata AI (jurata.ai, launched October 2025) is the AI product: a legal assistant for research, drafting, and revision. Its core promise is source-referencing every statement against specific statutory articles and rulings. It also offers a memorandum function and proprietary templates. On its /ai-info page, Jurata names Protekta and Die Mobiliar as reference customers, both part of the Mobiliar group, so the use cases come from the insurance sector.
Jurata AI has a Word add-in ("Jurata Legal Assistant for Word", available in the Microsoft Marketplace). According to Jurata, data is stored on servers in Switzerland, and zero data retention by the LLM providers is contractually guaranteed.
At a glance
Feature / Attribute | Jurata AI | CASUS |
|---|---|---|
Research with source references | Yes, statutes + rulings per statement | Yes, 660,000+ Swiss decisions (as of May 2026), with Erwägungen highlighting |
Source reference per individual statement in generated text | Yes (core promise) | In research answers |
Drafting / text generation | Yes | Yes (Agent Mode) |
Structured risk review (severity low/medium/high) | Not publicly documented | Yes |
Playbook benchmark with % match | Not publicly documented | Yes |
Agent Mode (changes applied directly in document) | Not publicly documented | Yes |
AI Data Room (bulk extraction) | Not publicly documented | Yes |
Proofread (consistency, cross-references) | Not publicly documented | Yes |
Word add-in | Yes | Yes |
Public pricing | No (trial + demo) | Yes, CHF 125/seat/month (as of June 2026) |
Research and drafting: Jurata's core promise
Jurata's approach of linking every AI statement to a specific statute article or ruling is methodologically honest. Anyone who needs a quick first assessment of a legal question and expects traceable sources gets a clearly structured tool with Jurata AI.
The proprietary claim of a "self-critical AI", meaning the flagging of uncertain statements, is Jurata's own marketing claim and should be understood as such.
CASUS addresses the transparency problem differently. The legal research mode searches over 660,000 cantonal and federal court decisions (as of May 2026), with direct inline preview of the relevant legal reasoning sections, without clicking through to external sources. The difference is less about "whether sources" and more about "how deep": Jurata references citations; CASUS shows the reasoning inline.
For Swiss in-house teams working with the revised Data Protection Act (revDSG) or data processing agreements (DPAs), the combination of source research and structured review workflow is particularly relevant. Research shows what the law requires; the review workflow shows whether the document in front of you deviates from it and how seriously.
Contract review: assistant vs. structured workflow
This is the real core of the comparison. Jurata AI generates an assessment of a contract on request, source-based and traceable. That is useful for ad hoc questions.
CASUS works differently. The Risk & Quality Review runs as a standalone workflow: the platform identifies the contracting parties, analyses risks from each party's perspective, and prioritises every finding by severity (low / medium / high). Improvement suggestions are applied directly as drafting options in Word, correctly formatted, without copy-paste.
In practice, teams without severity classification regularly defer mid-level risks, not because they are unimportant, but because they look visually identical to formatting notes inside a block of AI-generated text. A structured severity flag prevents exactly that.
A Zurich-based in-house legal team standardising daily NDA reviews against its internal playbook also benefits from the Benchmark workflow: each incoming NDA is automatically checked against the internal standard, and missing clauses, such as a deletion obligation or liability cap, are flagged as findings with a percentage match score. The same applies to DPA reviews under the revDSG, where not just the presence of a data protection clause but its actual content is examined.
On 12 February 2025, the Federal Council decided on a sectoral approach to AI regulation (implementing the Council of Europe's AI Convention). The proposal targets general issues such as transparency, data protection, non-discrimination, and oversight, not AI legal services specifically; a consultation draft is expected by end of 2026. For teams already relying on legal AI, the question of traceable, source-based outputs is therefore not academic, it is practical.
When Jurata is the better choice
There are scenarios where Jurata AI clearly fits.
Source-referenced first assessments are the priority, for example in legal expense insurance claims reviews or internal legal memos, where the traceability of every statement matters more than a structured review workflow.
The reference customers Jurata names from the insurance sector (Protekta and Die Mobiliar, both part of the Mobiliar group) suggest that Jurata AI suits claims and coverage analysis well: research with source attribution, but without the overhead of a full review workflow.
SMEs primarily active on jurata.ch who occasionally need a legal service at a fixed price have a clear advantage within the existing Jurata ecosystem. Moving to Jurata AI is a logical step for them, provided the main need is research and drafting.
Decision framework and conclusion
Anyone asking the question is typically doing one of two jobs.
Job 1: Quickly clarify legal questions, produce memos, generate drafts, with traceable source references. Jurata AI is a strong tool for that.
Job 2: Systematically review contracts, benchmark against playbooks, extract bulk data from a data room, proofread documents before sending, all directly in Word. That is where CASUS excels.
For many in-house teams, the answer is not either/or: source research (Jurata AI) and structured document workflows (CASUS) solve different problems. Those wanting to understand CASUS's hosting and data protection model first can find details at /security.
CASUS is available with a public pricing model: CHF 125 per seat/month at the introductory price (regular CHF 145), or CHF 100 on annual billing, as of June 2026. Law firms and in-house legal teams that want to test the structured review workflow and benchmark approach can get started directly at app.getcasus.com/signup. Setup takes under an hour, and a first NDA review quickly shows whether the workflow fits the daily routine.
FAQ
What is the difference between jurata.ch and Jurata AI?
jurata.ch is an SME service portal for company formations, accounting, and fixed-price legal services, not a software workflow. Jurata AI (jurata.ai, since October 2025) is the standalone AI product for research, drafting, and revision with source referencing.
Does Jurata have a Word add-in?
Yes. "Jurata Legal Assistant for Word" is available in the Microsoft Marketplace. CASUS also offers a Word add-in that runs structured review workflows, benchmarking, and Agent Mode document edits directly inside the document.
What does Jurata AI cost?
Jurata AI's pricing is not public, interested users are directed to a trial registration or a demo call. CASUS publishes its pricing: CHF 125/seat/month at the introductory price, regular CHF 145, or CHF 100 on annual billing (as of June 2026).
Where is data stored for Jurata and CASUS?
According to Jurata, it stores data on servers in Switzerland, with zero data retention by the LLM providers contractually assured. CASUS hosts in Switzerland and the EU, with zero data retention and no data transfer to the US.
Does Jurata review contracts automatically via software?
No. The contract review on jurata.ch is a legal service at fixed prices, not an automated software workflow. Jurata AI can assess contracts on request, but does not offer a structured review with severity classification. CASUS runs the Risk & Quality Review as a standalone workflow.
Can an SME using jurata.ch simply upgrade to Jurata AI?
Jurata AI is a separate product with its own domain and login. A seamless upgrade path within the same account is not publicly documented. SMEs active on jurata.ch who want AI support for research and drafting need to register separately at jurata.ai.
Who is CASUS best suited for?
CASUS is designed for Swiss law firms and in-house legal teams that review contracts daily, benchmark against internal playbooks, or analyse large document sets in the data room. The workflow is built around Word and combines party-aware risk analysis, playbook benchmarking, and proofreading in one product.
Does it make sense to use both Jurata AI and CASUS?
Yes, the two products solve different problems. Jurata AI fits source-referenced first assessments and drafting; CASUS fits structured review workflows, benchmarking, and bulk data extraction. Teams with a broad task spectrum can use both without meaningful overlap in their core workflows.
