AI applied to case law refers to the use of artificial intelligence to research, analyse and make use of court decisions. For legal professionals, the goal is not to replace legal reading, but to identify relevant rulings, case-law trends, courts' reasoning and decisions useful to a matter more quickly.
Silex applies legal artificial intelligence to the research, analysis and structuring of legal sources. The platform helps lawyers, jurists and in-house legal departments work faster, while placing source verification, legal method and professional obligations at the centre of the practice.
Why case law is difficult to research
Case-law research is rarely a simple keyword search. A lawyer or jurist must identify the relevant facts, characterise the legal question, find the applicable court decisions, distinguish landmark rulings from more circumstantial decisions, and check whether the solution remains current.
In Switzerland, the Federal Supreme Court publishes its landmark decisions in the Official Collection of Federal Supreme Court Rulings. Since 2007, all Federal Supreme Court decisions have generally been accessible online in anonymised form. Legal research can therefore cover thousands of decisions, with variations in language, area of law, date and type of proceedings.
To place this topic within the broader range of uses of legal AI, Silex offers an approach centred on sources, reasoning and verification.
What artificial intelligence brings to case-law research
Artificial intelligence tools can improve case-law research through natural language processing. Instead of formulating only exact keywords, the user can describe a situation in natural language: wrongful termination of a contract, a contested administrative decision, a disputed clause, social law, civil liability, procedure or a disciplinary sanction.
AI can then help turn this question into research angles, connect related concepts, extract important passages and propose a synthesis. Language models are particularly useful for moving from a large file to an initial map of the issues. But they are not enough on their own: the answer must always point back to verifiable decisions.
The legal analysis page explains how to use AI to structure facts, sources and reasoning, without confusing an automatic synthesis with a legal conclusion.
RAG, sources and court decisions: why verification remains essential
AI-assisted legal research often relies on an approach called Retrieval Augmented Generation, or RAG. The principle is simple: the system does not merely generate an answer from a model. It first searches for relevant sources, then relies on those sources to produce a contextualised answer.
In the legal field, this method is essential. A fluent but unsourced answer can be dangerous. A fabricated decision, a misquoted ruling, or foreign case law presented as applicable can undermine advice, a memo or an argument. Independent studies have also shown that general-purpose models are wrong in a majority of their answers to verifiable legal questions, whereas tools relying on prior research within structured sources significantly reduce this error rate. AI-assisted legal research must therefore be built around evidence: which decisions? which citations? what context? how current?
Swiss source: Federal Supreme Court decision databases.
Use cases for lawyers, jurists and legal professionals
- Case-law research
- Possible AI contribution: identify decisions close to a legal question.
- Professional oversight: read the rulings and verify their scope.
- Decision analysis
- Possible AI contribution: summarise the facts, procedure, reasoning and outcome.
- Professional oversight: check the reasoning and the references.
- Case-law trends
- Possible AI contribution: compare several decisions on the same topic.
- Professional oversight: verify the sample and the period studied.
- Preparing advice
- Possible AI contribution: structure the favourable and unfavourable arguments.
- Professional oversight: adapt to the mandate, the risk and the client.
For a law firm, these uses help reduce preparation time, particularly in matters where several decisions must be compared. The AI for lawyers page details the use cases specific to law firms.
Federal Supreme Court, cantonal courts and administrative decisions
Serious case-law research must define its scope: Federal Supreme Court, cantonal courts, administrative court, specialised administrative tribunals, published decisions, anonymised rulings or landmark case law. AI can help explore these bodies of case law, but it must not mix decisions from different levels without flagging it.
The Entscheidsuche platform allows searches of Swiss rulings, including those of the Federal Supreme Court and cantonal courts. It illustrates the importance of structured access to legal information. For a professional, the question is not only finding results, but knowing which ones are useful, current and applicable.
Swiss source: Entscheidsuche, Swiss decision search.
Generative AI and professional obligations
Generative artificial intelligence can automate preparatory tasks: summarising decisions, extracting passages, comparing lines of reasoning, proposing an outline or rephrasing a memo. But in the legal professions, this automation must be governed by professional obligations.
A lawyer cannot delegate their professional judgment to a model. They must verify sources, protect data, uphold professional secrecy and take responsibility for strategy. AI can be an efficiency accelerator, not a decision-maker. The Swiss Bar Association published guidelines in June 2024 on the use of artificial intelligence by lawyers, recalling that verifying results and upholding professional secrecy remain personal obligations of the mandated professional, whatever tool is used.
The Silex security page details the guarantees expected for professional use: Swiss hosting, confidentiality, zero training on client data and secure infrastructure.
How Silex transforms case-law research
Silex helps legal professionals formulate a question, identify relevant legal sources, analyse decisions and build a documented answer. The tool does not merely produce an answer: it provides the keys to understanding the reasoning and verifying the references.
For a lawyer, this can mean less time spent sifting through scattered results. For an in-house jurist, it can help prepare an internal memo or verify a position. For a litigation team, it can make it easier to compare several decisions on the same legal issue.
To discover the research, analysis and drafting features, see the Silex product page. Teams wanting to test Silex on their own case-law questions can book a demo.
Criteria for choosing an AI legal information solution
A legal information solution should not be assessed solely on the fluency of its answers. In law, quality is measured by the relevance of sources, traceability, data security and the professional's ability to retain control.
- Corpus: which decisions and legal sources are accessible?
- Traceability: can the cited rulings be opened, read and checked?
- Context: does the tool understand the area of law, the procedure and the question asked?
- Security: is the matter's data protected and excluded from training?
- Supervision: does the user retain final validation?
If you are comparing a general-purpose AI with a specialised legal platform, the ChatGPT vs Silex page highlights the differences in method, sources and security.
FAQ: AI and case law
Can AI find relevant case law?
Yes, provided it relies on reliable sources and the results are verifiable. It can help formulate a search, identify similar decisions and prepare a synthesis.
Can AI replace legal research carried out by a lawyer?
No. It can speed up case-law research, but the lawyer or jurist must read the decisions, verify their scope and adapt the analysis to the matter.
What is RAG in legal research?
Retrieval Augmented Generation involves first searching for sources, then generating an answer from those sources. In law, this method reduces the risk of unverifiable answers.
What risks should be monitored with AI and case law?
The main risks are fabricated decisions, incorrect citations, inapplicable sources, poorly protected confidential data and the absence of human supervision.
Is Silex suitable for case-law research?
Yes. Silex is designed for legal professionals who need legal research, verifiable sources, analysis and security within a Swiss environment.
