Skip to content
RechtsKI

← Back to the blog

Analyzing Federal Court Decisions with AI: Rethinking BGE Research

Classic BGE research via keyword search is time-consuming and incomplete. AI-powered semantic search finds relevant Federal Court decisions based on legal context – not just keywords.

Orlando Kahanek ·

Analyzing Federal Court Decisions with AI: Rethinking BGE Research

Researching Federal Court decisions is a core skill of every Swiss legal professional. Whether for case preparation, brief writing, or client counseling – the right BGE citation can make the difference between winning and losing a case.

But traditional research methods are reaching their limits. The Federal Court database (bger.ch) contains over 160,000 decisions – growing by approximately 8,000 new judgments annually.

The Limits of Traditional BGE Research

Problem 1: Keyword Dependency

Full-text search on bger.ch works on exact string matching. Searching for «lease termination» finds decisions with exactly those words – but not decisions discussing «dissolution of the tenancy» or «ending the lease relationship».

Problem 2: Context Blindness

Keyword search doesn't understand legal context. Searching «damages employment law» yields thousands of hits – but which concern wrongful termination (Art. 336a CO) vs. workplace accidents (AIA) vs. continued salary payment (Art. 324a CO)?

Problem 3: Language Barrier

The Federal Court rules in all three official languages. A German-speaking lawyer systematically misses relevant decisions in French or Italian.

Problem 4: Time Investment

A thorough BGE research on a complex legal issue typically takes 2-4 hours.

How Semantic AI Search Transforms Research

Semantic search is based on vector representations (embeddings) of texts. Instead of searching for exact strings, the system understands the meaning of a query and finds texts with similar meaning.

Example: Query: «Can an employer reduce salary if an employee arrives late?»

Keyword search finds: Only decisions containing «reduce salary» and «arrive late»

Semantic search finds:

  • BGE 130 III 213 (Salary deduction for breach of contractual obligations)
  • Judgment 4A_30/2020 (Limits of contractual penalties in employment law)
  • BGE 137 III 303 (Proportionality of employment law sanctions)
  • Relevant decisions in French (ATF) and Italian (DTF)

Practical Example: Tenancy Law Research

Initial question: A landlord wants to increase rent by 15% after a complete renovation. The tenant disputes the proportionality.

Traditional research (estimated 3 hours)

AI-assisted research (estimated 5 minutes)

Time savings: From 3 hours to 5 minutes – with simultaneously more comprehensive coverage.

Best Practices

1. Precise questions: The more specific your question, the more relevant the results 2. Source verification: Always verify cited BGE numbers on bger.ch 3. Supplementary research: Use AI as a starting point, not as the sole source 4. Think multilingual: The Federal Court rules in all official languages

Conclusion

AI-powered BGE research is not a future vision – it is available today and offers significant advantages over traditional keyword search.

---

RechtsKI offers AI-powered BGE research based on the complete federal law corpus. Every answer contains verifiable BGE numbers with direct links to the original decision.

Tags

BGEBundesgerichtSemantische SucheRechtsrechercheRAGFallanalyse