After months of heavy computation across our databases, based on 307 data pipelines, APIs, and direct court correspondence, we turned Swedish law from a pile of PDFs into something entirely new: a living network of connected information.
We have computationally mapped 276 million connections that link statutes, recitals, case law, and guidance at chapter, paragraph, and even word level. The network goes far beyond direct citations: we use LLMs to find analogies across doctrines, recurring legal mechanisms, emerging trends in lower-court cases and more.
Until now, law has been structured sequentially – page 1 to 500, doc after doc. Slow, costly, and blind to connections. That's no longer the case.
Qura’s connected data is a new way to access law. Select any paragraph in any law and gain access to all of its connections, giving you instant and complete oversight of a legal matter, from the EU directive all the way down to local lower-court argumentation.
This makes law ten times more accessible and easier to navigate.
Why is this groundbreaking for law?
The law is a maze. Hiring guides through that maze (lawyers) is a trillion-dollar industry. That is how complex it is.
Take one simple rule: “Issuers shall publicly disclose inside information as soon as possible.” Straightforward? Not at all. To even know what shapes this rule, you must be aware of:
- 40 MAR articles + 80 recitals with definitions, exceptions, carve-outs
- CJEU case law unpacking “precise information” and “significant effect”
- 27 national supervisors issuing their own recommendations, reports, sanctions
- National case law showing a sliver of practical application
- One sneaky sentence on page 34 in an ESMA guideline on mechanics of delay
The law is not just text and standalone facts in a gigantic pile of legal documents (Sweden alone has +10 million in active rule). It is a network of connected data points – supporting, contradicting, and reshaping each other depending on context. Change one specific circumstance and the applicable rule can flip entirely.
Now, you can click on that MAR sentence in Qura and instantly see all connections: from direct citations to vague references buried in ESMA guideline page 34 – all with clear summaries and smart categorization that replace dozens of hours of manual research work.
This is not only a game changer for lawyers using Qura; for the first time AI models don’t need infinite context windows to make sense of law – they can jump between the connections too. It is our biggest breakthrough in 1.5 years of building AI for legal reasoning. For the first time, we are seeing AI that can tackle complex legal problems. More on this soon.
