Legal research systems are getting better at finding relevant material. In cross-border work, however, difficulties can arise even before retrieval begins: a researcher may be looking for a concept the foreign legal system organises differently.

Foreign-law research often begins with translation. A lawyer identifies the concept that matters in the home jurisdiction, finds an apparently corresponding term in the target language and searches legislation, cases and commentary using that term.

This is sensible and frequently works. It can also build an initial assumption into the research process before any source has been retrieved: that the foreign system contains the same concept and organises the issue in broadly the same way.

Comparative lawyers are rightly wary of that assumption. Legal systems divide legal reality differently. A broad concept in one jurisdiction may correspond to several narrower concepts elsewhere. A doctrine that sits within contract law in one system may be addressed through a different doctrinal route in another. Sometimes there is a close equivalent. Sometimes the useful answer is a cluster of concepts rather than one translated label.

For legal AI research products, this creates an interesting data problem. Retrieval can only function against the underlying query, expansion and indexing strategy. Modern systems can reformulate queries and infer synonyms impressively, but foreign-law research may require conceptual expansion more than linguistic expansion.

Imagine a lawyer researching a US contract-law concept in Swedish law. The most useful Swedish search path may not be the term that most literally translates the English expression. It may involve several local concepts that divide the relevant legal issue differently. A system with access to explicit relationships between those concepts can use them to broaden or redirect the research.

This does not mean building a universal ontology of law before research can work. The practical question is more modest: “Which conceptual relationships are useful enough to make explicit?” For important legal concepts, a structured record can identify local terminology, related concepts, scope, purpose, legal effects and relevant distinctions. Comparative mappings can then indicate which concepts elsewhere are close, partial or structurally different.

This information can complement source retrieval. The sources remain authoritative; the structured layer helps the system identify what to look for and how to interpret the relationship between the research question and the foreign legal system.

The distinction is important because legal research quality is often discussed in terms of source coverage, citation accuracy and retrieval performance. All are essential. Cross-border research adds another dimension: conceptual coverage. A database can contain every relevant source and still be difficult to search if the researcher begins with the wrong conceptual map.

This is also where legal data strategy becomes inseparable from product strategy. Supporting another jurisdiction is not only a question of acquiring more documents. It involves deciding how users' questions will be connected to the way the jurisdiction actually structures its law.

The best research system may therefore be the one that first helps the user work out what the right foreign-law question is.