The question: can startups access relevant data sources?
At Legal Geek’s roundtable on data and AI, Jeroen Zweers asked seven experts about data challenges in legal tech. When it was Maurits’s turn: “As a startup, do you feel you can get access to relevant resources?”
The short answer: no.
SDU and Kluwer control licenses to serious legal publications in the Netherlands. That’s the market. If you’re building AI tools for Dutch law, you’re building with public court decisions and whatever else you can find.
What that forces you to do
When you don’t have comprehensive databases, you extract more value from what you have:
Visual extraction: We built complex AI prompting that pulls key information from court decisions and turns it into timelines and visualizations. Traditional legal publishers have more data. We had to make each decision more useful.
Social layer: Expert curation and user comments. A smaller, well-organized collection with context beats a comprehensive database no one can navigate.
Creative integration: Pull in legal blogs, academic papers, public sources. Add the context that makes individual decisions more useful.
User-centric design: Every piece of information has to count. Can’t rely on “we have everything” as the value proposition.
The toy analogy
Kids with too many toys. They don’t know what to do with them. Give them three toys and they get creative.
Limited data access works the same way. You’re forced to think about design and user experience instead of competing on volume.
What this means for legal AI
Three observations from the discussion:
-
Data moats aren’t everything. The publishers have the data. But having all the data doesn’t automatically make better tools.
-
Design matters more than expected. When users get value from 100 well-organized decisions instead of 10,000 poorly searchable ones, comprehensive loses.
-
Constraints create focus. We’ve gotten better at AI prompting and visualization because we had to. More data might have made us lazier.
The reality
Is limited data access a barrier? Yes.
Does it kill innovation? Apparently not.
Would we prefer open access to all legal data? Obviously.
But working within constraints has forced us to build tools that extract more insight from less data. Sometimes that leads to better products than having everything.
Maurits spoke at Legal Geek’s roundtable “Data and AI: Learning from each other’s journey” alongside experts from Nova, Rechtspraak, Kennedy Van der Laan, Pereira Tax, SDU, and Legalfly. The discussion was hosted by Jeroen Zweers (Noun) and Jelle van Veenen (Kennedy Van der Laan).
