800 Years of Legislation, Made Searchable by Meaning, for a Government Department

09 deep archive
A generative AI virtual assistant reduced regulatory lookup time by more than 90%, matching plain-language compliance questions to the right legislation across more than 800 years of accumulated regulation and multiple industry segments.

Table of Contents

Challenge

A government department was responsible for regulation spanning more than 800 years of legislation across multiple industry segments. Keyword search against that volume of material returned 90 or more irrelevant results per query, and identifying which rules actually applied to a given situation required specialist knowledge most users simply didn’t have.

Solution

NexaQuanta built a generative AI virtual assistant that understands regulatory intent, matching queries to regulations by meaning rather than keyword overlap. Users describe their business context or compliance question in natural language, and the assistant identifies and presents only the applicable regulatory requirements.

The system runs on Python and React JS, deployed on AWS cloud, using generative AI to interpret intent rather than relying on keyword indexing alone.

Results

Time to identify applicable regulations: reduced by more than 90%.

  • Compliance accuracy and consistency improved across multiple industry segments, regardless of which team was asking.
  • Most regulatory lookups no longer require a specialist intermediary.

What this means going forward

This wasn’t a faster search bar over the same keyword index. It replaced keyword matching with an understanding of regulatory intent, which is why it could cut through 800 years of accumulated legislation instead of just searching it faster. Telling someone which rule actually applies to them, not just which documents mention their search terms, is what took lookup time down by more than 90%.

Ready to see what’s possible in your workflow?

Explore our Agentic Work Solutions →

Share the Post:
Scroll to Top