Menu
Back to Case Studies

RegTech ·Confidential, EU enterprise compliance

Compliance opens an inbox of the few rule changes that touch the business, not the thousand that do not.

A large enterprise whose compliance team read every incoming EU regulatory update by hand, looking for the few that applied to their own entities and markets.

Where this fits

Best for compliance teams in regulated industries that must catch every relevant rule change without reading everything.

Every incoming regulation scored against the entities and markets it can touch. Delivered in four months.

  • Node and OpenAI on AWS
  • Keyword, industry and market matching
  • Legal text turned into action points

What we did

  • Ingested EU regulatory sources into one normalised queue
  • Matched each update against the client entities, industries and markets
  • Scored relevance so the inbox holds only what applies
  • Summarised dense legal text into points a compliance officer can act on

Result

The compliance team now opens a queue of rule changes already matched to the entities and markets they affect, each with a plain summary of what it requires. Reading and filtering that used to be done by hand arrives sorted, and relevant changes no longer sit unread in a feed nobody has time for.

Radar inbox: incoming EU regulations matched by keyword, industry and market against the organization's own entities. Monitored entity and market names redacted.

The hard part

In compliance the two kinds of error do not cost the same. Miss one rule that applies and there is a fine; surface ten that do not and the team stops trusting the tool and goes back to reading everything. So the matching cannot simply be tuned for accuracy, it has to lean deliberately toward over-inclusion with a review step, and every match has to show the reason it was made so a compliance officer can overrule it. A model that summarises a regulation confidently and wrongly is worse than no tool at all.

Stack
Node.jsOpenAIPostgreSQLAWSRetool
Focus
RegTechComplianceRegulatory IntelligenceAILegal

Facing a similar problem?

Tell us where it stalls today, and we'll tell you how we would approach it.