One engine. Every jurisdiction.
Paste one piece of content, pick 2–4 jurisdictions, and run it through the exact same detection and scoring pipeline. What changes below is only the regulatory data each jurisdiction supplies — citations, registration verification, and helpline routing — never the detection logic itself.
The scam-detection and scoring engine is completely separate from the regulatory data behind it. The engine never contains jurisdiction-specific rules — only the data layer changes: which firms are registered, which entities are on a warning list, which enforcement actions exist, and which helpline to route to.
Because of that separation, bringing a new country onboard is a data-integration task, not a rebuild of the detection logic:
- Connect that regulator's registry, warning list, and enforcement data.
- Register the new jurisdiction so it appears everywhere in the app.
- Add the local language, currency, and helpline details.
- Verify it against the same automated test suite every jurisdiction uses.
We keep a standalone proof of this: a brand-new, fictional regulator plugged into the exact same verification component every real jurisdiction above uses, with zero changes to the detection engine itself. Verifying a genuine registration and correctly rejecting an unregistered one both complete in well under a second.