Adverse Media Screening
Googling your customer, at industrial scale — finding the fraud conviction on page twelve of the search results.
Adverse media screening (also called negative news screening) is the process of searching news articles, court records, regulatory notices, and other public sources for damaging information about a customer or counterparty — fraud allegations, money-laundering investigations, sanctions exposure, corruption cases, organized-crime links. It is a standard component of anti-money-laundering programs and customer due diligence: watchlists only contain people already designated, while adverse media surfaces risk signals years before a name reaches an official list.
The screening problem is fundamentally a document-understanding problem at scale. Systems must match names across languages, transliterations, and spelling variants without drowning analysts in false positives (the customer named Mohammed Khan is not every Mohammed Khan in the news); distinguish the subject of an article from a witness, victim, or lawyer mentioned in it; classify the risk category and severity of what's alleged; and deduplicate the same underlying story syndicated across fifty outlets. Modern platforms use entity resolution, natural language processing, and increasingly large language models to read the articles the way an analyst would — determining who did what, when, and how credible and serious the source is.
Screening isn't one-and-done: obligations increasingly require ongoing monitoring, re-screening the customer base as new stories publish. The operational challenge is alert quality — a system that flags everything is as useless as one that flags nothing — so programs tune match logic and risk taxonomies, route alerts by severity, and document analyst dispositions. Those documented decisions matter: when a regulator asks why an alert was dismissed, the audit trail of what was found, assessed, and concluded is the evidence that the control actually operates.
Follow the money — the regulatory regime that makes banks read mountains of documents to prove their customers' money is clean.
Prove who you are before the account opens — the document-heavy front door of regulated finance.
Checking every name against every list — where a missed match carries the heaviest possible consequence.
Proof Perimeter runs document AI inside your own perimeter — with a provenance record on every field.
Book a demo