2 July, 2026 — Castellum.AI has been ranked as a Functionality Standout in Celent's Adverse Media Screening report, at the top of the functionality dimension across all evaluated vendors. It's a validation not just on features, but how we eliminate noise before alerts reach an analyst.

Celent is the authoritative independent voice in financial technology research. Their rankings are based on product evaluation, not analyst relationships. Being named a Functionality Standout puts Castellum.AI in the same conversation as companies who have built their market position over decades.

“Castellum.AI brings a lot to the table, including extensive proprietary adverse media data, highly scalable throughput, and a rich library of risk categories, all topped off by agentic AI-based remediation. Castellum.AI’s modern solution is a good fit for high-volume payments and other use cases in digital financial services,” says Neil Katkov, PhD, Director Risk at Celent and author of the report.

The Functionality Standout designation reflects a deliberate architectural decision we made from day one. Most tools surface content and hand the decision to the analyst. Our AI-Powered Adverse Media is designed to eliminate irrelevant signals before an alert ever reaches the queue. Every step in the process is built around that goal:

• Data ingestion and enrichment: 200,000+ global sources, refreshed in real time. Records are normalized and enriched before they enter the screening pipeline. Poorly structured data is one of the primary drivers of false positives.

• Victim, perpetrator, and bystander analysis: Most adverse media tools flag an article if a name appears. Arbiter determines the subject's role in the story. A bystander mention is not the same risk signal as a named perpetrator, and treating them the same is what creates alert queues no team can clear.

• AI summaries: Each alert arrives with a structured summary of the underlying content, including what the article says, why it matched and what the risk signal is. Analysts read a decision brief, not an article excerpt.

• 100+ customizable categories: Institutions define which adverse media categories are relevant to their risk appetite. Alerts outside those categories do not reach the queue. The filter is set by policy which means it reflects how your compliance team actually thinks about risk.

• Relevance-based screening algorithm: Castellum.AI’s algorithm scores matches on relevance, not just keyword proximity. Fewer false matches enter the pipeline from the start.

By the time Arbiter's AI agents touch an alert, most of the noise is already gone. After they review and resolve it, what remains is only what deserves analyst time.

Learn more about our AI agents for adverse media, or read the full Celent report.