AML screening that catches more — and clears faster.

Every Quick AML Check screens the name against official sanctions lists, criminal watchlists and politically exposed person records in one run. It returns source-linked candidates and point-in-time evidence for your audit trail. The Quick Check is deterministic, so no search data is sent to an AI model.

AI reduces false positives in the AI review stage inside Full Search. For each candidate match it compares the available identifiers and produces a plain-language rationale plus explicit false-positive indicators — a mismatched date of birth, a different nationality, weak name overlap. Instead of a flat queue of hits, your team sees which candidates are worth escalating and which have clear reasons to clear. We publish the mechanism, not an accuracy percentage, because reviewers should judge the reasoning themselves.

It depends on the product you choose. The deterministic Quick Check and deterministic monitoring send no data to AI models at all. Full Search, Adverse Media Check and Deep Research use AI, and when they do, screening runs on EU-hosted models with zero data retention. Data at rest never leaves the EU. You control which level runs on each case, so you decide exactly when AI is involved.

Yes. Products are chosen per check, not per account. Many teams run the deterministic Quick Check across high volumes and mass monitoring, then escalate a specific name to the AI-Enhanced or AI Review level when a case is higher-risk or a hit needs a documented judgement. You pay per check at each level's rate, with no monthly platform fee, so mixing levels simply matches spend to the risk of each case.

Fuzzy name matching scores string similarity — the edit distance between two names. It helps with typos but struggles with transliteration, aliases and cross-script names, and it cannot explain itself. Inside Full Search, AI-Enhanced matching expands each query into real-world variants across scripts and alias forms before matching, and the AI match review adds reasoning about whether a candidate is genuinely the same entity. You get wider recall and a readable justification, not an unexplained similarity number.

The AI review inside Full Search lists each candidate match with a short written rationale and explicit false-positive indicators — for example, a birth year that does not line up or a different country of nationality. Each candidate links back to its source record and the name variant that triggered it. The reasoning and indicators are included in the point-in-time evidence PDF, so the decision is documented for your file and defensible to an auditor.