What is sanctions screening?

Sanctions screening is the check of customers, counterparties and payments against official lists of people, companies, vessels and countries subject to financial sanctions, so a firm does not deal with or move money for them. It runs at onboarding, on every relevant payment and whenever lists change. A confirmed match has legal consequences, so a person always confirms it.

Why does sanctions screening matter once agents act?

Sanctions breaches are among the most serious failures a regulated firm can have, and many regimes treat them strictly, with little room for good intentions. That shapes what an agent may do. It can read a screening hit, compare identifiers, check the list entry and draft a view in seconds. It must not release a payment that a person has not cleared, and it must not decide a hit is a true match and act on it alone.

Screening is also time-pressured. A payment stopped by a possible match is waiting, and the customer usually is too. Agents shorten that wait by preparing the analyst's decision, which is where whether an AI agent can do sanctions screening gets its practical answer.

How does screening work in practice?

A screening tool compares names and other details against the lists using fuzzy matching, so spelling variations and transliterations are caught. Possible matches become alerts. An analyst then decides whether each is the listed party, using identifiers such as date of birth, nationality, address, registration number or vessel identifier.

Agents fit around that tool rather than replacing it. Our guides cover connecting an AI agent to your sanctions screening tool and triaging sanctions screening alerts with an AI agent, so the most urgent payment hits reach a person first.

An example from a payment queue

Take a business bank's outbound payment to a shipping agent. The screening tool stops it: the beneficiary's name is similar to a company on a sanctions list. The agent pulls the list entry, which gives a registration number and a country, and compares them with the beneficiary details in the payment and the customer's past payments to the same firm.

The registration numbers differ and the countries differ, but the payment references a vessel. The agent checks the vessel identifier against the list and finds no match, then drafts a recommendation to release with each identifier compared. An analyst reviews it and releases the payment. If the vessel had matched, the draft would have gone to a senior analyst with a recommendation to hold.

How is sanctions screening different from payee name verification?

Both compare names, so they are easy to confuse. Payee name verification checks that the name on a payment matches the name on the receiving account, to prevent misdirected payments and fraud. In the EU, the Instant Payments Regulation requires payment service providers to offer it. It protects the payer.

Sanctions screening checks whether the parties are legally prohibited, whatever the account says. A payment can pass payee verification perfectly and still be to a sanctioned party. The two checks need different sources, different agents and different owners.

Which terms sit next to sanctions screening?

• PEP screening: a status check that raises risk but does not prohibit the relationship.
• Asset freeze: the obligation to hold funds of a designated party, which follows a confirmed match.
• False positive: a possible match that is a different party; most sanctions alerts are these.

Where Pryme fits

In Pryme, the lists and the screening tool's results reach the agent as sealed or signed sources, while a beneficiary name typed into a payment is declared. The agent drafts a disposition at depth 2; releasing a stopped payment is held for a named analyst, the posture we take for all second-line compliance agents.