Review a screening hit

When FlowSentry screens your customers — and their directors and beneficial owners — against sanctions, PEP and internal lists, possible matches arrive as hits. This tutorial walks through reviewing one hit, from the queue to a decision, including the second person's confirmation that FlowSentry requires for true matches and higher-risk clears.

Before you start

• Reviewer (maker): Analyst, Senior analyst, MLRO or Deputy MLRO role.
• Checker: Senior analyst, MLRO, Deputy MLRO or Screening checker role — a different person — with a second factor for step-up.
• Your institution's four-eyes rules are on Screening › Four-eyes policy (/screening/policy).

Step 1: Choose a hit from the queue

1. Open Screening › Hits (/screening). The figures at the top show Open hits, hits Awaiting a checker and Statutory list hits.
2. Keep the default order, Risk (TFS first): hits on statutory lists — the UN Security Council list and Kenya's domestic designations — come first, then by score. A statutory hit open for more than a day shows as overdue.
3. Filter by category, list or status, or search by subject, entry or list. Tick Assigned to me to see only your own.
4. Open a hit and select Assign to me so colleagues know you are working it.

Warning: Targeted financial sanctions obligations apply without delay. Work statutory list hits first, every day.

Step 2: Compare the subject with the list entry

The hit page (/screening/hits/{id}) puts the two side by side:

• Side by side shows the subject as screened beside the list entry as it was listed in the version that raised the hit. The words that matched are highlighted.
• The Reading column shows, for each attribute, whether it agrees, is close or differs: date of birth (exact date, year or year range), nationality or country, identifier, and type (individual or organisation).
• Why this hit explains the match in plain English, for example: name 92% on a token match; date of birth differs by 12 years — likely false positive.
• Score breakdown shows how the score was built: token alignment, full-name similarity, phonetic similarity, and the threshold it crossed.
• Related parties, Other hits on the same subject and the Decision history by named user complete the picture.

For a second opinion, select Explain this match. The match explainer compares the attributes (computed by FlowSentry, not guessed) and suggests a decision and narrative, labelled AI draft — advisory. Use narrative fills only the narrative; the decision stays yours.

Note: The explanation never changes the engine's score, the threshold or whether a hit is raised. It helps you read the evidence faster.

Step 3: Gather context if needed

Open the customer's profile for identity documents, date of birth, nationality and addresses. A different date of birth or nationality, or a company versus an individual, often resolves a hit quickly. If you need more information, ask through your institution's usual process — never tell the customer why.

Step 4: Decide

Every decision needs a coded reason and a narrative.

• Clear (false positive) with an FP-… reason code when the customer is not the listed party. FlowSentry remembers the clear for that subject and list entry; the hit is not raised again unless the customer's details or the entry change.
• Propose true match with a TM-… reason code when the customer is the listed party. A second person must confirm it.
• Escalate to a case with an ESC-… reason code when it needs investigation. A case opens from a snapshot of what you reviewed and the case number shows on the hit.

Depending on your four-eyes policy, a clear may also need a checker. This always applies to hits on statutory (TFS) lists and to hits scoring above your institution's threshold. Such a clear becomes Clear proposed instead of Cleared. Some single-person clears are sampled for QA.

Step 5: The checker confirms

1. The checker sees the hit under Awaiting a checker, in My approvals and in their notifications.
2. They review the same evidence and the maker's narrative.
3. They select Confirm true match or Approve clear, completing step-up, or Reject proposal to send the hit back to open.

You cannot confirm your own proposal, and you cannot clear a hit you proposed as a true match.

A confirmed true match is recorded as True match. The customer's risk rating is recalculated, and your institution's sanctions response procedures apply.

Step 6: Known good customers

If a customer repeatedly hits the same list entry and is clearly not that person, a list manager can choose Propose exclusion on the hit, with a reason and an expiry of at most three years. A different person with list-approval rights approves it on Exclusions (/screening/exclusions). See Add a good-guy screening exclusion.

Tips

• Check the customer's profile before deciding — context resolves most false positives.
• If the page says "The list entry has changed (or been removed) since this hit was raised.", read the current entry before deciding.
• Use Name search (/screening/search) to look a name up against the lists without creating a hit.

Related

• Clear several screening hits at once
• Screening and risk FAQ
• Smarter screening and explainable risk