Matching poorly identified fraud suspects with customers in the master database
Fuzzy search and simplified match recording save time while improving the accuracy of the results
As part of dealing with Suspicious Activity Reports (a.k.a. SARs, the most important reports of potentially illegal activity), associates in the AML Compliance department at a small international bank match the suspects on each SAR to the bank’s customers. The manually-entered data on the SAR is notoriously unreliable — ridden with typos, inaccuracies, and abbreviations — which makes it difficult to match against the customer database. This process takes 5–15 minutes when done manually. Our product, SyNAPSE, automatically searches the customer database and returns the best potential matches, taking only a few seconds of the associates’ time.
User Story 1: An AML Compliance associate searches for the name listed on the SAR
- The associate manually copies and pastes the name of a suspect from the SAR into the web front end of the bank’s customer database
- The associate searches for likely variants of the name: abbreviations, nicknames, transposed names, and excluding middle names
- Potential customer matches are opened in a new browser tab
User Story 2: An AML Compliance associate searches for the ID number listed on the SAR
- The associate manually copies and pastes the ID number of a suspect from the SAR into the web front end of the bank’s customer database
- Potential customer matches are opened in a new browser tab
User Story 3: An AML Compliance associate compares all information listed on the SAR against the information for any potentially matching customers
- The associate reviews extended information for one customer — addresses, account numbers, and birth dates — and compares it against available information on the SAR
- The associate repeats this process for each potentially matching customer, switching back and forth between the two databases
- The associate copies and pastes the customer ID numbers of each verified match (there may be more than one) into a temporary document
SyNAPSE automates the search and displays the best potential customer matches
In the background, names and ID numbers from all new SARs are automatically compared against the customer database. The text is subjected to a fuzzy search that looks for near-matches of both the full name and each part of the name. Potential matches are recorded within the system for the associate to review at any time. This completely eliminates the work in User Stories 1 and 2.
The results of the automatic search are shown on a review screen with the name, ID number, and extended information for both the SAR suspect and the potentially matching customers. The associate can link a customer to a SAR with a single click, and unlink them just as easily in the event of an error. This reduces the work in User Story 3 to a few clicks, rather than repeatedly switching between windows and browser tabs.
