Triage analysis of fraud incidents

Many kinds of business have to deal with fraud, but it’s a particular problem for the financial and insurance industries. Being alerted early to suspicious activity, ranking incidents for further action and ruling out false positives are all required to manage fraud security and maintain customer satisfaction.

Hospital accident and emergency wards use the concept of triage – sorting people in order of seriousness, regardless of when they arrived. Exactly the same principle can be applied to fraud analysis. Instead of chasing up hundreds of small incidents, time and resources can be better spent concentrating on the large, costly cases and working one’s way down. It would be nice to treat every case as serious, but commercially it makes no sense. Alerts deemed noteworthy but of lower priority can be diverted to other departments. With our solution:

• Automatic data feeds from a variety of sources can be processed into a single unified data model, designed to support the triage analyst.
• The automated data stream can be further supplemented by ad hoc queries to other internal and external data sources.
• Link Explorer’s query-by-example feature makes a wide variety of complex parameterised queries available to the analyst in just a few clicks.
• The full power of Link Explorer’s network, temporal and spatial visualisations and analytics can then be rapidly and effectively applied to the query results to identify high risk cases via clusters and links, suspicious timelines and past behaviour, co-located fraud incidents etc.
• Integration with in-house systems allows fraud incidents to be assigned for investigation or moved to the appropriate upstream business processes.

As automated fraud detection continues to improve and identify more cases, triage analysis is becoming ever more important. By helping analysts to quickly identify those cases requiring further investigation Link Explorer improves the effectiveness of your fraud teams, helping to minimize losses and improve competitiveness.