TTEK: Modernizing Customs and Border Processing

In a customs and border environment, analysts and officers are required to make a decision on high-risk cargo and passengers in real time. Such ability helps them to identify high-risk cargo and passengers before arrival at the border, and conduct an inspection when there is a serious enough threat present. While there are numerous predictive solutions designed to operate on historical data, only a few provide the actionable insight essential for customs in the front end border operation where real-time decisions are made. With years of domain experience, TTEK endorses a blend of cutting-edge products with operational and academic subject matter expertise to assist nations with modernizing their border processing systems and methodologies. Leveraging a predictive model of known intelligence indicators that match an inbound shipment or passenger, the company can help to provide a strategic and operational target of a looming threat.

Although a new start-up, TTEK’s legacy is rooted in a former company named GreenLine Systems that was eminent for crafting deductive and inductive targeting solutions for several global customs administrations. Presently, TTEK collaborates with Data Science Ltd., a New Zealand-based firm that has a deep understanding of customs data and machine-learning approaches. “We bring in Data Science’s expertise on predictive analytics, and merge this with our knowledge of border/commercial trade processing to deliver a risk based solution for custom services worldwide to help these clients boost their ability to identify high-risk cargo as well as passengers,” explains Chris Thibedeau, CEO at TTEK.

Together with Data Science, TTEK develops a risk management application called ‘RiskLab.’ In the customs landscape, 90 nations use ASYCUDA (Automated System for Customs Data), a rudimentary UN customs processing system whose risk management capabilities are next to inexistent. What TTEK does is integrate RiskLab with ASYCUDA via an API for data access. “We use an array of machine learning based models on historical data to predict threats in real time,” notes Thibedeau. Once these models are developed based on historical data, they are installed to complement ASYCUDA to more effectively detect high-risk consignments. When inbound shipments fit the model, the container is flagged to the inspection team for closer scrutiny and a potential inspection.