Accurately detecting deception has been studied extensively in various sub-fields related to psychology and law enforcement, where it takes years to master the skills. To date, this skill set is primarily reserved for interrogation in law enforcement. However, the ability to detect such suspicious behavior is highly sought-after in financial and legal services.
Such organizations generally employ teleoperators or customer service representatives to meet the demand of a large customer base. These teleoperators are not usually behavioral experts, hence fraud costs organizations large amounts of money every year.
Using GPU-powered deep learning and state-of-the-art natural language understanding technology, this talk presents LexiQal, an automated explainable deception detection system, developed in cooperation with law enforcement behavioral experts, to provide automated fraud detection to support teleoperators in insurance and financial services.
Nigel Cannings, CTO at Intelligent Voice presented this session at NVIDIA GTC 2021.
If you would like to know more about how we can help you build a system to address this problem, email us at firstname.lastname@example.org.