our technology
Layered Voice Analysis (LVA)
Our service uses Layered Voice Analysis (LVA) to detect anger in recorded conversations.
Invented in 1997 by Amir Liberman and his company Nemesysco, LVA identifies various types of stress levels, cognitive processes, and emotional reactions that are reflected in different subtle properties of the voice. LVA uses a unique mathematical process to detect different types of patterns and anomalies in the speech flow and classify them in terms of stress, excitement, confusion and other relevant emotional states.
Extracting up to 151 vocal parameters, LVA ignores what a subject is saying and the language being used, focusing instead on the patterns and anomalies in the speech flow..
Language neutral
Most "sentiment analysis" products examine what a speaker says (using Natural Language Processing) to predict how the person feels. This approach only captures a fraction of actual communication, overlooking the sincerity, stress, and hidden intent in the human voice. This inadequacy is particularly acute in call center environments where unhappy callers may not explicitly express their concerns.
Layered Voice Analysis (LVA) relies on uncontrolled properties of the human voice making it independent of language and culture. Today, the technology is deployed in 60 countries, processing over 200 million conversations a month, in 140 languages.
Secure
With deep roots in homeland security, LVA meets the world's most rigorous standards for data privacy, security, and ethical AI, including SOC 2 Type certification. Data safeguards include encryption in transit and at rest, purpose-limited processing, and zero data retention.
Broadly deployed
LVA is currently deployed by governments, police, banks, insurance companies, corporate human resource departments, medical facilities, and call centers performing critical tasks such as security clearances, police investigations, fraud prevention, job applicant screening, and call center quality monitoring.