Cutting Analysis Time by 30%: How Phonexia Helped European Law Enforcement Fight Organized Crime

The unit had to investigate a network of seven people involved in money laundering and drug trafficking crimes. It was necessary to analyze 750 hours of intercepted audio recordings and an additional 120 hours of audio gathered from covert home surveillance. The number of cases running in parallel, combined with the scarcity of time and human resources, demanded an automated software solution that would significantly improve agents’ efficiency and speed up the extraction of important and actionable insights from the recordings. The ultimate goal was to reduce the time needed for call analysis and minimize the number of people involved. It was important to be able to solve such cases as quickly as possible to prevent any more damage being caused by these criminal activities.

Phonexia’s AI-powered voice biometrics and speech recognition technologies were chosen as the perfect fit for the software solution developed by a Phonexia partner. These technologies offered rich integration possibilities, on-premise development, and the most accurate speaker identification capability available on the market. Although the solution used the full range of Phonexia technologies, it was mainly the crucial combination of Phonexia Speaker Identification and Phonexia Keyword Spotting technologies that enabled the extremely effective way of analyzing speakers and the content in large amounts of audio.

“Despite Phonexia’s extensive range of speech and voice recognition technologies, Phonexia experts were always ready with an answer and very supportive in all areas. All clarifications were delivered even faster than the agreed period.”

Before the solution was implemented, typical call analysis — including speaker identification, manual transcriptions, extraction of personal names, meeting places, connections, and more — would take the team approximately 370 man-days to finish. This accounted for 60–70% of the entire case workload. Once the automated solution was in place, the speed of call analysis immediately improved. Even though the criminals were using different SIM cards, their voices remained the same.

The automated approach made it possible to discover all phone numbers used within the investigated group of criminals. Furthermore, at least 30 new phone numbers were connected to previously unknown criminal organizations, and new suspects were put on a list as points of interest for further investigation based on a combination of link analysis and speaker identification results. Phonexia technologies also enabled rapid filtering of intercepted calls so they could be prioritized and analyzed more effectively.

Call analysis sped up by 30%, resulting in an average saving of approximately 111 man-days on the investigated case. It was further discovered that 20% of all audio recordings didn’t include any speech at all. The Unit for the Suppression of Organized Crime was able to analyze intercepted calls much faster, allowing fewer analysts to respond to criminal cases more rapidly.