The Skorozvon service has introduced an AI-powered Answering Machine Detection (AMD) feature designed to identify advanced virtual assistants. These modern automated systems often use varied speech patterns, making them difficult to detect with traditional audio template matching. The new tool analyzes sound characteristics immediately after a connection is established, such as silence-to-speech transitions and speech activity, to terminate calls before they reach human operators.
According to internal research conducted by the company between 2025 and 2026, automated systems account for 73% of all calls. Previously, detection systems relied on comparing audio snippets against pre-recorded templates, which proved ineffective against newer, dynamic voice assistants. By filtering out these non-human interactions, the system aims to increase the number of productive conversations per hour and improve the accuracy of contact database reporting.
