AT&T built an AI system to prevent network outages. It reduced customer downtime by more than 12 million hours. (www.businessinsider.com)

AT&T has developed an advanced AI-driven end-to-end incident management system (EEIM) aimed at minimizing network outages, successfully reducing customer downtime by over 12 million hours. With 145 million wireless and 16 million broadband subscribers, the pressure to maintain service continuity is paramount, especially following significant outages, such as a major incident in February 2024 that impacted millions. The EEIM uses a blend of traditional machine learning and generative AI to quickly diagnose service disruptions, suggest remote resolutions, and proactively inform customers about outages, thereby reducing the influx of service requests. The significance of AT&T’s initiative lies not only in its direct impact on customer service but also in the technical advancements it represents within the telecom industry. By leveraging vast data resources—over 10 petabytes—and integrating technologies from providers like MongoDB, Microsoft Azure, and Snowflake, AT&T’s system is designed to adapt and scale with changing demands. The addition of AI agents allows for real-time interaction with customers, enhancing the overall responsiveness to outages. This proactive approach sets a new standard for operational efficiency in telecom, showing how AI can fundamentally reshape service and incident management in high-stakes environments.