AT&T is moving beyond traditional AI applications, deploying nearly 500 autonomous agent workflows that redesign entire processes rather than just automate tasks. This approach has helped the telecom giant accelerate service delivery, cut operational costs, and improve customer interactions at scale.
- Nearly 500 agentic AI workflows are in production, with plans to triple next year
- AI processes 45 billion tokens daily and executes over 1 million transactions monthly
- Agentic automation cuts provisioning times and reduces manual mainframe interactions
What happened
AT&T has significantly expanded its use of agentic AI to transform business operations. The company now runs close to 500 autonomous agent workflows that automate and optimize entire business processes rather than isolated tasks. These workflows handle a variety of functions, including customer provisioning, sales prospecting, and mainframe system interactions. The AI platform processes around 45 billion tokens daily and executes more than one million transactions every month.
This transformation is coordinated by the Business Agent Foundry, established in early 2025. The foundry collaborates with different business units to map and streamline processes, removing unnecessary steps and dependencies while embedding AI capabilities directly into workflows. This approach replaces the traditional model of many specialized agents with a smaller set of multi-functional agents that operate based on centrally stored business logic.
Why it matters
AT&T’s agentic AI initiative addresses long-standing challenges associated with legacy processes and applications, which are costly to maintain and complex for employees to navigate. By redesigning workflows end to end, the company reduces manual effort and accelerates service delivery—such as halving provisioning time for dedicated internet connections where fiber is present. This results in strong business benefits including labor savings, increased revenue opportunities, operational efficiencies, and improved customer satisfaction.
Importantly, AT&T has implemented strict governance to ensure AI workflows impacting customers meet rigorous security, privacy, and compliance standards. Human oversight remains integral until performance stabilizes above 90% accuracy, and continuous monitoring guards against performance drift. This disciplined approach enables AT&T to realize robust returns on AI investments and more confidently scale autonomous agents across the enterprise.
What to watch next
AT&T plans to triple the number of agentic AI workflows over the coming year as it continues to identify and redesign additional business processes. This expansion will likely require further consolidation and integration of data infrastructure and cloud connectivity to support reliable AI execution at scale. The ongoing maturation of AI governance and operational frameworks will also be critical to balancing innovation speed with risk management.
The company’s success in deploying autonomous agents for complex workflows like sales prospecting and customer provisioning could serve as a model for other large enterprises seeking to modernize legacy systems. Observers will want to monitor how effectively AT&T scales agentic AI across diverse business units and whether it can replicate the current levels of labor savings and customer satisfaction gains on a broader basis.