AT&T Integrates Autonomous Agents to Transform Business Operations
AT&T Corp. is advancing its use of artificial intelligence beyond enhancing employee productivity, increasingly implementing autonomous agents to fundamentally redesign its business operations. The telecom giant currently has nearly 500 autonomous workflows in production, with plans to triple that number by next year.
As reported by Luke Corbin, AT&T’s assistant vice president of digital transformation and AI, the company processes approximately 45 billion AI tokens daily. Its autonomous AI platform manages over 1 million transactions each month, with many workflows now functioning autonomously to perform tasks such as customer provisioning and sales prospecting.
During a recent event hosted by Amazon Web Services Inc. in Boston, Corbin explained that AT&T’s strategy focuses on reimagining entire business processes rather than merely automating individual tasks. The company has swiftly transitioned from initial experiments with generative AI to carefully identifying business processes that can be redesigned to leverage these agentic capabilities.
Central to this transformation is the Business Agent Foundry, a team established in early 2025 focused on collaborating with different business units. Instead of gathering isolated automation suggestions from employees, the Foundry maps out business processes using industry frameworks to identify areas for simplification, including unnecessary steps, organizational handoffs, and dependencies on outdated applications, according to Corbin.
One notable initiative involved automating interactions with AT&T’s mainframe systems. Previously, modifications to customer fiber installation orders required employees to manually retrieve tickets from queues, update records, and integrate them back into workflows. With agentic automation, AT&T has streamlined many manual aspects of this process, resulting in reduced delays and faster service delivery.
Corbin remarked, “It was kind of cool having some of the most advanced technology talking to some of the oldest that we have.” In another example of efficiency, AT&T has halved the time it takes to provision dedicated internet connections in buildings with pre-installed fiber infrastructure.
Innovative Agent Architecture
AT&T’s architecture deviates from traditional methods that often utilize numerous independent, specialized agents. Instead, the company employs a smaller set of agents that execute business workflows guided by centrally managed business logic.
Governance plays a critical role in this process. While employees can independently create personal productivity agents, workflows that affect customers must undergo security, privacy, and compliance reviews by the Foundry. “We’ve democratized agents to the point where people can use them for personal productivity,” Corbin stated, “but not to the extent that it would drive business workflows that could affect customers or meaningfully impact the business.”
Every production workflow starts with human oversight. Only after achieving an accuracy level in the mid- to high 90% range within a defined risk profile can agents operate autonomously, with continuous performance monitoring in place to detect any deviations.
AT&T measures success through various metrics, including labor savings, revenue generation, operating costs, and customer satisfaction. Corbin noted that investments in production AI have yielded fivefold returns within a year.
As a 151-year-old company with a long history of computer use, AT&T faces the challenge of legacy processes and applications. “The cost of maintaining those applications, training employees to use them, and ongoing maintenance is very high,” Corbin remarked. “When you can offset some of that with simpler generative workflows, you’ve got a business case that’s built for itself.” Sales prospecting has seen significant benefits, with agents helping identify promising customers by validating locations and checking fiber availability before sales representatives invest time in pursuing leads.
Consolidating Data for Scalability
Expanding these capabilities requires more than just AI models; AT&T must also consolidate fragmented data and ensure reliable connectivity across data centers, cloud infrastructure, and mobile workers. “We’ve been through a lot of mergers and acquisitions,” Corbin mentioned, highlighting the complexities of their data infrastructure.
Encouraging employees to submit ideas for utilizing agents has mitigated concerns about AI replacing jobs, along with the appeal of eliminating inefficient processes. “There are a tremendous number of people in the organization who have knowledge and passion about the business processes that they used to use but didn’t like,” he said.
Looking forward, Corbin emphasized a shift from merely improving existing processes to leveraging agents for launching services, penetrating new markets, and supporting fresh sales channels. He believes that the responsibility for AI-driven business transformation should not be confined to a specialized team like the Business Agent Foundry, but rather distributed throughout the organization.
“I feel privileged to have the team that I do,” Corbin concluded, “but in the long run, my team shouldn’t exist. What we do should be completely distributed.”

