AT&T uses agentic AI to reinvent business processes from the ground up

AT&T Corp. is moving beyond using artificial intelligence to improve employee productivity, increasingly applying autonomous agents to redesign how its business operates. Nearly 500 agentic workflows are already in production, and the telecom giant plans to triple that number next year.

AT&T now processes about 45 billion AI tokens daily across the company, according to Luke Corbin (pictured, below), assistant vice president of digital transformation and AI at AT&T Business. Its agentic AI platform handles more than 1 million transactions monthly. Many of those workflows now operate autonomously, handling tasks ranging from customer provisioning to sales prospecting.

Speaking at an Amazon Web Services Inc. event in Boston earlier this month, Corbin said the company’s approach emphasizes rethinking entire processes rather than simply automating individual tasks. AT&T has moved quickly from experimenting with generative AI to systematically identifying business processes that can be redesigned around agentic capabilities.

At the heart of the transformation is Business Agent Foundry, a group established in early 2025 to work with individual business units. Rather than soliciting isolated automation ideas from employees, the foundry maps business processes using industry frameworks and identifies opportunities to eliminate unnecessary steps, organizational handoffs and legacy application dependencies, Corbin said.

One early initiative involved automating interactions with AT&T’s mainframe computers. Changes to customer fiber installation orders previously required employees to retrieve tickets from queues, manually update records and return them to the workflow.

Agentic automation eliminated portions of that manual process, reducing delays and accelerating service delivery.

“It was kind of cool having some of the most advanced technology talking to some of the oldest that we have,” Corbin said.

In another example, AT&T halved the time required to provision dedicated internet connections in buildings where fiber infrastructure was already installed.

A different approach

AT&T’s architecture breaks from conventional approaches that build large numbers of independent, specialized agents. Instead, the company maintains a relatively small set of agents that execute business workflows using centrally managed business logic.

“We store the business logic in context or in memory,” Corbin said.

Retrieval agents identify applicable knowledge, while planning agents determine the actions, tools and data needed to complete a task.

Governance is central to the process. Individual employees can independently create personal productivity agents, but workflows affecting customers must pass through the foundry and undergo security, privacy and compliance reviews.

“We’ve democratized agents to the point where people can use them for personal productivity, but not to where you would drive business workflows that could touch your customers or materially have an impact on your business,” Corbin said.

Every production workflow initially includes human oversight. Only after accuracy reaches the mid- to high 90% range within a permissible risk profile can agents operate autonomously. Continuous monitoring evaluates performance and detects drift.

Success metrics include labor savings, revenue generation, operating costs and customer satisfaction. Corbin said production AI investments are achieving fivefold returns within a year.

A 151-year-old company that has been using computers since they were invented “is built on a lot of legacy processes and applications,” Corbin said. “The cost of maintaining those applications, training employees to use them and ongoing maintenance is very high. When you can offset some of that with simpler generative workflows, you’ve got a business case that’s built for itself.”

Sales prospecting is one beneficiary. Agents identify promising customers by validating their locations and checking fiber availability before representatives invest time pursuing sales opportunities.

Harmonizing data

Scaling up those capabilities requires more than AI models. AT&T has had to consolidate fragmented data and provide reliable connectivity across data centers, cloud infrastructure and mobile workers.

“We’ve been through a lot of mergers and acquisitions,” Corbin said. “You can imagine what our data infrastructure looks like.”

Encouraging employees to come forward with ideas for using agents has helped limit resistance from those who worry AI could replace them. So has the opportunity to replace inefficient processes.

“There are a tremendous number of people in the organization who have knowledge and passion about the business process that they used to use but that they didn’t like,” he said.

Corbin said AT&T is now shifting its focus from improving existing processes toward using agents to launch services, enter markets and support new sales channels.

Ultimately, Corbin said responsibility for AI-driven business transformation should be distributed throughout the organization, rather than concentrated in a specialized team like Business Agent Foundry.

“I feel privileged to have the team that I do, but in the long run, my team shouldn’t exist,” he said. “What we do should be completely distributed.”

Photo: Flickr CC