AT&T Reshapes Business Processes with Agentic AI, Processing 45 Billion Tokens Daily
AT&T Corp. is moving beyond using artificial intelligence to boost employee productivity, increasingly deploying autonomous agents to redesign how its business operations run. With nearly 500 agentic workflows…
AT&T Corp. is moving beyond using artificial intelligence to boost employee productivity, increasingly deploying autonomous agents to redesign how its business operations run. With nearly 500 agentic workflows in production, the telecom giant plans to triple that number next year. Luke Corbin, assistant vice president of digital transformation and AI at AT&T Business, said AT&T currently processes about 45 billion AI tokens company-wide each day. Its agentic AI platform handles over 1 million transactions per month. Many of these workflows now run autonomously, handling tasks ranging from customer onboarding to sales lead generation.
Earlier this month at an Amazon Web Services event in Boston, Corbin said the company's approach emphasizes rethinking entire processes rather than simply automating individual tasks. AT&T has rapidly moved from experimenting with generative AI to systematically identifying business processes that can be redesigned around agentic capabilities.
At the core of this transformation is the "Business Agent Foundry," a team established in early 2025 that collaborates with business units. Corbin said the foundry does not solicit isolated automation ideas from employees but instead uses industry frameworks to map business processes and identify opportunities to eliminate unnecessary steps, handoffs, and legacy application dependencies.
One early initiative involved automating interactions with AT&T's mainframe systems. Changes to customer fiber installation orders previously required employees to pull tickets from a queue, manually update records, and return them to the workflow. Agentic automation eliminated parts of that manual process, reducing latency and speeding up service delivery. "It's kind of cool to have some of our most advanced technologies talking to some of our oldest technologies," Corbin said.
In another example, AT&T cut in half the time required to activate dedicated internet connections in buildings with existing fiber infrastructure.
A different approach
AT&T's architecture breaks from the traditional method of building numerous standalone, specialized agents. Instead, the company maintains a relatively lean set of agents that use centrally managed business logic to execute workflows. "We store business logic in context or memory," Corbin said. Retrieval agents identify applicable knowledge, while planning agents determine the actions, tools, and data needed to complete tasks.
Governance is central to the process. Individual employees can independently create personal productivity agents, but workflows that affect customers must pass through the foundry and undergo security, privacy, and compliance reviews. "We've democratized agents to the point where employees can use them for personal productivity, but we have not democratized them to the point where they can drive business workflows that might touch customers or materially impact the business," Corbin said.
Every production workflow initially includes human oversight. Agents only run autonomously after accuracy reaches the mid-to-high 90% range within acceptable risk parameters. Continuous monitoring evaluates performance and detects drift. Success metrics include labor savings, revenue growth, operating costs, and customer satisfaction. Corbin said production AI investments delivered a fivefold return within one year.
Corbin said the 151-year-old company, which has used computers since their invention, "is built on a lot of legacy processes and applications. The cost of maintaining those applications, training employees on them, and keeping them up is very high. When you can offset some of that with simpler generative workflows, the business case speaks for itself."
Sales lead generation is one beneficiary. Agents identify promising customers by verifying prospect locations and checking fiber availability before sales representatives invest time in pursuing opportunities.
Orchestrating data
Scaling these 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 M&A. You can imagine what our data infrastructure looks like," Corbin said.
Encouraging employees to proactively propose agent use cases helps reduce resistance from those worried AI might replace them. Opportunities to replace inefficient processes with agents serve the same purpose. "There are a lot of people in the organization who have knowledge and passion about business processes they've used in the past but didn't like," he said.
Corbin said AT&T is now shifting focus from improving existing processes to using agents to launch services, enter markets, and support new sales channels. Ultimately, Corbin said, responsibility for AI-driven business transformation should be distributed across the organization rather than concentrated in a dedicated team like the Business Agent Foundry. "I'm honored to have the team we have, but in the long run, my team shouldn't exist. What we do should be fully distributed," he said.
insigtX content is informational and educational, not investment advice.