AI Agent News/stories/at-t-uses-agentic-ai-to-rebuild-business-processes-with-nearly-500-workflows-in.md

Infrastructure

AT&T uses agentic AI to rebuild business processes, with nearly 500 workflows in production

AT&T says it has nearly 500 agentic workflows in production and plans to triple that number next year, applying autonomous agents to redesign how its business operates. The company processes about 45 billion AI tokens daily and its agentic platform handles more than 1 million transactions a month.

Source: siliconangle.com/2026/11/14/att-uses-agentic-ai-to-reinvent-business-processes-from-the-ground-up

AT&T is moving past using AI for employee productivity and is applying autonomous agents to rework its business processes. Luke Corbin, assistant vice president of digital transformation and AI at AT&T Business, said the company's approach emphasizes rethinking entire processes rather than automating individual tasks. The Business Agent Foundry, formed in early 2025, maps processes using industry frameworks and looks for steps, handoffs and legacy dependencies to remove.

AT&T's design uses a relatively small set of agents that run workflows against centrally managed business logic, rather than many specialized agents. Retrieval agents find relevant knowledge and planning agents decide the actions, tools and data a task needs. Workflows that touch customers must pass through the foundry and security, privacy and compliance reviews, and every production workflow starts with human oversight.

Corbin said agents can move to autonomous operation only after accuracy reaches the mid- to high 90% range within an acceptable risk profile, and that continuous monitoring checks for drift. He also said AI investments in production are returning fivefold within a year. Scaling the work required consolidating fragmented data across data centers, cloud infrastructure and mobile workers.

Why it matters if you build agents

Developers building agents for enterprises can see a concrete pattern: a small set of agents driven by centrally managed business logic, with retrieval and planning agents, plus gated autonomy tied to measured accuracy.

Teams should plan for human review on every production workflow at first, and for drift monitoring before granting any agent autonomy over customer-facing tasks.

Get stories like this every weekday

One short email with the day's agent tooling news.

Subscribe