Bringing Agentic Power to Bear Across AT&T’s Entire Business

We’ve been making big strides in our journey delivering the value of AI to our people, technology vendors and customers. We’re celebrating our wins and leading the charge building an AI-focused telco industry.

We’ve recently earned two new milestones in AI industry benchmarks: the Spider 2.0 text to SQL accuracy leaderboard, and the GSMA Open-Telco LLM leader board. These achievements highlight how we’re not only proficient in using AI technologies: we can build them too.

As we look to what’s next for us in AI, we’re centered on one word: “agentic.” Last year, I talked about how important AI agents – or as we like to call them, “autonomous assistants” – will be to the next wave of AI innovation. That time is now, as AI agents are fulfilling the promise of what AI can do for companies.

Agents move AI from the information economy into the action economy. They go way beyond generating content, which is Gen AI’s specialty (and something we’re using to improve network operations, among lots of other things). Agents take it further by planning and executing a task from beginning to end, with human guidance and intervention when needed. It combines goal understanding, multi-step planning, and tool orchestration all in one package.

We’ve already started to test how AI agents can bring value directly to our customers. Now, we’re expanding how we build agentic AI tools for our employees, putting more power directly in the hands of our teams to build custom AI agents with an update to our internal Ask AT&T tool.

Ask AT&T Workflows gives teams a graphical drag-and-drop agent builder. Teams across AT&T can build agents to automate time-consuming tasks in their own unique ways, which allows them to spend more time focusing on problems that only humans can solve: strategically serving our customers, solving complex challenges and planning for the future.

The first in-production tool built with Ask AT&T Workflows is already proving its value at scale. Built together by the AT&T Business and technology development teams, AI agents can take customer service update requests, synchronize data across systems, and auto-install information in real time. It improves the customer experience and reduces wait times, while allowing employees to focus on higher priority actions.

A human always oversees the “chain reaction” of agents (and has required human checkpoints throughout the process). These agents work within AT&T’s ecosystem, able to pull information and data from Ask Docs and Ask Data, our proprietary GenAI tools that securely house AT&T information, to give uniquely informed insights and actions. Every action the agents take is logged, and when one agent passes off a workload to the next, data isolation, retention policies and role-based access are enforced, keeping data privacy and security top of mind along the full process.

As another example, let’s look at how a network engineer can set up AI agents to help them resolve network alerts and get customers reconnected faster if a network incident occurs. An agent can correlate telemetry to identify the location of the alert’s issue, pull recent change logs, check for known issues, and open a trouble ticket. Then, another agent could propose a path to resolution, writing new code to patch the issue. The engineer would watch over all of it, making sure the agents are performing as expected and taking the right actions. Once it’s revolved, a third agent could compile artifacts and a summary, which the engineer reviews to plan preventative measures for the future and keep customers connected.

This is just one example of infinite possibilities. These agents give our employees a powerful tool: the ability to turn regular language prompts into an artificial assistant that can carry out step-by-step instructions.

The end result? Less time is spent on the minutiae and busy work, and more time spent building the future of customer experience.