Image of a chat interaction between a person and Agentforce with a phone icon to the top right.
AT&T logo

AT&T aims to take the hassle out of in-store phone upgrades with Agentforce

Two AI agents turn customer context into recommendations, helping retail experts work toward a 10-minute sales cycle.

September 14, 2026

The Results

8
weeks from ideation to pilot-ready
67 %
shorter sales cycle targeted

Context spread across systems

AT&T is one of the world’s largest telecommunications companies, driving world-class connectivity across wireless and broadband networks for millions of consumer and business customers. When customers are ready for a device upgrade, they can visit an AT&T retail store for help and guidance from a retail expert.

Before, the AT&T experts had access to customer information, but getting the context they needed took significant manual effort. Account details, purchase history, network data, and past interactions lived across multiple systems, making experts search their tablets while customers waited.

To find the right device or offer, AT&T had to piece together context from the customer’s account, device ecosystem, and purchase history to determine the best fit, making the interaction dependent on each expert’s experience and judgment.

Altogether, that added time to the in-store sales process, which could take more than 30 minutes. For AT&T, reimagining the upgrade experience wasn’t just an efficiency challenge – it was a relationship imperative.  Their customers’ time is valuable, and they want to make every minute in the store count.

Personalized insights at a glance

Agentforce brings the insights and guidance directly into their in-store workflow. A Briefing Agent summarizes what matters for each customer and their visit, while a Cart Optimization Agent recommends relevant devices and offers, helping experts quickly understand the customer and confidently guide the conversation.

“The vision is to give our people a Ph.D.-level understanding of the customer’s relationship with AT&T and our products, so every expert has what they need in one place to create the best possible experience,” explained John Miller, Vice President of Consumer and Business Solutions.

For example, when a customer checks in, the Briefing Agent surfaces relevant details such as their tenure, account standing, upgrade eligibility, and service usage right on the expert’s tablet. With that information at a glance, the retail expert can greet the customer already knowing what matters, helping reduce the need for customers to explain information AT&T already knows. The expert can focus on what the customer is saying and spend the conversation helping rather than searching.

If a customer indicates that they’re there to upgrade their phone during check-in, that choice signals the Cart Optimization Agent to prepare relevant device and offer recommendations based on their individual needs and history. When the retail expert connects with the customer, the recommendations and rationale behind them help guide the conversation, while the expert helps the customer make an informed final choice.

Real-time data, guided reasoning

Behind AT&T’s in-store app, Salesforce connects real-time customer data with Agentforce reasoning — embedding trusted agentic intelligence directly into AT&T’s retail workflow. Customer selections within their check-in journey set the parameters for Agentforce to then suggest the latest offers and options that meet their needs.

At check-in, customer data is pulled together in real time via Data 360 rather than pre-stored, in line with AT&T’s data requirements. Data 360 harmonizes the customer, account, order, purchase, network, and interaction data needed from systems including Sales Cloud and Service Cloud, and Snowflake and Azure through Zero Copy. Data 360 data graphs retrieve the relevant records on demand with sub-second retrieval, while MuleSoft connects more than 50 of AT&T’s distinct legacy backend systems. Together, they give Agentforce a complete, actionable view of the customer while the underlying data remains in AT&T’s environment.

The Briefing Agent adds the customer to the “In Store Now” queue experts use to connect with waiting customers. If the customer selected an upgrade during check-in, the Cart Optimization Agent runs in parallel, with results and recommendations ready for the retail expert to review with  the customer.

There, the Cart Optimization Agent weighs signals across the customer’s history to rank options. If someone consistently buys iPhones and orange cases but maxes out their usual low-storage device, for example, an iPhone 17 Pro in Cosmic Orange with 512GB might be the “best pick,” alongside “good” and “OK” picks. Each includes a rationale, giving the sales rep options to discuss and the discretion to make the final recommendation.

This architecture diagram, titled "How AT&T grounds AI agents with trusted context," illustrates how AT&T integrates various data sources and systems to power its AI agents.

Faster, more consistent selling

AT&T moved from ideation to working pilot in about eight weeks with help from Salesforce Professional Services. Their goal is to cut the retail sales cycle by 67%, from more than 30 minutes to 10 minutes. 

It’s a shift from “heads down” searching to “heads up” service, giving retail experts the customer context and recommendations they need to focus on the conversation rather than searching across multiple systems. And by putting that same intelligence and guidance in every expert’s hands, AT&T aims to give customers a consistent experience regardless of who helps them or how long they’ve been in the role. “When customers feel like we know who they are and understand what they need, we become more than someone selling them a product — we become a trusted adviser. That’s how you create a customer for life,” said Miller. 

AT&T is already imagining what comes next to make the in-store experience even better. For example, upon arriving at the store, customers could speak with Agentforce Voice – just like an in-store associate – to check in and describe what they need, versus just a chat experience. Agentforce could then combine insights from that conversation with existing customer context, giving the retail expert a richer picture to seamlessly continue the customer’s upgrade journey.

Why AT&T chose Salesforce

The image depicts a 3D digital illustration of a faceted, low-poly white human head positioned on top of a grid of white and iridescent purple cubes.

Extensible platform

AT&T could build on their existing MuleSoft APIs and Salesforce investments rather than starting from scratch, providing a foundation for the new agentic experience.

The image depicts a 3D graphic of two smartphones and data cylinders enclosed within a circular frame topped by an Agentforce icon.

Headless flexibility

Agentforce works behind AT&T’s existing in-store app instead of requiring a separate conversational interface. The headless design lets clicks and check-in selections trigger the right Salesforce workflows while experts stay in a familiar experience.

Icon of an iridescent, glowing, pixelated cube.

Data readiness with Zero Copy

AT&T didn't need to move their entire data estate into Salesforce. MuleSoft connects legacy backend systems, while Data 360 and Zero Copy with Snowflake harmonize order, purchase, and interaction data in real time — giving AT&T a unified customer view in AT&T’s existing environment.

An icon of a brown leather briefcase with the blue Salesforce cloud on it.

Salesforce Professional Services

“This was truly one team — there was no ‘us and them.’ AT&T brought our vision and front-end expertise, Salesforce brought deep data and agentic expertise, and together we moved at lightning speed to go from idea to pilot-ready in just eight weeks,” said Miller.

Get started