Key Takeaways
- AT&T, Southwest, Canada Goose, and others share lessons they learned from their agentic deployments.
- Common themes include using AI to turn cost centers into value centers, building trust into agentic systems, and the importance of evolving along with customer needs.
Salesforce customers across industries recently described their agentic rollouts, sharing what worked, what surprised them, and what they’d tell others just getting started on their own AI journeys.
A clear pattern emerged. Customers seeing real results don’t just deploy. They listen to their users, define success, and build the guardrails to sustain it. Here are three lessons that come up again and again.
AI unlocks the value in the human side of your business
A consistent theme across Salesforce customers is that delegating time-consuming tasks to agents gives people the opportunity to offer more value as humans. When agents take on routine tasks, humans get to do the work that only they can handle.
At Canada Goose, human staffers no longer spend their days answering customer queries about when a jacket will arrive or how to exchange a sweater for a different size. Agentforce handles that now, autonomously resolving 89% of routine messaging inquiries and 15% of calls. Instead, the company’s “Style Experts” focus on personalization, personal shopping, and moments of connection.
“We wanted to make sure that AI wasn’t removing the human; it was actually elevating the human,” said Dennis Liut, the company’s Head of Global Customer Experience and Revenue. “And now we have an Experience Center that used to be a call center that is bringing in millions in revenue.”
Canon is also pivoting its service organization from a cost center to a value center, as agents field questions that don’t require a human touch. “There are all kinds of opportunities to leverage this data — to harness that voice of the customer,” said Bill Duval, VP of Service Information Management and Technology. “Driving customer effort down while capturing that information and driving product development — it’s incredibly valuable to us.”
AT&T’s strategy centers on using AI to make human interactions better, not replace them. That includes its “agent briefing,” which uses check-in details captured through AT&T’s retail app to collect and package a customer’s key info so that a retail expert knows who they’re helping and why before they even connect. Making it work requires investing in data quality and governance up front so the briefing delivers accurate, relevant context. It also means being deliberate about where AI enters the conversation and where it steps back.
The goal, said John Miller, VP of Consumer and Business Solutions, is proactive customer service. Anticipate what someone needs before they ask, and equip the human on the other end of the interaction to deliver.
Discover what customers actually want — and keep asking
Several customers warned against over-engineering the rollout based on internal assumptions.
“When you start off, you have a lot of hypotheses around what customers want. And you quickly learn through doing what they actually need,” said Rudi Khoury, Chief Digital Officer at Fisher & Paykel. That creates a powerful argument for diving in while remaining ready to iterate. “Just get started and learn directly from your end user.”
Southwest Airlines applies the same principle to its roadmap, using real chat data to shape where it deploys next. Actual customer questions, not internal guesses, drive prioritization.
“We’re using a lot of the data that we’re finding in the chat experience to help drive and know where our customers want to be serviced in this experience,” said Megan Rauber, Manager of Customer 360 for the airline. “We’re continuing to see our CSAT score improve week over week as we’re continuing to introduce new capabilities.”
The lesson is that deployment isn’t a one-time decision; it’s an ongoing conversation with your customers.
Build in the control and trust you need
In many regulated industries, trust matters more than speed or efficiency. The customers seeing the most durable success build governance into their implementations from the start.
AT&T gave itself the ability to turn agents on or off at any time, implementing a trust layer and its own LLM for anything touching proprietary IP.
Sammons Financial Group took a similarly disciplined approach. The company ran through more than 200 guardrails and tests before going live and made a deliberate choice to prevent the model from learning from live conversations, effectively eliminating drift in a compliance-sensitive environment. To ensure safety, Andrew Walling, AVP of Capability Planning and Delivery, noted that the company implemented a fail-safe: “We have a kill switch, a supervisor agent that’s constantly listening to all of those phone calls, checking the values in the CRM against what was said on the line.”
Takeda is automating the administrative burden that consumes its sales reps’ time while remaining carefully within the constraints that regulate the pharmaceutical industry. That balance requires staying current as the technology itself keeps advancing. “AI is moving at such a rapid pace. You’ve got to keep up with it,” said Samer Ansari, Head of Commercial Innovation and External Experience.
It’s a useful reminder that governance isn’t a constraint on ambition. It’s what makes ambition sustainable.
