AI successes make headlines. The messy work of AI adoption rarely does.
At RetailClub’s AI Festival, Aaron Rose, President and Chief Commercial Officer at At Home, described a feeling familiar to many retail leaders: everyone sees what other companies are doing with AI and wonders whether they are falling behind. What is harder to see is the experimentation, uncertainty, and organizational work behind each success story.
That contrast captured much of the conversation at the festival. Sessions explored shopping agents, new roles, and changing customer relationships, but the practical question was how to turn all that possibility into progress.
Here are the conversations that showed what progress looks like behind the scenes.
Focus on a Problem Worth Solving
When everyone else appears to be moving faster, every new AI opportunity can feel urgent. During “The AI-First Marketing Organization” panel, Rose described how At Home’s early efforts became a “whack-a-mole” approach: plenty of activity, without enough focus on what would deliver a return.
At Home responded by forming a cross-functional AI Working Council and setting six-month roadmaps for its workstreams. Teams now weigh potential projects against the effort required and the value they could create. Julie Bowerman, former CMO of Kellanova, joined the panel and described a similar need to build a pipeline of use cases, then choose which opportunities deserve investment.
Sarah Engel, President of January Digital, offered marketing teams a practical starting point: choose one important workflow, decide what AI will handle and what a person still owns, then measure whether the new approach works over 90 days. The results can tell the team whether to scale the change, revise it, or move on.
Redesign the Work, Not Just the Tool Stack
Choosing where to focus leads to a harder question: how does the work need to change? At RetailClub, Greg Pulsifer, SVP of eCommerce at Sam’s Club, described a people-led, technology-powered approach. Sam’s Club has put that idea to work as members leave its stores. Instead of having employees check receipts against shopping carts at the exit, its AI system verifies purchases. Members spend less time waiting, and employees have more time to help them. In clubs using the system, Sam’s Club reported that members exited 23% faster.
Marketing teams face a version of that same bottleneck. In a retail leader panel with Dr. Janet Sherlock, CEO and founder of Org.Works, Engel pointed to content approval. Many teams still rely on one person to review every asset before it goes live. As AI increases output, that person can slow the process. If reviews are rushed, work that needs a closer look may slip through.
A routine price update and a major brand campaign do not need the same level of oversight. Teams can redesign the process so routine work moves faster and people spend their attention where judgment matters most. That is how AI changes the work, rather than simply adding to it.
Make Skills and Ownership Part of the Plan
Much of AI adoption is happening inside existing jobs before anyone decides what those jobs should become. In the retail leader discussion, Engel pointed to a lifecycle specialist using AI to draft copy variations or a social team member using it to sort creative options. Those employees may become the people colleagues turn to for help before their training or authority reflects their new responsibilities.
The need to prepare teams reaches well beyond marketing. At RetailClub, Target’s Prat Vemana, chief information and product officer, and Sarah Travis, chief digital and revenue officer, spoke as retail leaders navigating AI’s impact across technology and commerce. Target’s Store Companion offers a separate example of employee AI adoption. The retailer built the assistant using store teams’ real questions and process documents, then improved it with feedback from pilot stores. Making it useful required understanding employees’ work before putting answers at their fingertips.
Bowerman described another part of that effort: embedding AI in the marketing capability plan at Kellogg and taking a similar approach across marketing and sales at Glanbia. A shared understanding and practical skills matter when employees are starting from different levels of experience and confidence.
Leaders need to notice where responsibilities have shifted, prepare people to handle them, and give them the authority to apply what they learn.
Trust Breaks Quietly
The more work AI takes on, the more small mistakes matter. At RetailClub, Ekta Chopra, Chief Technology and AI Officer at E.L.F. BEAUTY, described the work behind agentic shopping: establishing an agent’s identity, connecting it to a customer and loyalty program, enabling payments, and securing the experience. Each connection has to work for a seemingly effortless interaction to earn a customer’s confidence.
Trust also depends on what a brand says. An AI-generated message can be technically correct but still sound wrong for the brand or miss the customer’s context. Bowerman emphasized shared awareness of AI risk and governance, so teams can apply standards consistently rather than rely on one person to catch every problem.
As Engel put it, “Trust breaks quietly.” One off-brand message or frustrating interaction may seem small. Repeated across more customer touchpoints, those misses can change how people experience the brand. The work behind adoption has to protect that relationship as AI becomes part of more interactions.
Measure Progress in the Work Itself
A headline can show what a retailer launched. It rarely shows the ideas a team set aside, the process it rebuilt, or the responsibilities it had to clarify along the way.
A more useful measure is whether the organization solved a worthwhile problem and can sustain the change. Does the new workflow help employees do their jobs? Does it improve the customer experience? Is the result strong enough to build on?
RetailClub AI Festival made one thing clear: AI adoption may be messy behind the scenes, but the retailers making progress are working through that mess with purpose.
