Organizations continue to tout the efficiencies and cost savings they’ve enjoyed after implementing AI. What they tend not to discuss, however, is how consumers feel about AI taking over aspects of the customer experience.

That might be because customers are less enthusiastic about AI than businesses are. For instance, a recent Metrigy study found that 80% of the 503 consumers surveyed preferred interacting with a human over an AI agent.

“No customer should live through a brand’s efficiency gains,” said Elizabeth Garry, VP, Ecommerce at jewelry brand Pandora, while addressing tradeoffs between brand experience and costs during a session at June’s CommerceNext Growth Show in New York City. During the panel discussion “Are Robots Stealing Your Brand Soul? Keep It Human in the Age of Agentic,” Garry and other executives explained how they worked to optimize AI without destroying their brands’ unique qualities and human elements.

Begin With the Brand, Not the Tech

Authentic Brands Group, whose diverse portfolio includes Brooks Brothers, Juicy Couture and Reebok, builds and maintains a distinct knowledge base for each brand, said Adnan Somani, EVP, Digital Technology. While each knowledge base includes everything from history to emoji usage for the particular brand, humans are still responsible for fine-tuning any outputs, whether they’re customer care emails or imagery.

This goes hand in hand with “identifying what is your soul as a brand,” said Oscar Diaz, SVP, Technology & Digital Innovation at G-III Apparel Group, parent company of DKNY and Wilsons Leather, among others.

Only once you’ve nailed down what consumers want from your brand and why customers continue to buy from it can you introduce AI to enhance rather than override what makes the brand unique. “Over-engineering for the sake of commodity or cost-optimization devalues your brand,” Diaz added. “And that’s where the risk is.”

Implementing AI without first experimenting and iterating presents another kind of risk. “The people who put one baby toe into the water are the ones who end up with AI content that looks the same as everyone else’s,” said Brian Schiff, Co-Founder and CEO of customer support technology provider Flip. “The ones who really lean in and learn it and spend the time to play with it, understand it, tweak it, experiment, those are the ones who are able to use the power it has to create amazing experiences.”

Balancing Efficiency and Empathy

Applying AI judiciously, rather than simply automating every possible process and interaction, is critical to maintaining a brand’s heart and soul. At Authentic Brands, “we like the AI taking it 90% of the way there,” Somani said, “and the human takes it to 100% and presents something they’re really proud of.”

When working with clients to automate customer service with AI, Flip encourages the marketing and customer experience teams to determine ahead of time which topics and scenarios can and cannot be delegated to AI. They’ll also settle on the parameters for when to escalate from machine to human agent.

Using AI to implement these rules can provide brands with unprecedented scalability and flexibility, Schiff said; while LLM agents handle routine “Where is my order?” and “How do I reset my password?” queries, humans can focus on frustrated customers with additional concerns, balancing efficiency with empathy.

Seeking a similar balance with its creative, Pandora refuses to use AI-driven model imagery. “We believe that psychologically it’s really important for someone to look into the eyes of someone else and see them wearing our pieces,” Garry said.

However, the brand is testing AI to create 360-degree views and other product imagery: “I think it’s a really interesting example of how AI can complement something that should remain intimate and human while amplifying the usability, the styling capabilities that would be really expensive and really tedious to produce.”

Both instances show the importance of prioritizing connection and trust to ensure AI doesn’t erode a brand’s soul. “Anything that’s going to break your customer’s trust is something we can’t automate,” Diaz said. “AI can scale things, but humans can give you accountability. I think that’s the connectivity people want as consumers. Once you break that trust, it’s hard to gain back.”