AI is heading into its biggest holiday shopping season yet, and as brands prepare for holiday shopping, AI assistants are creating another path to discovery. Consumers can describe what they need and receive product recommendations without starting with a brand name, says Amperity.
That raises a question beyond how to appear in an AI-generated answer: Once AI introduces a shopper to a brand, what makes that shopper return?
In an interview with Retail Insider, Derek Slager, Co-Founder and Co-CEO of Amperity, talked about why brands’ response to AI discovery needs to extend beyond search optimization. His perspective: Getting recommended creates an opportunity, but converting that introduction into a lasting relationship depends on how well a brand understands and serves the customer after the first purchase.
How could AI-powered product recommendations change the way consumers discover brands during the holiday shopping season, and what does this mean for traditional search and brand familiarity?
AI assistants give shoppers another way to describe what they need and evaluate products, including products from brands they have never bought before. For holiday shopping, that might be a gift for someone with a particular interest, within a specific budget. The recommendation can begin with that need rather than a familiar brand name.
Traditional search and brand familiarity still matter. Retailers should also make product information accurate, consistent, and useful wherever shoppers encounter it. Clear specifications and availability help someone decide whether a product fits their needs, regardless of how they found it.
The harder problem begins when the shopper reaches the retailer. An assistant’s recommendation does not mean the retailer receives the conversation that produced it.
The retailer has to work from the context it can actually access and is permitted to use. It needs to recognize the customer where possible, understand what the current interaction tells it, and decide how to respond. Discovery creates the opportunity. What happens next determines whether the customer has a reason to return.
Amperity’s research found that 60 per cent of consumers said AI led them to choose a brand they hadn’t previously considered. What does this suggest about how AI is influencing consumer purchasing decisions and brand loyalty?
It suggests AI can change which brands make it into the consideration set. It does not establish that an AI recommendation creates loyalty. Those are different outcomes, and retailers need to measure them separately.
Someone may try an unfamiliar brand because a recommendation fits an immediate need. Whether they buy again depends on the product and the experience that follows.
If the next interaction ignores what they just bought, the retailer has learned very little from the introduction.
That distinction matters operationally. Teams should look beyond the initial conversion to repeat purchase, retention, and customer value over time. Where referral information is available, they can compare customer cohorts while accounting for differences in promotions and purchase timing. A spike in first orders alone does not tell you whether the relationship is getting stronger.
When AI introduces a shopper to a brand, what customer data and insights are most important for turning that initial purchase into a lasting relationship?
Start with identity: can the retailer reliably connect this interaction to the right customer?
Someone arriving through an AI assistant may already shop in the retailer’s stores or belong to its loyalty program. A new route to the brand does not necessarily mean a new customer.
Once that connection is established, purchase history, loyalty activity, stated preferences, recent behavior, and permissions give the next decision a foundation. History tells you something about the relationship. Current signals tell you what may have changed. They need to be interpreted together.
A winter jacket purchase illustrates the limit of the transaction alone. It could be a gift, a replacement, or something the customer bought for themselves. Additional evidence can help, but an inference should not become a fact simply because an AI system generated it confidently.
The useful question is whether the context is sufficient for the decision you are about to make. You may have enough evidence to stop promoting the jacket someone just bought, while still lacking enough information to infer their personal style.
How can retailers use purchase history, customer preferences and other data to make product recommendations more relevant without overwhelming consumers or compromising their trust?
A relevant recommendation can still be a poor customer decision if the timing, frequency, or permissions are wrong. Those controls need to be part of the workflow that delivers the offer.
Purchase history can help suppress a recently purchased item or suggest a complementary product. Recent browsing and stated preferences can change that interpretation. But those signals have to reach the systems making the next decision. If a purchase is recorded in one system while another keeps sending the same promotion, the customer experiences the disconnect.
Permissions and channel preferences need to travel with the context. Teams should set clear rules for eligible data, contact frequency, and which actions require human review. Customers should understand how their information is used and have meaningful ways to change their preferences.
Teams also need to inspect why an offer was selected and stop or correct the workflow when the evidence is wrong. That is how trust becomes something the system can enforce and the team can verify.
With the holiday shopping season approaching, what should retailers prioritize to turn AI-driven discovery into repeat purchases and long-term customer relationships?
I would start with the path from a completed purchase to the next customer interaction. Can the retailer recognize the person, make the purchase available to the relevant systems, and change the next message before it goes out? That is a practical test teams can run before the holiday rush.
A trusted customer profile is essential, but the work cannot stop there. If acting on a new signal requires an analyst to assemble data, a marketer to interpret it, and an operator to rebuild an audience, the opportunity may pass before the response is ready.
Retailers should identify those handoffs and establish approved workflows for common post-purchase decisions. Gift purchases deserve particular care. The buyer may not be the end user, so the follow-up should reflect what the retailer knows without inventing preferences for either person.
Then measure whether the changed experience earns another purchase. Compare repeat purchase and retention against a credible baseline, accounting for promotions and seasonality. Engagement can help diagnose what happened, but more messages or clicks alone do not demonstrate a stronger relationship.
The holiday priority is to close the distance between learning something useful about a customer and putting an appropriate response into action, with people accountable for the decision.
More from Retail Insider:
