How agentic commerce will test loyalty and returns

Who pushed the buy button? It’s an interesting question here in the early days of agentic commerce, especially as it pertains to returns and fraud.

Retailers need to prepare for scenarios of artificial intelligence agents mistakenly buying for their users, increasing impulse buys or misleading consumers when researching and buying items. These aren’t typical situations yet, but as consumers increase adoption of AI and use agents to shop for them, they could become common complaints at the returns desk.

An IAB study found 80% of surveyed consumers expect to use AI more in their shopping, with nearly 9 in 10 saying AI recommendations feel personalized and make shopping more enjoyable. Of course, trust remains an issue, as the majority of consumers in the survey are double-checking AI results and doing further research into recommendations.

Nevertheless, a symbiotic AI agent and consumer relationship is on the horizon. How will returns and consumer loyalty be put to the test?

Policies default to ‘you are your bot’

In response to AI buying features, retailers are starting to draw a line in the sand on returns, saying shoppers are responsible for their AI agents. It’s a move that makes sense for retailers — reminiscent of the old retail adage, “you break it, you buy it.”

Consider a consumer asking a tool to find an “absorbent, low-cost diaper that prevents rash.” They receive a few options, review the product breakdown, and then empower an AI agent to purchase and have a store deliver a bulk box of those diapers. Later, if an agent automatically reorders that diaper brand without a consumer realizing it, the consumer will have a tough time returning it, and loyalty is at risk.

The reality is, for retailers, whether an AI agent or a human finger pushed the buy button, the purchase looks the same to their systems. If a consumer connects an AI browser plug-in to a retailer’s account and tells it to buy diapers, the retailer sees a purchase from that consumer’s account, on their device, from that consumer’s IP address. Currently, there’s no easy way to distinguish agent-driven purchases from manual ones.

Consumer loyalty and returns get complicated

Agentic commerce and consumers increasingly relying on AI-driven recommendations present a new layer of complication for retailers to contend with. In returns, retailers already deal with consumers feeling deceived by a product photo or vague product description; hence, returning the item. 

What if consumers now claim the AI chat engine provided false details? What if they say they didn’t tell an AI bot to buy a product? What if they direct AI agents to buy multiple colors and sizes just so they can try more at home and return later?

Retailers saw a version of these nuanced situations when one-click checkout arrived: impulse purchases went up, and so did returns. Agentic commerce could have the same effect, and that’s a problem. AI could remove that very human moment of hesitation for a consumer, whether they want to buy or not buy, and decide for them. Returns will go up, but retailers will push back, and consumer loyalty could be tested.

AI protects retailers from AI 

Agentic purchasing decisions are a new challenge for retailers, but at the same time it’s nothing new, as retailers have been adjusting to AI since its inception. Retail crime groups have been using AI to scan return policies for loopholes in returns and using generative AI to doctor receipts or photos of product to falsely claim it was damaged upon delivery.

The continued response from retailers is to double down on their own AI tools. AI can help retailers spot patterns across millions of transactions, such as a spike in returns tied to a specific referral source or a cluster of purchases linked to one identity hiding behind several logins. Unusual patterns get detected automatically, such as a consumer with a high rate of buying, trying items and returning them.

Retailers with transaction data and shopper profile insights, across in-store, online and customer service centers, unified and visible across departments, can use AI to detect fraud and abuse while also using AI to provide improved CX and make better decisions. They can support store operations teams to stop returns at the counter. Digital teams can stop claims and refunds at the call center. They can support loss prevention teams by digging deep into ORC patterns and reducing criminal activity.

It’s true that retailers may not currently have a clear way to tell which purchases came from an AI agent or a consumer, but visible, connected data strategies and AI can support retailers in reducing false returns claims, identifying abusive behaviors and protecting loyalty.

Vishal Patel is chief product and AI officer at Appriss Retail.