How Important Is it For Retailers To Engage With AI Models Proactively?

In an extensive report issued by end-to-end platform and network solutions company Rithum, data points surrounding today’s consumer and their engagement with AI tools during the purchase journey were examined.

One central theme presented was that, especially as AI moves from its nascent stage in retail integration into an early form of maturity, it is vital for brands and businesses to control their own narratives by proactively engaging with LLM platforms.

“LLMs are powerful tools to attract more customers — as long as your product content, pricing, inventory data, and other details are accurate and consistent across where LLMs look. If you don’t control the data story about your products, then AI will make one up for you — whether it’s right or not,” the report authors began.

Stating that the emergence and solidification of AI in the purchasing journey represents a “new kind of ecommerce land grab,” the Rithum report suggested that logo recognition — and even customer loyalty — may be waning in overall importance compared to this new and disruptive force.

“A new brand can displace a household name when its product data gives AI better information, and the retailers losing ground often have no idea it’s happening. A brand can drop from a top citation to near-invisible in a matter of weeks when a competitor publishes a more structured product page, a Reddit
thread surfaces full of complaints, or a pricing inconsistency makes AI stop trusting the data,” the report read.

“By the time the pattern shows up in revenue data, the LLM has already formed a recommendation habit. A shopper who’s bought from an AI-recommended competitor once has now taught the algorithm what they buy — and the next recommendation builds on the last. The brand that wasn’t in the first recommendation is increasingly unlikely to appear in the ones that follow,” it added.

Other interesting data points brought forth by the report:

• LLMs appear to be creating a new product research and recommendations ecosystem beyond traditional search methods: About one-fifth (19%) of users currently buy from brands they’d never heard about previously, about 13% consider themselves more likely to switch retailers or products following engagement with an LLM, and about one-third (32%) spend less time browsing other websites when working with an LLM.
• Brand websites may be losing traction overall: Over half of shoppers (53%) polled indicated that they currently trust AI tools equally as much as brand websites. When double-checking AI data, search engines are the most common option (28%), followed by online reviews (19%), friends and family (17%), and one’s own personal experience (17%). Brand websites come in at just 5%.
• AI usage for shopping purposes roughly increases alongside income lines: The percentage of respondents who answered that they’d used an AI tool for shopping moved upward, roughly in lines with increased income. Approximately 56% of those earning less than $30,000 annually said they did so, 70% of those earning $30,000 to $49,999, 68% of those earning between $50,000 and $69,000, 77% of those earning $70,000 to $99,999, 84% of those earning $100,000 to $149,999 annually, and 81% of those earning above $150,000 per year.

“AI offers up information about your brand no matter what. The question is: Did you design your story, or did you leave it to chance? By the time a bad recommendation lands, you’ve already lost credibility,” the report concluded.

Quinnipiac, Walmart Data Suggests Trust in (Some) Agentic AI Streams May Not Be as Solid as Suggested

The report also comes as Quinnipiac released its own major study into American trust (or lack thereof) surrounding AI tools. And while that poll did not exclusively name AI-assisted shopping as a measurable activity, it did note that a majority of those polled (51%) had used an LLM to research topics of interest (which may include shopping).

However, one note of contrast between Rithum’s findings and Quinnipiacs, at least loosely: While Rithum suggested that trust in AI was rising, the Quinnipiac numbers didn’t show as strong of a case in this regard. A full 76% of Americans polled stated that businesses were not being transparent enough in their AI usage, and nearly the same cohort (74%) said that government was failing in its duty to properly regulate AI.

“Americans are not rejecting AI outright, but they are sending a warning. Too much uncertainty, too little trust, too little regulation, and too much fear about jobs,” said Dr. Chetan Jaiswal of the Quinnipiac University School of Computing and Engineering.

Further, recent Walmart data suggested that ChatGPT conversion rates as compared to in-house e-comm conversion rates are abysmal, with the blue-and-yellow brand pulling back to refocus efforts on its own Sparky agent, and a more curated, customized linking of ChatGPT interest to its own sales channel.