Consumers are increasingly comfortable using AI agents for product recommendations and price comparisons, but remain hesitant to let the technology complete purchases on their behalf, creating an opportunity for brands to improve product data and build consumer trust, according to recent Circana data.
The findings come after multinationals like Mars predict auto-replenishment will become a cornerstone of ecommerce but brick-and-mortar will remain key for discovery and trial. Yet, consumers are looking to agentic commerce to simplify decisions rather than automate them, according to Sally Lyons Wyatt, Circana’s Global Chief Advisor, Consumer Goods & Foodservice Insights.
“People are happy to let AI handle the hassle, just not the moment the money leaves their wallet,” she said.
While 34% of shoppers are influenced by AI to make decisions and more than half (53%) are comfortable using it to compare prices, only 16% of consumers trust AI agents to purchase products without their approval, per Circana.
Better product data builds trust
Building trust in AI agents requires updating data, according to Kiara Barrett, Circana’s Global Head of Thought Leadership.
“Many AI trust problems are actually product data, pricing and inventory problems,” she said.
Agentic commerce relies on more nuanced information compared to traditional SEO searches that rely on keywords to lift brands to the top of a query. For brands to show up as the answer in an AI search means thinking like an AI agent.
This means providing product details around pricing, nutrient claims, ingredients, packaging materials and other explicit details that answer specific shopper questions.
“First and foremost, fix your data accuracy. Get it right. Make sure you’ve got the right price. You’ve got the right inventory availability,” Barrett said, adding that stale prices, expired sales or out-of-stock inventory contributes to a negative shopping experience.
“If the agent sends someone to a store for a sale that ended yesterday, the experience is simply wrong,” she emphasized.
Identifying which data points to include also means understanding what consumers are searching for within a brand’s respective category.
“Those are the types of language that people are typing into that box, and that’s the type of language that brands and retailers need to be embedding into the products,” she said.
Why transparency matters in AI recommendations
Generic generative AI tools (like ChatGPT or Gemini) garner more consumer trust than retailer-site AI assistants because “consumers may suspect that the retailer assistant is optimizing for what the retailer wants to sell, rather than what is generally best for them,” Barrett explained.
She adds: “The trusted retailer does not automatically create a trusted agent, and the agent has to earn trust based on its own behavior.”
Barrett recommends that retailers explain why a product is recommended, differentiate sponsored placements, offer alternative options and disclose when recommendations are limited due to their own assortment.
“Cite your sources, show your receipts,” she said.
In the future, Barrett notes that future autonomous replenishment adoption “will be earned task by task, not adopted as one giant leap.”
