Salesforce survey finds shoppers trust AI advice but resist automated purchases
Retailer-run agents are associated with bigger orders and stronger loyalty, but shoppers’ concerns about privacy and unwanted charges limit the handoff.
Listen to this story
The audio brief
Story brief
3 key pointsSalesforce’s Nov. 12 survey of more than 2,500 shoppers in the U.S., U.K., Canada and Australia finds personal AI agents gaining influence as product researchers, while shoppers remain reluctant to delegate payment. Although 74% trust AI recommendations, only 19% of personal-agent users had tried one at checkout and 9% granted unrestricted access to financial details. Branded agents may offer retailers a commercial advantage, but Salesforce’s spending comparisons are associations, not proof of cause; accurate product, price and inventory data are the practical requirement.
- 01
Thirty-one percent of shoppers had tried a personal agent; among those users, 28% assign tasks multiple times a day.
- 02
Twenty-seven percent of personal-agent users say an agent bought something too expensive, while 22% report a blocked purchase; only 32% report no major errors.
- 03
Thirty-nine percent of shoppers had used a retailer’s agent; 80% of those users reported a good experience, and 75% of satisfied users felt more loyal.
AI is earning a place in shoppers’ decisions without earning control of their wallets. Salesforce’s holiday-shopping research, published November 12, finds that almost one in five shoppers starts with AI before going to a store. Yet 76% aren’t interested in letting a personal AI agent make purchases for them—a sharp boundary between advice and authority.
The new Salesforce Consumer Shopping survey covers more than 2,500 shoppers in the United States, United Kingdom, Canada and Australia. The release combines their responses with the company’s commerce metrics and holiday forecasts. Those are different kinds of evidence: reported attitudes and experiences, observed spending comparisons, and predictions about purchases still ahead.
An adviser before a buyer
Personal agents—the assistants shoppers give tasks to—are mainly being used to narrow choices. Thirty-one percent of consumers surveyed had tried one. Among personal-agent users, 28% assign tasks multiple times a day. The most common shopping jobs involve gathering and interpreting information rather than committing money:
- Compare prices across products.
- Analyze specifications and reviews for differences such as sizing quirks or technical details.
- Discover products, including brainstorming gifts that meet a shopper’s criteria.
That advisory role can shape which brands reach consideration. Salesforce found that 74% of shoppers trust AI recommendations, while 30% of AI shoppers say the tools frequently introduce unfamiliar brands. The influence also runs in reverse: 11% of shoppers say they would skip a purchase if an AI assistant advised against it.
Of personal AI agent users surveyed, 19% had deployed an agent at checkout.
Of personal AI agent users surveyed, 9% gave agents unrestricted access to their financial details.
The risks shoppers want to contain
The biggest barriers to letting agents buy are unwanted sharing of purchase history or personal data, and accidental or unwanted charges. Some users are separating the agent’s spending from their other finances: Salesforce says 36% of personal-agent users have set up a separate credit card for their agent.
Mistakes are already part of users’ experience. Twenty-seven percent of personal-agent users say their agent bought something too expensive, and 22% say it was blocked from completing a purchase. Only 32% report no major errors. Salesforce also says most shoppers would abandon an AI assistant after two or three mistakes.
Retailers’ agents have a different role
The research separately examines agents offered by brands on their ecommerce sites. Thirty-nine percent of shoppers had used one. Of those users, 80% reported a good experience; among that satisfied group, 75% felt more loyal to the brand afterward. That loyalty figure applies to satisfied users, not everyone who encountered a retailer’s agent.
Salesforce reports that shoppers engaging with a brand’s agent had an 8% higher average order value and bought 23% more items than the average order. These are spending associations, not proof that the agents caused larger purchases. The company also says shoppers arriving through AI referrals were more than twice as likely to buy as typical ecommerce shoppers.
An earlier Salesforce holiday forecast, published July 20, provides historical context rather than independent confirmation. It put 2025 holiday sales growth at 6.2% for retailers using branded shopping agents, versus 3.9% for those without them. That comparison concerns last year’s sales, not results from the approaching holiday season.
For this season, Salesforce projects AI will drive $75.8 billion in Cyber Week spending, up 13% from 2025. Its practical advice to retailers focuses on the information agents consult: accurate product data, current prices and inventory, and retailer-owned agents capable of answering shoppers’ detailed questions.
Sources
- salesforce.com2026 Holiday Predictions: A Unified Yet Fractured Journey
- salesforce.comNew Data: Holiday Shoppers Will Rely on Agents. Retailers Must Be Ready.
Reader comments
Newest comments first. Replies stay oldest first.