More than one in three Americans bought a product based on an AI recommendation over the past year, according to research shared by venture capital firm a16z. Yet rising consumer interest in AI-assisted shopping has not translated into widespread trust in autonomous purchasing, and some major retailers are moving to block automated agents from their platforms.

The research, published as part of a16z's Charts of the Week series, pairs consumer survey data with network-level measurements of how automated agents are being treated across the web.

Trust Gap: Consumers Take the Advice but Keep the Checkout

Around 25% of survey respondents said they would feel comfortable letting AI buy various consumer goods on their behalf, excluding luxury products. By contrast, 42% said they would not trust AI to make purchases on their behalf.

The survey results coincide with signs that automated agents are meeting stiffer resistance online. Cloudflare, a web infrastructure provider that handles traffic for a large share of websites, reports that the share of HTTP 403 “Forbidden” responses — the status code a server returns when refusing access — affecting agents and bots has roughly doubled since the summer. The trend suggests increasing friction between automated shopping tools and websites, although the data does not establish agentic commerce as the sole cause.

Taken together, the figures expose a central challenge for agentic commerce: AI can help consumers compare products, but handing over the final purchasing decision remains difficult. Americans largely accept AI as a research assistant while stopping short of granting it spending authority.

Retailers Have Reasons to Resist

Agentic commerce allows AI systems to perform shopping tasks on behalf of consumers, potentially including product selection and checkout. Some major retailers have restricted automated agents from accessing their platforms, and their resistance has clear commercial underpinnings.

Shopping platforms depend on customer visits, advertising exposure, and direct relationships with customers — elements that autonomous agents could weaken. If consumers purchase through AI intermediaries, retailers may lose valuable opportunities to influence buying decisions.

Fraud and returns add further complications. Data from the National Retail Federation, the largest US retail trade group, shows that return rates and total return costs have roughly doubled and tripled, respectively, since 2019. Retailers may worry that automated purchasing could increase mistaken orders and disputes over transactions.

These concerns leave AI shopping facing both technical restrictions and commercial resistance. Better safeguards could help, but consumers must also trust agents to spend their money correctly.

In an X post published on October 16, 2026, a16z summarized its findings:

More than 1 in 3 Americans bought something based on an AI recommendation in the past year

44% say AI helps them find better value for money

34% would pay more for a product if AI showed it was the best option for them

Charts of the Week: pic.twitter.com/AUVnx49tGV

— a16z (@a16z) October 16, 2026

The full post is available on X: a16z (@a16z)

AI Recommends, Humans Still Check Out

Consumer behavior suggests AI currently plays a stronger role in product research than in final purchasing. According to a16z, 44% of Americans say AI helps them find better value for money, while 34% would pay more for a product if AI identified it as their best option. That willingness to rely on AI for recommendations contrasts with reluctance to authorize purchases.

Similarweb, a digital analytics firm, provides the clickstream data cited by a16z, which suggests shopping journeys increasingly include AI before checkout. However, most AI shopping sessions remain in the middle of the purchasing process, between product discovery and payment. For agentic commerce, the implication is that the technology is embedded in browsing and comparison stages, while payment remains in human hands.

The distinction matters for businesses measuring advertising effectiveness. When AI influences research but consumers complete purchases later, identifying which recommendation drove a sale becomes harder. The cited Similarweb data only runs through June, so current behavior may differ.

Social Commerce Shows the Limits of Automation

Social commerce presents another challenge for autonomous shopping. Data from Marketplace Pulse, an e-commerce research firm, shows that the top 1% of TikTok Shop sellers generate roughly 60% of sales, while the 8.4% of sellers producing at least 1,000 videos capture around 60% of overall sales.

These figures show how human-facing content can drive commerce at scale. AI agents may streamline product comparisons, but replacing discovery, persuasion, and personal preferences presents a different challenge.

A Gradual Transition, Not Overnight Automation

For now, the evidence points to a gradual transition rather than fully automated shopping. AI recommendations are gaining consumer acceptance, but retailer restrictions and lingering concerns about purchasing autonomy continue to limit agentic commerce. Signals worth tracking from here include whether Cloudflare's-blocking measurements continue to climb, whether future a16z survey waves show more Americans willing to delegate purchases, and whether updated Similarweb clickstream data — which currently runs only through June — shows AI sessions moving closer to the point of payment.