TL;DR
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
PYMNTS Intelligence reports that Amazon accounted for 59% of consumers’ most recent AI-assisted purchases, while it represented 29% of online retail research starts. The figures come from separate distributions, not a tracked shopper funnel, but point to a gap between AI-led product discovery and where shoppers complete purchases.
Amazon accounted for 59% of consumers’ most recent AI-assisted purchases, according to new PYMNTS Intelligence data, suggesting that AI tools may be changing how shoppers find products faster than where they buy them. The report compares that figure with Amazon’s 29% share of online retail research starts and 49% share of consumers’ main purchase completions; the figures are separate distributions, not a tracked path from search to checkout.
The findings appear in PYMNTS Intelligence’s report, Will the 2026 Shopping Season Go Agentic?, published October 30. The report says roughly 132 million U.S. adults have bought something with AI assistance, while 58.4 million now start online retail research using an AI tool. Separately, 57.5 million AI-assisted shoppers said they use AI to research products they previously searched for on Amazon.
But product discovery does not necessarily lead to a different checkout destination. PYMNTS reports that 54.5 million consumers still buy on Amazon even when an AI tool found the product. That figure and the 59% purchase share describe different measures; the report does not provide a shopper-level conversion rate linking AI research directly to Amazon purchases.
The report also finds that consumers’ willingness to give an AI agent authority to both shop and buy has fallen. The share saying they would allow that declined from 32% in September 2025 to 24% in August 2026. Only 36.8 million consumers would allow an AI application to hold their payment card, according to the report. These results suggest that consumers may accept AI for research while remaining more cautious about delegating payment and purchasing decisions.
AI Finds Products, Amazon Closes Sales
The numbers highlight a potential divide in online shopping: AI may increasingly influence discovery, while established retailers retain advantages at checkout. Amazon already has shoppers’ stored payment credentials, delivery preferences, customer accounts and a familiar returns process. Those existing services can make completing a purchase there simpler than moving from an AI recommendation to a retailer unfamiliar to the shopper. This is an explanation offered by the report, not a measured breakdown of why each consumer chose Amazon.
That distinction matters to retailers, AI platform operators and payments companies seeking a role in agentic commerce. If shoppers use an AI assistant to compare products but return to an established retailer to pay, the assistant may shape product consideration without controlling the transaction. Checkout access, authentication and consumer trust could prove as important as the quality of an AI recommendation.
Merchant readiness is another part of the challenge. In the PYMNTS Intelligence survey, 87% of merchants viewed agentic commerce as an opportunity, and 90% had begun work on products and catalogs. Businesses reported spending an average of 2% of annual revenue preparing for it, while only 23% said they were very or fully prepared to integrate with AI platforms or agent networks. The gap suggests that interest and investment have not yet translated into broad operational readiness.
Products Worth Considering
Ledger Stax Crypto Wallet with Touchscreen
Ledger Flex + Folio Crypto Hardware Wallet
Ledger Nano X Bluetooth Crypto Wallet
As an affiliate, we earn on qualifying purchases.
Consumers Delegate in Stages
The report describes consumers’ approach to AI shopping as a “ladder” of delegation, rather than an all-or-nothing choice. A shopper might ask an AI tool to find options, compare features or recommend a product, but still decide which item to buy and complete payment personally. Giving an agent permission to execute the purchase involves a different level of trust and financial authority.
That distinction helps put the adoption figures in perspective. While 132 million U.S. adults had made an AI-assisted purchase, the report says 62.9 million would let an agent both shop and buy on their behalf, and just 36.8 million would allow an AI application to hold their card. Using AI during a purchase is not the same as handing it control over the transaction.
The report’s comparisons between Amazon research, purchase completion and AI-assisted purchases are useful indicators of changing behavior, but they do not establish that individual consumers moved from an AI search directly to an Amazon checkout. The data show a contrast across reported measures, not a verified end-to-end shopping journey.
“The report describes consumers’ willingness to delegate shopping tasks as a “ladder” rather than a switch.”
— PYMNTS Intelligence report, “Will the 2026 Shopping Season Go Agentic?”
The Data Does Not Track Checkout Paths
The report’s 29%, 49% and 59% figures come from separate distributions. They do not show that a particular group of consumers began research with AI and then completed a purchase on Amazon. The data therefore indicate a difference between reported research and purchase measures, but do not quantify how often AI recommendations result in Amazon sales.
The published material also does not establish which product categories account for the AI-assisted purchases, how Amazon’s share compares with individual rival retailers, or how much of the reported behavior reflects repeat customers and existing account features. Merchant survey responses describe stated preparation and outlook, not the eventual performance of agentic commerce.
It remains unclear how quickly consumers will become comfortable authorizing AI agents to pay, which safeguards they will expect, and whether AI platforms or retailers will control those permissions. The report’s figures capture reported behavior and attitudes up to August 2026; they do not settle how shopping patterns will change during the 2026 holiday season.
Holiday Season Tests Agentic Shopping
The report frames the 2026 shopping season as a test of whether AI tools can move beyond helping consumers find and compare products to participating more directly in purchases. The figures released so far show substantial use of AI in shopping, alongside lower willingness to let an agent both shop and buy or hold a payment card.
Retailers, AI platforms and payment providers will need to address how an agent’s authority is set, verified and limited if shoppers are to delegate more of checkout. The report points to controls such as spending caps, approved merchants and requests for shopper authorization when a purchase falls outside preset rules. It does not identify a common standard or announce a specific industry rollout.
Further evidence will be needed to determine whether AI-assisted discovery changes where consumers ultimately purchase, and whether merchants’ current preparations lead to integrations shoppers use. How many consumers authorize AI to complete a transaction, rather than only help with research, remains a key measure to watch.
Key Questions
What does the 59% figure mean?
PYMNTS Intelligence says Amazon accounted for 59% of consumers’ most recent AI-assisted purchases. The report does not describe this as Amazon’s share of all online shopping.
Does the report show shoppers moving from AI research to Amazon checkout?
No. The research-start, purchase-completion and AI-assisted-purchase figures come from separate distributions, not a tracked funnel following individual shoppers from discovery to checkout.
How many consumers are using AI to shop?
The report estimates that roughly 132 million U.S. adults have bought something with AI assistance. It also says 58.4 million now begin online retail research using an AI tool.
Are consumers willing to let AI agents make purchases?
Willingness appears more limited than general AI-assisted shopping. The share who said they would let an agent both shop and buy fell from 32% in September 2025 to 24% in August 2026, according to PYMNTS Intelligence.
What is still unknown about the findings?
The report does not establish how often an AI recommendation directly leads to an Amazon purchase, or how behavior varies by product category. It also does not show whether reported merchant preparations will lead to widespread agent integrations.
Source: pymnts.com
