Agentic checkout is live. Yet almost nobody is using it

AI agents handle just 3% of transactions in the UK and US, according to a Checkout.com survey of 400 merchants and 12,005 consumers, while 89% of merchants say they are actively preparing for agentic commerce. Nearly a quarter of consumers (24%) say they will never delegate a purchase to AI, and 27% trust no organization to operate a shopping agent for them. OpenAI launched Instant Checkout in September 2025 with Etsy sellers and later moved it toward merchant apps in March 2026.

18 min read Editor’s note: The numbers in this piece come from Adobe Analytics, Checkout.com, Shopify, Salesforce, court filings, and public announcements. We’ve flagged vendor data throughout for transparency . TL;DR - Merchants in the UK and US report that AI agents are involved in roughly 3% of transactions https://www.checkout.com/newsroom/consumer-demand-for-ai-shopping-is-forming-fast-but-trust-for-agentic-commerce-is-still-catching-up today, while 89% say they are actively preparing for agentic commerce. - On the other side, nearly a quarter of consumers say they will never delegate a purchase to AI, and 27% trust no organisation to operate a shopping agent for them. - OpenAI started moving Instant Checkout https://www.digitalcommerce360.com/2026/03/06/openai-shifts-checkout-plans-agentic-commerce-strategy/ toward merchant apps in March 2026, about six months after launching it. - Meanwhile on the referral side, Adobe recorded AI-referred retail visitors converting 54% better https://www.digitalcommerce360.com/2026/06/17/adobe-ai-referred-traffic-to-retail-sites-doubles-in-a-year/ than non-AI traffic in May 2026, after converting at roughly half the rate a year earlier. - Shopify reported AI referred orders growing close to 13x year-on-year https://www.shopify.com/enterprise/blog/ai-search-insights in Q1 2026, with more than half of those sessions landing straight on a product page. - For retailers, roughly a third of the content on an average retail product page is unreadable to large language models https://business.adobe.com/blog/ai-traffic-surge-retail-sites-not-machine-readable . Table of Contents Agentic Checkout: Less Clicking, More Doing In September 2025, OpenAI put a purchase flow inside a chat window. Instant Checkout https://openai.com/index/buy-it-in-chatgpt/ launched with US Etsy sellers, and OpenAI named a million Shopify merchants including Glossier, SKIMS, Spanx and Vuori as next in line. The plumbing was the Agentic Commerce Protocol https://stripe.com/newsroom/news/stripe-openai-instant-checkout , an open standard co-developed with Stripe. Perplexity followed in November, opening its PayPal powered checkout https://www.cnbc.com/2025/11/19/perplexity-ai-online-shopping-paypal.html to free US users after a year of restricting it to Pro subscribers. Microsoft arrived at NRF in January with Copilot Checkout https://news.microsoft.com/source/2026/01/08/microsoft-propels-retail-forward-with-agentic-ai-capabilities-that-power-intelligent-automation-for-every-retail-function/ , auto enrolling Shopify merchants. Google announced the Universal Commerce Protocol https://blog.google/products/ads-commerce/agentic-commerce-ai-tools-protocol-retailers-platforms/ the same week, co-developed with Shopify, bringing purchases into AI Mode and Gemini. So, four surfaces with four payment stacks. An enormous amount of conference stage time and a great many slides with the word agentic on them. Then somebody decided to do a study. Checkout.com commissioned Census wide to survey 12,005 consumers across the UK, US, Brazil, China, France and the UAE, along with 400 heads of payment at consumer facing merchants, with fieldwork in early March 2026. Those merchants put AI agents at 3% of transactions . Eighty-nine per cent said they were preparing anyway https://www.checkout.com/newsroom/consumer-demand-for-ai-shopping-is-forming-fast-but-trust-for-agentic-commerce-is-still-catching-up . Hold those two numbers next to each other and you have the shape of the market as it currently stands: a super expensive readiness programme aimed at a sliver of volume. The consumer hesitation is predictable. 24% of respondents said they will never delegate a purchase to AI. Another 27% said there is no organisation they would trust to run a shopping agent on their behalf. What’s interesting here is that people did not really refuse the assistant itself, just the idea of handing it the card. The March 2026 Plot Twist On 5 March 2026, Digital Commerce 360 reported that OpenAI was sidelining Instant Checkout https://www.digitalcommerce360.com/2026/03/06/openai-shifts-checkout-plans-agentic-commerce-strategy/ in favour of purchases through retailer apps inside ChatGPT. A spokesperson said the company was evolving how it approaches commerce in ChatGPT. Reporting on the shift put the number of live merchants at around a dozen, and narrowed the future scope to a small set of large integrated retailers such as Target, Instacart, Expedia and Booking.com. Checkout.com, which sells into this category and has no obvious reason to talk it down, read the move as structural rather than directional https://www.checkout.com/blog/openai-agentic-commerce-shift : AI handles discovery and intent, merchants keep control of checkout. This is exactly the opposite of the story most vendors told in late 2025. Five days later, a federal judge added a second constraint. In Amazon’s suit against Perplexity, Judge Maxine Chesney granted a preliminary injunction https://www.cnbc.com/2026/03/10/amazon-wins-court-order-to-block-perplexitys-ai-shopping-agent.html blocking the Comet browser agent from password protected areas of Amazon’s site. The contention was that Comet accessed accounts with the user’s permission, but without authorisation https://www.geekwire.com/2026/judge-blocks-perplexitys-ai-bot-from-shopping-on-amazon-in-early-test-of-agentic-commerce/ from Amazon. Your customer wanting an agent to shop for them, in other words, does not by itself give that agent the right to operate on a merchant’s systems. Agent access https://industrycontents.com/ai-right-to-explanation/ is now something to be negotiated instead of taken for granted. 29 SEP 2025 OpenAI launches Instant Checkout with Etsy, on ACP with Stripe 19 NOV 2025 Perplexity opens PayPal checkout to free US users JAN 2026 Copilot Checkout at NRF. Google and Shopify announce UCP the same week 5 MAR 2026 RETREAT OpenAI moves Instant Checkout toward merchant apps. About a dozen merchants live 10 MAR 2026 CONSTRAINT Amazon wins preliminary injunction against Perplexity’s Comet 16 APR 2026 GROWTH Adobe: AI referred retail traffic up 393% year on year in Q1 20 MAY 2026 GROWTH Google expands UCP and Universal Cart at Google Marketing Live Four rails, one shelf This is why betting on a single checkout integration is a poor use of a quarter. There are at least four live standards, backed by companies that compete with each other, and none has won. | Standard | Backed by | What it actually does | |---|---|---| | UCP https://www.shopify.com/news/ai-commerce-at-scale AP2 https://cloud.google.com/blog/products/ai-machine-learning/announcing-agents-to-payments-ap2-protocol Mastercard’s position, stated in January 2026, is that it participates in all of them. That is the tell. When the networks refuse to pick, the merchants underneath them do not need to either. Every one of these standards, though, sits on top of the same requirement: a structured, accurate, machine readable product feed. OpenAI’s own documentation says merchants must supply a regularly refreshed CSV or JSON feed https://developers.openai.com/commerce/guides/key-concepts covering identifiers, descriptions, pricing, inventory, media and fulfilment before anything else works. Google is asking for the same thing through Merchant Center. Microsoft through Microsoft Merchant Center. So even though the checkout buttons differ, the shelf underneath is shared. The Money’s New Hiding Spot While the checkout story stalled, something less photogenic happened to the traffic. In March 2025, visitors arriving at US retail sites from AI assistants converted 38% worse than everyone else. By March 2026 they converted 42% better. By May 2026, per Adobe’s most recent read, 54% better, with 53% higher revenue per visit and 23% more pages viewed. Adobe bases this on more than a trillion visits https://techcrunch.com/2026/04/16/ai-traffic-to-us-retailers-rose-393-in-q1-and-its-boosting-their-revenue-too/ to US retail sites. AI referred visitors, US retail Adobe Analytics, 2026. Over 1 trillion visits. Compared against all non AI traffic in aggregate. A sign flip that complete, in twelve months, in a channel that large, does not happen often. Two buts before you quote this internally. Adobe sells an LLM Optimizer product, and this data was published alongside it. The comparison is against non-AI traffic in aggregate, which blends paid search, email, affiliate and organic, so it is not a clean head to head with Google. Which is why the second dataset matters more than the first. Shopify’s Q1 2026 commerce data https://www.shopify.com/enterprise/blog/ai-search-insights , gathered independently across a different merchant base, found AI referred shoppers converting close to 50% higher than organic search visitors with 14% higher average order values. Referral sessions from AI chatbots grew more than 8x year on year and, more importantly, orders from them grew nearly 13x. The mechanism sits in one number from that same report: more than half of AI referred sessions start directly on a product detail page, against roughly 20% for organic search. The assistant already did the comparison shopping https://industrycontents.com/ai-shopping-agents-ignore-star-ratings/ . The visitor lands pre-qualified, at the bottom of a funnel you did not have to build. Sessions starting on a product page Rather than a homepage or category page Shopify commerce data, Q1 2026. Salesforce, measuring a third population of 1.5 billion shoppers, put AI and agent influence at 20% of global retail sales https://www.salesforce.com/news/stories/2025-holiday-shopping-data/ over holiday 2025, worth around $262 billion, and found retailers running their own shopper agents grew sales 59% faster than those that did not. Three vendors, three methodologies, three commercial interests in the answer, and they agree on the same direction and magnitude. The visibility gap that needs fixing Adobe also ran its Content Visibility Checker https://www.marketingweek.com/ai-traffic-is-surging-but-many-retail-sites-arent-readable/ across the US retail sector, scoring how much of a page an LLM can actually read. Homepages averaged 75%. Category pages, 74%. Individual product pages came in at 66%. So the page most likely to convert is the page least likely to be readable by the channel converting best. The best performing retailers scored 82.5% on home-pages and the worst, 54.2%. Page content a model can read US retail averages. Higher is better. PRODUCT PAGES BY CATEGORY, MAY 2026 Adobe Content Visibility Checker, 2026. Best retailers 82.5% on homepages, worst 54.2%. The cause is usually architecture rather than laziness. A decade of headless and composable frontend work produced sites where the base HTML is a shell and everything meaningful reviews, specs, FAQs, social proof hydrates afterwards through JavaScript. Human browsers wait for it and most generative engines take the base HTML and leave. By category, Adobe’s May data put cosmetics at 63% readable, electronics at 56%, sporting goods and apparel at 51% each. The Surface Audit Framework: five steps It’s called Surface Audit because it treats every assistant as a shelf, you either appear on or do not, and because it deliberately avoids committing to any one checkout rail. Five steps, in order. None of them requires a protocol decision. 01 Score What a model can read on 20 product pages RAIL AGNOSTIC02 Feed Fix the catalogue before the frontend RAIL AGNOSTIC03 Measure Instrument the channel, state its limits RAIL AGNOSTIC04 Publish Write what gets cited, not what ranks RAIL AGNOSTIC05 Connect Cheapest checkout integration, no more RAIL SPECIFIC 1. Score what the machines can actually read A third of your product page content may be invisible to the systems doing the recommending. You cannot optimise what you have not measured. Take twenty product pages spanning your best and worst sellers. Fetch each with JavaScript disabled and read what comes back. Anything missing from that raw HTML specs trapped in images, reviews loaded by a third party widget, sizing tables rendered client side is missing from the assistant’s view of you. Score each page, record the baseline, and hand engineering a ranked list. Below is what you should check on each page: | Element | Common failure | What the assistant cannot then do | |---|---|---| | Price and availability | Injected by JavaScript after load | Include you in any query with a budget or an urgency | | Specifications | Rendered inside an image or a PDF | Match you to a query with a constraint, which is most of them | | Reviews and ratings | Third party widget in an iframe | Answer “is it any good” without leaving your page | | Variants and sizing | Loaded on click | Confirm you have the size or colour asked for | | Shipping and returns | Behind a modal or an accordion script | Compare you on the thing shoppers abandon over | | Product schema | Missing, or generated flat by a plugin | Resolve what the page is about with any confidence | 2. Fix the feed before the frontend Every standard on the table ACP, UCP, Copilot reads from a structured product feed. Feed work is the only investment here that pays regardless of which rail wins. Get identifiers, pricing, live inventory, fulfilment options and media complete and refreshed daily. Then add the attributes assistants use to disqualify you: what the product is not for, compatible accessories, materials, care requirements, return window. Google has started accepting answers to common product questions https://blog.google/products/ads-commerce/agentic-commerce-ai-tools-protocol-retailers-platforms/ as feed attributes in Merchant Center. Treat that as the direction of travel. Sequence it so the cheap work happens first: | Tier | What it covers | Effort | |---|---|---| Table stakes | GTIN or MPN, title, description, price, currency, availability, image, brand, condition | Usually already there. Audit for completeness, not existence | Disqualifiers | Materials, dimensions, compatibility, care, age suitability, what it is not for | Merchandising time. This is where most catalogues are thin | Commercial | Shipping cost and speed by region, return window, warranty, stock depth | Systems work, and the tier assistants use to break a tie | Freshness | Daily refresh minimum, real time for price and stock | Engineering. Wrong prices in a feed are worse than no feed | 3. Instrument the channel honestly If AI traffic is buried in Direct and Referral, no one in your finance meeting believes the conversion numbers above. GA4 added a native AI Assistant channel on 13 May 2026, with no configuration required. Use it, then note its three limits out loud in your reporting. It counts forward only, so your earlier trend stays buried. Google’s own AI Overviews and AI Mode clicks land in Organic Search. And a large share of assistant traffic arrives with no referrer at all, which means Direct. Keep a parallel custom channel group running, and treat every AI number as a floor. | Limit | Effect on your reporting | Workaround | |---|---|---| | Counts forward from 13 May 2026 only | No year on year comparison, and no baseline before that date | Annotate the date in GA4 and keep any custom channel group you already had running in parallel | | Google’s own AI surfaces are filed under Organic Search | AI Overviews and AI Mode clicks never appear in the AI Assistant channel | Treat the AI channel as non-Google AI only, and say so on the slide | | Referrer-less sessions land in Direct | An unknown share of assistant traffic is invisible | Report every AI figure as a floor rather than a total | Agent traffic https://industrycontents.com/context-rot-switching-ai-tools/ distorts paid media reporting in much the same way, which we looked at separately in advertising in the age of AI agents https://industrycontents.com/ai-agent-traffic-ad-spend/ . 4. Publish the content to get cited https://industrycontents.com/language-models-flatten-brand-positioning/ instead of ranked Assistants cite comparison and buying guide content at higher rates than product pages, because that is the format that answers the question being asked. Write the comparisons you have been avoiding, including the ones where a competitor wins on a specific dimension. Open each section with a declarative answer to the question in the heading rather than a wind up. Add the specificity that lets a model match you to a narrow query, price bands, use cases, who the product suits badly. Vague superlatives give a model nothing to work with. Four formats worth building first, in order of how often assistants reach for them: Head to head comparisons , including against competitors. The version where you lose on one dimension is the version that gets quoted, because it reads as evidence rather than marketing. Buying guides organised by constraint , not by product. Budget, room size, skin type, skill level, whatever your category’s real filter is. Compatibility and fit pages. Explicit lists of what works with what. These answer a query type nobody writes content for and everybody searches. Who this is wrong for. The single most citable paragraph you can write, and almost nobody publishes it. 5. Take the cheapest checkout integration available, and no more At 3% of transactions, a bespoke integration cannot earn back a quarter of engineering time. But refusing to appear at all costs you the option. If you are on Shopify, most of this is a toggle in Admin rather than a project. Microsoft auto enrols Shopify merchants https://searchengineland.com/microsoft-launches-copilot-checkout-and-brand-agents-467175 in Copilot Checkout with an opt out. If you are not on Shopify, route through PayPal or Stripe rather than building against a protocol directly. Check the economics before you switch anything on: trade press has put OpenAI’s merchant fee at 4% of completed purchases https://www.techradar.com/pro/what-openais-4-percent-checkout-fee-means-for-the-future-of-commerce , on top of normal processing, though OpenAI itself has only ever described it publicly as a small fee. Perplexity’s merchant programme has run at zero commission. Before switching anything on, get answers to these four. They decide whether the integration is worth having at all: Who is merchant of record? This determines who owns the customer data, who handles the refund, and whose name appears on the statement. What is the total take rate? Platform fee plus your normal processing. Compare it against your contribution margin per order, not against your revenue. What data comes back with the order? Email consent, marketing permission and returning customer identity are the three that determine whether this is a sale or a relationship. How do you turn it off? Auto enrolment is common. Know the exit before the entrance. Where this thesis could be wrong Microsoft reported that shopping journeys involving Copilot are 33% shorter https://www.axios.com/2026/01/08/microsoft-ai-copilot-checkout than traditional search paths, produce 53% more purchases within thirty minutes, and are 194% more likely to end in a purchase when intent is present. Copilot apps have more than 100 million monthly active users. If those figures hold across a wider merchant base, the 3% ceiling rises fast. It also assumes the agent reports what actually happened, and one study found agents confirming orders the event log never recorded https://industrycontents.com/ai-agent-journey-hallucinations/ . A third of respondents in the Checkout.com study expect at least 10% of their purchases to be AI driven within a year, and 72% of merchants think consumers will adopt faster than merchants are ready for, though stated intent has never been a reliable predictor of what people actually do. Universal Cart is rolling out across Nike, Sephora, Target, Ulta, Walmart, Wayfair and Shopify merchants including Fenty and Steve Madden, announced at Google Marketing Live 2026 https://searchengineland.com/google-expands-universal-commerce-protocol-and-launches-new-agentic-shopping-tools-478113 . Google can absorb years of unprofitable agentic commerce in a way OpenAI cannot. If checkout inside AI Mode gets genuinely good, the distribution behind it is a different order of magnitude to anything discussed above. None of which changes step one. Every one of those scenarios still requires that a model can read your product page. Your first 90 days Weeks 1 to 2. Run the readability score on twenty product pages. Confirm the GA4 AI Assistant channel is reporting and annotate the date. Establish your baseline before anyone changes anything. Weeks 3 to 6. Fix feed completeness first, server rendered product content second. Ship the two comparison pieces your category most obviously lacks. Weeks 7 to 12. Switch on whichever checkout integrations are a configuration change rather than a build. Tag those orders separately. Re-score the same twenty pages and report the delta, not the absolute. Then wait ninety days before drawing any strategic conclusion from the numbers. The channel is moving too fast for a four week read to mean anything. Related reading AI agents are confirming orders that were never placed https://industrycontents.com/ai-agent-journey-hallucinations/ , on what happens when the assistant misstates your customer’s cart. Advertising in the age of AI agents https://industrycontents.com/ai-agent-traffic-ad-spend/ , on what happens to paid media once bots account for the majority of web traffic. AI can quote you. Can it vouch for you? https://industrycontents.com/ai-mention-vs-recommendation/ , on the reputation signals that decide whether an agent recommends you at all. The brand protection stack for multi marketplace brands https://industrycontents.com/brand-protection-stack-scaling-sellers/ , on keeping listings accurate and defended once you sell in more than one place. What small sellers can do about copycat listings https://industrycontents.com/diy-ai-steps-to-stop-copycats/ , the same problem without a legal budget. FAQ Agentic checkout is a process that allows shoppers to finalise purchases directly within an AI assistant without navigating to the merchant’s website. The assistant securely collects necessary buyer and payment information, passing it to the merchant to handle fulfilment and order processing. While the merchant remains the merchant of record, you do lose the direct on-site session, which impacts your ability to perform email capture and post-purchase upsells. Because consent handling for off-site transactions is still evolving, it is recommended to consult with your legal team regarding marketing permissions. No, these protocols are designed to function together rather than in competition. They operate at different layers of the transaction stack, with ACP and UCP managing the commerce handoff while others handle authorisation and agent identity. ChatGPT has shifted toward a more focused approach, moving away from broad product listings in chats to apps built by individual retailers. The scope of the Agentic Commerce Protocol has narrowed to prioritise larger integrated retailers over the initial mass-merchant rollout. Not settled, and the first ruling went against the agent. In March 2026 a federal judge granted Amazon a preliminary injunction against Perplexity’s Comet browser, finding that it accessed accounts with the user’s permission but without authorisation from Amazon, and that Amazon was likely to succeed on claims under the Computer Fraud and Abuse Act. Perplexity has appealed. For merchants, the practical read is that agent access is becoming something platforms grant rather than something agents take.