Something fundamental is shifting in how people buy things online, and it is moving faster than most businesses have had time to notice.

AI agents — software systems capable of reasoning, planning, and acting autonomously — are beginning to handle the shopping journey on behalf of users. Not just recommending products, but searching, comparing options across merchants, and evaluating shipping terms and return policies. Autonomous checkout is where this is heading. But the shift that is already reshaping commerce is happening in what gets considered, what gets compared, and what gets ignored.

For brand owners, marketing directors, and ecommerce leads, the question is no longer whether agentic commerce will affect your business. It is whether your business will be visible, selectable, and executable when AI agents become a significant share of your customer base.

This guide is written for decision-makers who need to understand not just what agentic commerce is, but what it means for their specific business  and what to do about it.

What Is Agentic Commerce?

Agentic Commerce is when AI handles the shopping journey on your behalf. That include searching, comparing, and evaluating options so you don’t have to.

Take a simple example. Instead of spending hours researching a smartphone, a user says or type:

“Find me a phone with a sharp camera under $150 and fast delivery.”

The AI scans websites, reads reviews, compares prices, checks delivery timelines, and returns a recommendation (or in some cases, completes the purchase outright)

Today, AI is mostly deciding. The part where it buys without you is still being built with the emerging protocols such as UCP.

This is what separates agentic commerce from traditional e-commerce. Before, people used to navigate the internet themselves. Now AI is beginning to do it for them.

And it goes beyond shopping. It includes hotel bookings, subscriptions, grocery reorders, appointment scheduling or  any repetitive commercial task.

How Big Is This New Industry?

For any decision-maker evaluating how much attention and resource to give this topic, the numbers are the most honest place to start.

McKinsey’s research projects the US B2C retail market alone could see up to $1 trillion in revenue orchestrated through agentic commerce by 2030, with global projections reaching $3 to $5 trillion. To put that in context, total global ecommerce revenue is forecast to reach $6.88 trillion in 2026 according to Shopify’s Global Ecommerce Sales Growth Report. Agentic commerce is not a niche within ecommerce — it is on a trajectory to become a defining share of it.

The consumer adoption signals are already visible. Adobe Digital Insights reported that AI-driven visits to US retail sites grew 393% year-over-year in Q1 2026, and that those visitors converted 42% better than shoppers arriving through traditional channels. That conversion premium matters enormously because it suggests that users arriving via AI agents arrive with clearer intent and higher purchase readiness than the average browser-based shopper.

A 2026 IBM Institute for Business Value and National Retail Federation study surveying over 18,000 consumers across 23 countries found that 45% of consumers already use AI in some part of their buying journey including researching products, interpreting reviews, and hunting for deals. That figure is not measuring fringe early adopters. It is measuring nearly half of all consumers.

The trust barrier remains the most significant friction point. Juniper Research’s April 2026 study identified trust as the number one barrier to agentic commerce deployment, ahead of all technical concerns. Consumers are willing to let AI assist their decisions. They are more cautious about letting AI complete transactions entirely on their behalf. That caution is rational, and the industry — from Visa and Mastercard to Google and OpenAI — is actively building the authentication and accountability infrastructure designed to address it.

Who Is Building Agentic Commerce and How Fast

Agentic commerce is being built by the largest technology companies, payment networks, and ecommerce platforms in the world — and much of it is already live, but mostly in experiments stages.

Google launched its Universal Commerce Protocol in January 2026, co-developed with Shopify, Target, Wayfair, and backed by Home Depot, Lowe’s, Best Buy, Visa, and Mastercard. The protocol is designed to standardize how AI agents interact with merchants through checkout, fulfillment, and order management — making Google’s AI surfaces, including AI Mode in Search and the Gemini app, capable of completing purchases directly within the conversation interface. This is not a roadmap item. It is in active pilot with US merchants today.

OpenAI’s journey is more instructive than a straightforward success story. It launched Instant Checkout in September 2025, positioning ChatGPT as a native purchasing destination powered by the Agentic Commerce Protocol co-developed with Stripe. Within weeks, the feature was pulled back. Users were engaging with product discovery inside ChatGPT enthusiastically, but not converting to purchases. The infrastructure proved harder to build than anticipated — merchant onboarding stalled, product data accuracy was inconsistent, and a sales tax compliance gap had gone unaddressed.

OpenAI has since refocused on product discovery and search within ChatGPT, routing actual transactions to connected retailer apps rather than completing them natively. The ACP protocol itself remains active and in development. The lesson for decision-makers is an important one: even with 800 to 900 million weekly active users and the resources of one of the world’s most capitalised AI companies, getting agentic commerce right at the transaction layer is genuinely hard. Consumer willingness to discover through AI is ahead of consumer willingness to transact through AI — and the gap between those two behaviours is where the real infrastructure work is happening right now.

On the payments side, Visa launched its Intelligent Commerce program, building token systems that allow AI agents to transact securely on behalf of users with pre-authorized spending limits and verifiable identity trails. Mastercard launched its Agentic Payments Program in parallel, working with Microsoft and other AI platforms to ensure that payments made within AI environments are safe, transparent, and accountable at every stage. Both networks understand that their role in agentic commerce is not just processing transactions — it is providing the trust infrastructure that makes autonomous purchasing acceptable to consumers and regulators alike.

Anthropic published the Model Context Protocol, an open standard that allows AI agents to share context, memory, and intent across different tools and platforms — effectively giving agents the ability to carry what they know about a user from one environment to another without losing continuity.

Shopify, meanwhile, has been building agentic commerce capabilities directly into its merchant infrastructure, positioning itself as the commerce layer that connects AI surfaces to merchant inventory and checkout systems. For the millions of merchants on Shopify’s platform, some of this groundwork is being laid on their behalf — though the degree to which individual merchants are discoverable and selectable within these systems still depends heavily on their own product data quality and configuration.

The pace matters here. This is not a five-year horizon. The foundational protocols, payment rails, and AI surfaces are operational now. What is still being built is the scale, the consumer trust layer, and the merchant readiness on the other side of the equation.

What Agentic Commerce Means for Your Business Specifically

The implications of agentic commerce differ depending on where you sit in your organisation. Brand and marketing teams face a different set of challenges to ecommerce and infrastructure teams, though both ultimately need to solve for the same outcome — being visible, selectable, and executable when AI agents become a meaningful share of your customer base.

For brand owners and marketing directors

The most immediate shift is in how discoverability works. AI agents do not browse your website. They query structured data directly. That means your homepage, your brand photography, your editorial content, and your carefully crafted product descriptions written for human attention spans play no role in whether an agent selects your product. What matters is whether your product data is accurate, complete, structured, and accessible in a format machines can parse.

This changes the SEO conversation significantly. Traditional SEO optimized for search engine rankings and human click-through rates. Agentic commerce requires a parallel discipline — Generative Engine Optimization — focused on making your products legible to AI recommendation systems. Product titles need to be specific and attribute-rich. Descriptions need to be structured, not persuasive. Policies around returns, shipping, and availability need to be clear and machine-readable, because those are the fields agents evaluate when comparing your offer against a competitor’s.

The customer ownership question also deserves serious attention. Google’s UCP and OpenAI’s ACP both affirm that merchants remain the seller of record and retain their customer data. But as AI surfaces handle more of the discovery and evaluation journey, the brand touchpoints a customer experiences before purchase will increasingly belong to the platform, not to you. Building owned audience assets — email lists, loyalty programmes, direct channels — becomes more strategically important as AI ecosystems mediate more of the top of the funnel.

Advertising attribution will also need to evolve. When an AI agent influences discovery, evaluation, and purchase across a single conversation, traditional last-click models cannot accurately attribute the sale. Marketing teams that adapt their measurement frameworks now — building towards multi-touch and intent-signal-based attribution — will have cleaner data when agentic commerce scales.

For ecommerce leads and heads of digital infrastructure

The technical mandate is more concrete. According to Stripe’s technical field guide for agentic commerce, agents query product catalogs, check inventory, compare pricing, and trigger checkout through API calls — and they expect responses in under 200 milliseconds. If your infrastructure is built for human browsing patterns rather than machine query volumes, it will fail the performance threshold agents require, and your products will be skipped.

The starting point is product data. Shopify’s executive guide on agentic commerce is direct on this point: ensure key fields — title, description, price, inventory, weight, variants, shipping options, return policy — live in structured, machine-readable fields, not embedded in marketing copy or page layouts. A product whose specifications are described in a long-form paragraph rather than discrete structured attributes is effectively invisible to an agent evaluating it programmatically.

After product data, the next priority is API accessibility. Both UCP and ACP require merchants to expose commerce APIs — endpoints that agents can call directly to query your catalog, check real-time inventory, retrieve shipping timelines, and trigger checkout flows. If you are on Shopify, much of this infrastructure is being built on your behalf. If you operate on a custom stack, this is an active engineering workstream that needs to be scoped and resourced now.

Webhook configuration matters too. Google’s UCP requires webhooks that push order status updates back to the platform in real time. Without these, your integration review will not be completed and you will not be confirmed as live on Google’s AI surfaces.

How to Prepare for Agentic Commerce With Actual Steps

The good news for decision-makers is that preparing for agentic commerce does not require rebuilding your entire technology stack. It requires getting foundational things right in a specific order, with clarity about what is urgent now and what can be phased.

Step 1: Fix your product data before anything else

This is the step most businesses underestimate and the one that determines everything downstream. Shopify’s agentic commerce guide is explicit: AI agents rely on structured data. Product information — title, price, material, dimensions, availability, variants, return policy — must live in discrete, machine-readable fields, not embedded in marketing copy or long-form descriptions. If a human has to read a paragraph to understand what a product is, an agent cannot evaluate it at all. Audit your product catalog for completeness and consistency before attempting any protocol integration. Every missing attribute, every inconsistent price, and every vague product title is a reason an agent will select a competitor instead.

Step 2: Ensure your APIs are accessible and performant

Stripe’s technical guide sets the benchmark: agents expect API responses in under 200 milliseconds. They query products, check inventory, retrieve shipping estimates, and trigger checkout programmatically. If your infrastructure cannot serve those calls reliably and at speed, agents will time out and move on. Map your existing API endpoints against what UCP and ACP require — product catalog, real-time inventory, pricing, checkout, and order status — and identify where gaps exist. If you are on Shopify, much of this is handled through the platform’s native infrastructure. If you are on a custom stack, this is an engineering project that needs to be scoped and resourced explicitly.

Step 3: Start optimizing for AI discovery, not just search rankings

Retail TouchPoints advises implementing structured data standards — specifically JSON-LD and Schema.org Product schemas — so AI systems can extract unambiguous product information from your pages. This is the foundation of Generative Engine Optimization: ensuring that when an AI agent evaluates whether your product matches a user’s intent, it has clear, structured, machine-readable signals to work with rather than marketing prose it has to interpret. Review pages, return policy pages, and shipping information pages all benefit from this treatment, not just product pages.

Step 4: Strengthen trust signals systematically

Agents evaluate merchants as well as products. Review coverage, transparent return policies, clear customer support pathways, and accurate business information all contribute to the trust signals that influence agent selection. According to Boston Consulting Group, agentic commerce will hinge on merchants’ ability to provide structured, explainable data and clear consent pathways for both humans and their AI agents. Anything that reduces uncertainty — for a human shopper or an AI system evaluating your offer — works in your favour.

Step 5: Build fraud controls designed for agent traffic

Standard fraud detection systems were built around human behavioural signals. Agent traffic looks different — higher query volumes, faster transaction sequences, less browsing context — and will trigger false positives in systems not calibrated for it. Stripe’s Shared Payment Token system scopes each token to a specific seller, bounded by time and amount, with full transaction observability. Work with your payments provider to understand how their fraud systems distinguish between legitimate agent traffic and malicious bots, and ensure your anomaly detection is tuned for the new transaction patterns agentic commerce generates.

Step 6: Diversify your acquisition channels while you build

None of the above removes the platform dependency risk inherent in agentic commerce. As you build your presence in AI ecosystems, invest in parallel in the channels you own outright — email, loyalty programmes, direct relationships, and community. The brands best positioned for agentic commerce in the long run will be those that can benefit from AI-driven visibility without becoming wholly dependent on any single platform’s rules or commercial arrangements.

Bottom Line

Agentic Commerce represents a major transition in how the digital economy operates.

For businesses, this means the internet is no longer only human-facing. It is becoming increasingly agent-facing too.

Brands that prepare for this future by building AI-readable websites, structured systems, transparent policies, and future-ready digital infrastructure will position themselves ahead of competitors.

📌 Editorial Note

At DMi Agency, we have made it our business to understand what AI is doing to search, discovery, and commerce before it becomes obvious to everyone else. The businesses we work with don’t wait for the shift to arrive. They prepare for it.