Agentic commerce is online shopping where an AI agent, not the customer - searches, compares, and buys. It has moved from demos to live transactions: AI assistants now complete purchases inside ChatGPT, Google AI Mode, Microsoft Copilot, and others, through open protocols such as OpenAI and Stripe's ACP and Google and Shopify's UCP. For retailers, the implication is stark: when an agent does the shopping, the customer may never see your website, and the agent decides what to recommend based on structured product data, not visual design. Getting ready is less about picking a protocol and more about having the clean, machine-readable data foundation that every protocol depends on.

This is the forward edge of everything we've written about AI search visibility and product data readiness. Discovery was the first wave; transaction is the second - and it's already underway.

What "agentic commerce" actually means

In a traditional online purchase, a person browses, compares, adds to basket, and checks out. In agentic commerce, an AI agent does that work on the customer's behalf. The shopper expresses intent in natural language - "a quiet, energy-efficient washing machine under £600 with next-day delivery" - and the agent queries inventory, pricing, and shipping across retailers, selects the best match against those constraints, and completes the purchase, often without the customer ever opening a product page.

These aren't the rules-based support chatbots retailers have used for a decade. They are autonomous agents that search, compare, and transact. And the shift this forces is fundamental: for twenty years e-commerce has been optimised for human browsers; now it has to be legible to AI agents that decide on structured data rather than design.

This is no longer theoretical

The pace of 2026 has surprised even close observers. The headline developments:

• AI checkout is live at scale. ChatGPT's instant-checkout capability has been operating since late 2025, letting users buy directly inside conversations, and is now broadly available in the US. Microsoft Copilot, Perplexity, and Google's AI Mode have all added purchasing.

• The protocols arrived - and there are two. OpenAI and Stripe's Agentic Commerce Protocol (ACP) went live in late 2025 with partners including Shopify, Etsy, Instacart, and DoorDash. Google then launched the Universal Commerce Protocol (UCP) at NRF in January 2026, co-developed with Shopify and backed by a large coalition spanning major retailers, card networks, and platforms. Both standardise how agents discover, transact, and handle post-purchase.

• The payment networks moved fast. Through early 2026, Mastercard ran its first live agentic transactions, Visa launched a trusted-agent protocol commercially, and American Express introduced agentic developer tooling - including purchase protection for registered AI agents. Agent-initiated payments are being supported at scale.

• Live data access is standardising. Beyond discovery, the Model Context Protocol (MCP) lets AI systems query a retailer's source-of-truth systems directly - inventory, pricing - rather than scraping rendered pages.

The financial case behind the rush is large. McKinsey estimates agentic AI will influence $3–5 trillion in global retail commerce by 2030, and Morgan Stanley projects nearly half of online shoppers will use AI shopping agents by 2030, making up around a quarter of their spending. Early signals back it: Adobe reported AI-referred visitors converting at a markedly higher rate than traditional search visitors over Black Friday 2025, and Salesforce data suggested retailers with AI-agent integration saw several times better sales growth during Cyber Week than those without.

The uncomfortable shift: the customer never sees your store

The strategic consequence is worth stating plainly. When an agent searches a product graph, compares options, and checks out on the customer's behalf, your carefully designed homepage, category pages, and merchandising are bypassed. The agent's decision rests on whether your data is complete, structured, current, and machine-readable - and whether your store is even reachable through a protocol at all. As one widely-shared industry warning put it, if an agent cannot read your store, it simply recommends a competitor whose data it can.

This is the same root truth as AI search, raised to a higher stake: in discovery, weak data costs you a citation; in agentic commerce, it costs you the transaction.

You don't have to pick a side - but you do have to be ready

There's a temptation to treat this as a platform bet: back ACP, or back UCP, and wait. That's the wrong frame, for two reasons.

First, the market is consolidating around interoperability, not a single winner. Most brands will need to be reachable through both major protocols, because customers will arrive through different AI front-ends. Betting on one is a needless risk.

Second - and this is the point that should reassure constrained teams - the foundation beneath every protocol is the same. Whichever agent is asking, it needs clean, structured, accurate product data with real-time pricing and availability. Get that right and you are ready for ACP, UCP, and whatever comes next. Get it wrong and no protocol integration will save you, because the agent still can't trust your data.

That reframes the work from a risky platform commitment into a no-regret foundation: invest in the data layer that all of them share. It also means you rarely need to replatform. Industry guidance is converging on an adapter approach - exposing existing infrastructure to agentic protocols through a translation layer rather than rebuilding core systems - which is exactly the kind of phased, low-integration path we set out in how to make AI-readiness progress when your roadmap is already full.

The under-discussed risk: the agentic readiness gap

Most retailers racing into this are focused on discovery - being visible to agents. Far fewer are asking whether the rest of the operation can keep up once an agent actually buys. When a machine transacts in seconds against real-time constraints, the pressure moves downstream: is your inventory data accurate to the minute? Can your fulfilment promise what the agent committed to? Does your returns and post-purchase data hold up?

This distance between digital presence and operational capability is the agentic readiness gap, and it's where many retailers will be caught out. It's also why agentic readiness is a data-and-operations discipline, not a marketing campaign - and why a clear-eyed assessment of where you actually stand is worth more than rushing to bolt on a checkout button.

What this means for UK retailers

UK retailers are exposed to both the opportunity and the gap. Demand is ahead of the European curve - British shoppers are the most confident AI adopters in Europe, around 93% have used tools such as ChatGPT in the past year, and chat-based platforms already drive over 50 million monthly shopping-intent visits in the UK. But readiness lags: among larger UK retailers, 54% cite legacy-system integration and skills gaps as a leading barrier, and only about 17% of European retailers have scaled AI across multiple functions, versus 28% in North America. The encouraging part is that the highest-leverage preparation - clean, structured, governed product data - sits largely in the data layer and can progress alongside an existing roadmap, without a disruptive rebuild.

How to get ready

A sensible, low-risk sequence:

1. Audit your data and your reachability. Assess product-data completeness, structured-data coverage, and real-time accuracy - and map which parts of your commerce lifecycle (browse, cart, checkout, returns) are actually exposed through APIs versus UI-only and therefore invisible to agents.

2. Fix the shared foundation first. Prioritise the clean, structured, current data that every protocol depends on, starting with your highest-value categories.

3. Adopt an adapter mindset. Plan to expose existing systems to agentic protocols through a translation layer, supporting interoperability across both major standards rather than betting on one.

4. Close the readiness gap. Make sure inventory accuracy, fulfilment, and post-purchase data can honour what an agent commits to.

How VE3 helps

VE3 is a global technology consultancy specialising in data, AI, cloud, and digital transformation. We help retailers prepare for agentic commerce from the foundation up - building the clean, structured, governed product data that every protocol relies on, assessing API and data reachability, and planning low-integration adapter approaches that avoid a disruptive replatform. Our position is deliberately vendor-neutral: we help you become reachable and trustworthy to any agent, on any protocol, rather than betting your strategy on one. A scoped data-readiness audit is the lowest-risk way to see where you stand against the agentic shift - and what to fix first.

Want to know whether AI agents can find, trust, and transact with your products? Talk to us about a scoped data-readiness audit.