Zero-Click Commerce

For two decades, digital commerce has been organised around a single assumption: the customer will visit a product page. Every element of the eCommerce stack - search engine optimisation, product photography, review systems, comparison tools, checkout flows, abandoned cart recovery - exists to attract a human to a page and persuade them to click "buy." The entire industry is built on clicks.

Zero-click commerce eliminates the click. When an AI shopping agent receives a mandate - "reorder my usual coffee beans," "find me a birthday gift for my partner under £50," "replace my running shoes with the same model in the next size up" - and executes that mandate without the human ever seeing a product page, the foundational assumption of digital commerce collapses. The product page still exists, but the customer never visits it. The reviews still exist, but the customer never reads them. The checkout flow still exists, but the customer never navigates it.

This is not a theoretical scenario. It is the logical endpoint of trends already visible in 2026. Perplexity Shopping surfaces product recommendations within conversational AI responses - the user never leaves the chat interface. Google's AI Overviews increasingly include purchase links that bypass traditional search results. Amazon's Subscribe & Save automates repeat purchases with no human interaction beyond the initial setup. Each of these systems moves commerce closer to zero-click execution.

The AXD Institute considers zero-click commerce the defining design challenge of the agentic era - not because it is the most common form of commerce today, but because it is the form that most completely exposes the inadequacy of traditional experience design. When the customer is absent, every assumption of interface-based design must be reconsidered.

In traditional commerce, the act of purchasing includes an implicit trust verification step. The customer sees the product, reads the description, checks the price, evaluates the merchant, and makes a conscious decision to proceed. This verification is so embedded in the purchase process that it is invisible - but it is doing critical trust work. The customer is continuously assessing: Is this the right product? Is this a fair price? Is this a trustworthy merchant? Do I actually want this?

Zero-click commerce removes every one of these verification steps. The agent makes all of these assessments on the human's behalf, using criteria that may or may not align with what the human would have decided if they had been present. This creates what the AXD Institute calls the trust transfer problem: the trust that was previously distributed across the entire purchase journey must now be concentrated entirely in the human-agent relationship.

The four layers of trust become critical in zero-click commerce:

Competence trust must be absolute, not approximate. In a zero-click purchase, there is no human review to catch a competence failure. If the agent selects the wrong product, the human discovers the error only when the package arrives. The cost of competence failure is higher because the feedback loop is longer.

Integrity trust must be verifiable, not assumed. When the human never sees the alternatives the agent considered, they cannot assess whether the agent's selection was genuinely optimal or influenced by hidden incentives. Zero-click commerce demands radical transparency about agent decision-making - not at the point of purchase (the human is absent) but in post-purchase reporting and audit trails.

Benevolence trust must be proactive, not reactive. A benevolent agent in a zero-click context must anticipate harms the human cannot see: counterfeit products, subscription traps, dynamic pricing exploitation, merchant fraud. The agent must act as a fiduciary - placing the human's interest above its own operational efficiency.

Predictability trust must be earned through transparency. The human must be able to understand, after the fact, why the agent made the choices it made. This requires structured agent observability - decision logs, reasoning traces, and outcome reports that make invisible actions legible.

If the customer never sees the product page, everything that matters must be specified in advance. This is the delegation design challenge of zero-click commerce - and it is fundamentally different from the delegation design of assisted shopping.

In assisted shopping, the human can correct course during the process. "No, not that one - something more like this." In zero-click commerce, the mandate must be complete before execution begins. This places extraordinary demands on the quality of outcome specification - the structured expression of what the human wants, what constraints apply, and what trade-offs are acceptable.

Explicit mandates work well for repeat purchases and commodity goods. "Reorder the same coffee beans" is a clear, executable mandate. "Buy me something nice" is not. The design challenge is building systems that can handle the spectrum between these extremes - translating vague human intent into actionable agent instructions without requiring the human to think like a database query.

Learned preferences fill the gap between explicit mandates and human intent. Over time, a well-designed agent learns that "something nice for my partner" means "a book by a literary fiction author we have not read, hardcover, under £25, from an independent bookshop if possible." This learning must be transparent, correctable, and bounded - the agent should never extrapolate beyond its evidence or assume that past preferences predict future desires with certainty.

Escalation boundaries define when zero-click becomes one-click. Not every purchase should be zero-click. The agent must know when to pause and ask: when the price exceeds a threshold, when the product category is unfamiliar, when the merchant has low trust signals, when the mandate is ambiguous. Designing these boundaries - the interrupt frequency calculus - is essential to preventing zero-click commerce from becoming zero-oversight commerce.

Zero-click commerce inverts the merchant's competitive challenge. In traditional commerce, the merchant competes for human attention - through advertising, brand storytelling, visual merchandising, and checkout optimisation. In zero-click commerce, the merchant competes for agent selection - through structured data, machine-verifiable trust signals, and programmatic transaction surfaces.

The Merchant's Readiness analysis identifies four capabilities that become critical in a zero-click world:

Machine-readable product data. If the agent never visits your product page, your product information must be available through structured data feeds, APIs, and schema markup. Product descriptions written for human persuasion are invisible to agents. Specifications, attributes, compatibility data, and availability status - expressed in machine-interpretable formats - are what agents consume. Signal Clarity is the first pillar of readiness.

Verifiable trust signals. Agents cannot evaluate brand storytelling. They evaluate data: fulfilment accuracy rates, return percentages, dispute resolution records, customer satisfaction scores. Merchants that invest in building machine-verifiable reputation via reliability will be selected by agents. Those that rely on brand recognition alone will be bypassed.

API-accessible commerce. Zero-click commerce requires the entire transaction - product discovery, price verification, inventory check, purchase execution, order confirmation - to be completable through APIs without any human-facing interface. The Engagement Architecture essay details the technical requirements.

Agent-compatible payment processing. The emerging agent payments ecosystem - including Visa Intelligent Commerce, Mastercard Agent Pay, Google's Agent Payments Protocol (AP2), and the x402 protocol - is building the financial infrastructure for zero-click commerce. Merchants that integrate with these systems early will be accessible to agents. Those that require human-facing checkout will be excluded from zero-click transactions.

Zero-click commerce is the purest expression of what the AXD Institute calls absent-state design - the practice of designing experiences for moments when the human is not present. This concept, introduced in the AXD Manifesto, holds that "the most consequential experiences happen when no one is watching."

In zero-click commerce, the human is absent during the entire purchase cycle. They do not see the options considered. They do not evaluate the trade-offs made. They do not experience the transaction. They experience only the outcome - the product that arrives, the money that was spent, the confirmation that appears in their inbox. The quality of the experience is determined entirely by the quality of the delegation design and the trust architecture that governs it.

This is why the AXD Institute argues that zero-click commerce is not merely a commercial innovation - it is a design paradigm shift. Traditional UX design assumes the user is present and interacting. Zero-click commerce assumes the user is absent and the agent is acting. Every design principle must be reconsidered:

Instead of usability (can the human complete the task?), the design challenge is delegability (can the human express the task clearly enough for the agent to complete it?). Instead of feedback (does the interface respond to human actions?), the challenge is observability (can the human understand what the agent did after the fact?). Instead of error prevention (does the interface prevent human mistakes?), the challenge is trust recovery (can the system heal when the agent makes a mistake?).

Zero-click commerce is where Agentic Experience Design proves its necessity. Without a design discipline built for absent-state interactions, zero-click commerce will be designed by engineers optimising for efficiency rather than designers optimising for trust. The result will be systems that work - until they fail, and no one knows how to recover.