The Merchant's Readiness

The platform vendor's perspective, exemplified by Shopware's Agentic Commerce Cheat Sheet, excels at defining operational metrics. It correctly identifies KPIs like Data Fill Rate and Update Speed as critical. In Agentic Experience Design (AXD) terms, these map directly to the foundational pillars of Signal Clarity and Reputation via Reliability. A high data fill rate ensures an agent can understand a product, while rapid update speeds ensure the agent receives timely and accurate information, building a reputation for reliability. These metrics are necessary, providing the informational bedrock for any agentic system, but they only address the machine-readable aspects of commerce, not the human-centric conditions for delegation.

The vendor playbook’s limitations become apparent in its "emerging metrics," specifically the Bot Trust Score. This metric, designed to measure consumer confidence in AI agents, correctly identifies trust as the currency of agentic commerce. However, it treats trust as a sentiment to be measured - a survey response or a behavioural ratio - rather than as a system to be engineered. It asks "how much do customers trust us?" but fails to ask "what have we built to make ourselves trustworthy?" This approach mistakes a lagging indicator (sentiment) for the leading structural condition (trust architecture), leaving merchants to monitor trust levels without providing the tools to influence them.

The proposed implementation roadmaps further expose this structural gap. They offer a pragmatic, phased approach to operational rollout - auditing data, piloting use cases, and scaling - but consistently omit the parallel work of designing for trust. The guidance to "build trust at each step" is offered without specifying the mechanisms. An AXD-informed approach would integrate trust architecture design into every phase: assessing trust baselines during preparation, testing trust formation and recovery protocols during piloting, and monitoring for trust erosion during scaling. Without this integration, the roadmap leads to a functionally complete but socially inert system.

If this thesis is correct, organisations must fundamentally reframe their approach to agentic commerce readiness. The work is not sequential - operations first, trust later - but parallel and deeply integrated. Product leaders and designers must treat trust architecture as a co-equal discipline alongside API development and data syndication. This means expanding project plans to include a dedicated workstream focused on designing, testing, and scaling the mechanisms of trust for each specific use case. The goal is not merely to achieve a high Bot Trust Score, but to build the underlying systems that earn it.

Practically, this requires a new set of questions to be embedded in the development lifecycle. During the initial assessment, teams must ask: "What are the consequence levels of this delegation, and what initial trust signals are required?" In the pilot phase, the focus must shift from "Does the agent work?" to "How do we gracefully handle failure and recover the user's trust?" For scaling, the challenge becomes architectural: "How do our systems for trust calibration and erosion detection scale with our user base?" By making trust architecture a core competency, organisations can move beyond building systems that are merely capable of autonomous action to creating experiences that humans are confident in delegating to.

This is the merchant’s true readiness challenge: not technical capability, but trust competency.