Agentic Experience Design

The Agentic Commerce Ecosystem

Partners

The age of agentic commerce is not being built by any single organisation. It is emerging from the convergence of protocols, platforms, and research that together form the infrastructure upon which autonomous agents will discover, negotiate, transact, and act on behalf of humans. These are the organisations whose work most directly shapes the landscape that Agentic Experience Design exists to address. Each is advancing a different layer of the agentic stack - from agent communication and tool connectivity to payment rails and economic analysis - and together they define the design challenges that AXD practitioners must understand.

The Agentic Stack

The agentic commerce infrastructure is organised in complementary layers. Anthropic's MCP provides the vertical connectivity between agents and tools. Google's A2A enables horizontal agent-to-agent communication. OpenAI's ACP governs the commercial conversation between agents and merchants. Stripe's x402 provides the payment execution layer. Google's UCP and Shopify's merchant infrastructure make products discoverable to agents. IBM's watsonx orchestrates multi-agent enterprise workflows. And McKinsey quantifies the economic imperative driving it all.

Each layer creates distinct AXD design challenges - trust architecture, delegation design, observability, and recovery - that practitioners must address to build agentic experiences that humans can trust.

01

Google

Agent-to-Agent Protocol & Universal Commerce Protocol

Google has positioned itself at the centre of the agentic commerce infrastructure through two complementary protocols. The Agent-to-Agent Protocol (A2A) establishes the open standard for inter-agent communication - enabling autonomous systems built on different frameworks to discover, negotiate with, and delegate tasks to one another. The Universal Commerce Protocol (UCP), launched in January 2026 with Shopify as a co-development partner, creates the transactional layer that transforms AI conversations into completed purchases. Together, A2A and UCP represent the most comprehensive attempt to build the communication and commerce rails upon which agentic experiences will operate. From an AXD perspective, Google's protocols address the foundational coordination problem: how agents discover capabilities, establish trust, and execute delegated commercial intent across organisational boundaries.

Agentic Protocols
A2AUCP
Related Observatory Essays
Issue 022: The Agentic ProtocolsIssue 035: Universal Commerce Protocol
Google Universal Commerce Protocol
02

Shopify

Agentic Commerce at Scale

Shopify has emerged as the merchant infrastructure layer for the agentic age. Through Sidekick - its AI-native commerce assistant - and the development of Agentic Storefronts, Shopify is reimagining what it means for a store to serve customers who are not human. As a co-developer of Google's Universal Commerce Protocol (UCP), Shopify ensures that its millions of merchants are natively accessible to AI agents across every major platform. The January 2026 announcement that Shopify now enables native commerce at scale across all major AI channels represents a fundamental shift: from merchants optimising for human attention to merchants optimising for agent discovery and trust. From an AXD perspective, Shopify's approach validates the discipline's core thesis - that the machine customer requires entirely new design patterns for product discovery, trust signalling, and transaction completion.

Agentic Protocols
UCP
Related Observatory Essays
Issue 006: The Machine CustomerIssue 035: Universal Commerce Protocol
Shopify: AI Commerce at Scale
03

Stripe

Agentic Payments & the x402 Protocol

Stripe has defined the payment infrastructure for agentic commerce through its Agentic Commerce Suite and the x402 protocol. The x402 protocol - named after the HTTP 402 "Payment Required" status code - enables machine-native payments where AI agents can autonomously authorise and execute transactions without human intervention at the point of sale. Stripe's Five Levels of Agentic Commerce framework provides the most widely cited maturity model for understanding how organisations progress from AI-assisted browsing to fully autonomous purchasing. The Agentic Commerce Suite, launched in December 2025, allows merchants to connect product catalogues directly to AI agents. From an AXD perspective, Stripe addresses the critical payment trust problem: how delegated financial authority is scoped, constrained, and made recoverable when agents transact on behalf of humans.

Agentic Protocols
x402
Related Observatory Essays
Issue 023: Agent Payments ProtocolIssue 038: Stripe's Five LevelsIssue 039: The x402 Protocol
Stripe Agentic Commerce Suite
04

OpenAI

Agentic Commerce Protocol & Operator

OpenAI has moved decisively from language model provider to agentic commerce platform. The Agentic Commerce Protocol (ACP) - an open standard enabling conversations between buyers, their AI agents, and businesses to complete purchases - represents OpenAI's vision for how commercial transactions should work in the agentic age. Operator, launched in January 2025, demonstrated the first consumer-facing autonomous agent capable of browsing, comparing, and purchasing products on behalf of users across partners including Uber, Priceline, and DoorDash. The "Buy it in ChatGPT" feature extends this into native commerce within the conversational interface. From an AXD perspective, OpenAI's approach foregrounds the conversation-as-transaction model - where the boundary between information retrieval and commercial action dissolves, making delegation design and trust architecture essential rather than optional.

Agentic Protocols
ACP
Related Observatory Essays
Issue 022: The Agentic ProtocolsIssue 021: Designing for AI Autonomy
OpenAI Agentic Commerce Protocol
05

Anthropic

Model Context Protocol (MCP)

Anthropic has established the foundational connectivity layer for the agentic ecosystem through the Model Context Protocol (MCP). Released as an open standard in November 2024, MCP provides the secure, two-way connections between AI agents and external data sources, tools, and enterprise systems that make autonomous action possible. Where A2A governs agent-to-agent communication and ACP governs agent-to-merchant transactions, MCP governs the vertical relationship between an agent and the tools it needs to act in the world. The protocol has been adopted across the industry as the de facto standard for agent-tool integration. From an AXD perspective, MCP addresses the capability trust problem - how humans can understand, verify, and constrain the tools and data sources their agents access. Without MCP's structured connectivity, agents would operate as opaque systems with unverifiable capabilities, making trust architecture impossible.

Agentic Protocols
MCP
Related Observatory Essays
Issue 022: The Agentic Protocols
Model Context Protocol
06

IBM

watsonx Orchestrate & Enterprise Agentic AI

IBM has positioned watsonx Orchestrate as the enterprise orchestration platform for multi-agent systems. Where other partners focus on protocols and consumer-facing agents, IBM addresses the organisational complexity of deploying agentic AI at enterprise scale - coordinating multiple specialised agents across business functions, managing workflows that span departments, and maintaining governance over autonomous decision-making. The October 2025 announcement of five new agentic capabilities in watsonx Orchestrate - including domain-specific agents for HR, finance, and procurement - demonstrates IBM's thesis that enterprise agentic AI requires structured orchestration rather than individual agent deployment. From an AXD perspective, IBM's approach validates the discipline's emphasis on organisational trust architecture: the design of systems where multiple agents operate within defined authority boundaries, with clear escalation paths and human oversight mechanisms.

Related Observatory Essays
Issue 022: The Agentic Protocols
IBM watsonx Orchestrate
07

McKinsey Global Institute

Agentic AI Economic Impact Research

McKinsey Global Institute provides the economic evidence base that underpins the urgency of Agentic Experience Design. Their June 2025 research estimates that agentic AI has the potential to unlock $2.6 trillion to $4.4 trillion in additional value beyond existing AI deployments - with marketing, sales, and customer operations identified as the highest-impact domains. The November 2025 State of AI survey reveals that agentic AI adoption is accelerating faster than any previous AI capability wave, with organisations deploying autonomous agents across customer service, procurement, and supply chain management. From an AXD perspective, McKinsey's research quantifies the design challenge: as trillions of dollars in economic activity shifts to agent-mediated transactions, the organisations that design trust-governed, well-delegated agentic experiences will capture disproportionate value. The economic case for AXD is inseparable from McKinsey's evidence.

Related Observatory Essays
Issue 031: Financial Services ReadinessIssue 032: Retail Readiness
McKinsey: Seizing the Agentic AI Advantage
The AXD Imperative

These organisations are building the infrastructure. AXD exists to ensure that what gets built on top of it is designed for trust. Every protocol, every platform, every economic projection points to the same conclusion: autonomous agents will mediate an increasing share of human commercial activity. The question is not whether this will happen, but whether the experiences will be designed with the intentionality, trust architecture, and human oversight that the stakes demand.

That is the work of Agentic Experience Design.