🌱 The impact of AI on commerce

A McKinsey report sets out how agentic commerce - where AI agents search for and purchase products on behalf of consumers - might progress. It claims that driving forces are already in place for this, including the readiness of agents to make decisions and the suitability of infrastructure such as MCP, A2A, etc.

It sets out a 6-level automation curve:

• Level 0: Programmatic Convenience (Set & Forget)
• Level 1: Assist (Cognitive Sidekick)
• Level 2: Assemble (Personal Shopper)
• Level 3: Authorize (Supervised Executor)
• Level 4: Autonomize (Intent Steward)
• Level 5: Network Autonomy (Multi-Agent Commerce)

Gartner predicts that by 2028, 90% of B2B buying will be AI agent intermediated, pushing over $15 trillion of B2B spend through AI agent exchanges. Source

E-commerce businesses must take action to ensure that they stay viable. These include:

• Data Quality. Since AI agents don't browse web pages but query structured data, the quality of data determines whether an AI can find the company. This requires clean cataloguess, consistent metadata, and real-time inventory.
• Machine readability. Exposing catalogues, pricing, inventory, shipping, promotions, and return logic as APIs will enable agents to engage.
• Agent authentication. This includes supporting authenticated APIs for agents being developed by Visa and Mastercard, purchase authorisations constrained by budget, time, and category, and auditable transaction logs.

Sources:
Source