---
id: "concept-a2a-commerce"
type: "concept"
source_timestamps: ["¶1", "¶2"]
tags: ["commerce", "automation", "ai-agents"]
related: ["concept-headless-bot-site", "framework-ai-agent-spectrum", "claim-intermediaries-compress-margins"]
definition: "A commerce model where autonomous AI agents interact directly with vendor systems to research, evaluate, and purchase products on behalf of human consumers."
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-nm-97-retailers-ai-shoppers"
sourceUrl: "https://hbr.org/2025/10/what-should-retailers-do-about-ai-shoppers"
sourceTitle: "What Should Retailers Do About AI Shoppers?"
---
# Agent-to-Agent (A2A) Commerce

## Agent-to-Agent (A2A) Commerce

**Definition:** A commerce model where autonomous AI agents interact directly with vendor systems to research, evaluate, and purchase products on behalf of human consumers.

Agent-to-Agent (A2A) commerce describes a paradigm in which the primary buying conversation occurs between two digital entities: a consumer's autonomous shopping agent and a vendor's digital systems. In this model, AI agents handle the entire customer journey — product discovery, research, evaluation, price comparison, and even checkout — often **without a single human click**.

Early manifestations include Amazon's *Buy for Me* (see [[entity-amazon-d97]]), and integrations by [[entity-perplexity-d97]], ChatGPT (see [[entity-openai-d97]]), and Gemini. While *full* delegation of the final purchase is currently limited, the automation of the **upper and middle funnel** is already disrupting traditional e-commerce and threatens to bypass vendor websites entirely.

This shift is the engine behind the [[framework-ai-agent-spectrum]] (how retailers should posture toward agents), the [[concept-retailers-prisoners-dilemma]] (the strategic trap it creates), and the economics warning in [[claim-intermediaries-compress-margins]]. The most aggressive infrastructure response is the [[concept-headless-bot-site]].

### Enrichment grounding
Independent analyses corroborate the structural nature of this shift. McKinsey frames *agentic commerce* as agents orchestrating intent-driven, personalized shopping flows and projects up to **$1 trillion** in orchestrated US B2C retail revenue by 2030. Deloitte and Bain agree agents create new pathways between shoppers and products that can bypass retailers' traditional digital channels. Kibo adds a useful distinction: **B2A** (Business-to-Agent — making data legible to agents via structured feeds/APIs) precedes **A2A** (buyer's and seller's agents transacting directly). See [[quote-first-buying-conversation]] for the source's framing.


## Related across articles
- [[concept-agentic-commerce-d5]]
- [[concept-agentic-commerce-d14]]
- [[concept-agentic-commerce-d15]]