McKinsey's Agentic Commerce Forecast: Why 2025 Is the Last Year of Shopping As We Know It
McKinsey just published what might be the definitive analysis of agentic commerce — and the headline is stark: 2025 will likely be the last year consumers shop as they do now.
The report, "The Agentic Commerce Opportunity: How AI agents are ushering in a new era for consumers and merchants", projects $1 trillion in US B2C retail revenue mediated by AI agents by 2030, with global figures reaching $3-5 trillion. But the numbers only tell part of the story. What's more interesting is how McKinsey thinks this unfolds — and what it means for merchants preparing for the shift.
From SEO to ACO: The Optimization Battleground Moves
McKinsey frames the transition as a fundamental platform shift. Just as businesses moved from traditional advertising to SEO when Google became the primary discovery channel, commerce is now moving from SEO to Agentic Commerce Optimization (ACO).
The logic is straightforward: if consumers stop browsing and start delegating purchases to AI agents, then being discoverable by those agents becomes the entire game. An AI agent doesn't scroll through search results. It synthesizes, compares, and decides based on structured data, real-time availability, and contextual relevance.
The merchants who win will be those who optimize for agentic discovery — not just human eyeballs.
The OpenAI-Stripe Protocol: Infrastructure Matters
One detail from McKinsey's analysis that hasn't gotten enough attention: the Agentic Commerce Protocol co-developed by OpenAI and Stripe. This isn't a product feature. It's infrastructure.
The protocol defines how AI agents communicate with merchant systems to discover products, verify pricing, process payments, and hand off fulfillment. OpenAI made it open-source, which signals something important: agentic commerce will be an industry-wide shift, not a walled garden controlled by one platform.
For merchants, this is good news. It means you won't need to build separate integrations for ChatGPT, Gemini, Perplexity, and whatever comes next. A single protocol-compliant integration should work across agents.
The Automation Curve: Not All Categories Move at Once
McKinsey's companion piece on "The Automation Curve in Agentic Commerce" breaks down which product categories will see the fastest agent adoption. High-frequency, low-consideration purchases — household staples, replenishment items, gifts — are already being delegated to agents. Complex, high-consideration purchases (furniture, enterprise software, professional services) will take longer.
This has implications for how you prioritize. If you sell consumables or replenishment goods, agentic readiness is urgent. If you're in considered purchases, you have more time — but not much.
What Merchants Should Do Now
McKinsey identifies the same infrastructure gaps we've been talking about at CHATTERgo:
-
Data quality is the foundation. Agents amplify your data problems. Clean product catalogs, accurate inventory, and comprehensive knowledge bases aren't nice-to-haves anymore.
-
Real-time connectivity is non-negotiable. An agent that recommends an out-of-stock product loses trust — and that trust doesn't come back.
-
Merchant control matters. As agents gain more autonomy, you need visibility into what they're saying and doing on your behalf.
The CHATTERgo platform was built for exactly this moment: Shopify and Magento integration for real-time data, knowledge base ingestion, multi-channel deployment, and merchant control over AI behavior.
The Bottom Line
McKinsey's analysis isn't speculative. The $1 trillion figure is based on current adoption curves, existing technology, and consumer behavior that's already shifting. One in five Cyber Week 2025 orders involved an AI agent. That was four months ago.
The question isn't whether agentic commerce will reshape retail. It's whether your store will be ready when the agents come shopping.
Source: McKinsey & Company — "The Agentic Commerce Opportunity: How AI agents are ushering in a new era for consumers and merchants", October 2025.