Every few months, a new report lands with a headline figure so large it stops meaning anything. "$3 trillion in agentic commerce by 2030" is one of those figures. It comes from McKinsey's October 2025 analysis of how AI agents will reshape purchasing behavior. It is almost certainly directionally correct. It is also almost certainly useless to a merchant running a $1 million Shopify store unless someone translates it.
This piece is that translation.
What the number actually represents
McKinsey's projection covers the total value of commerce influenced or executed by AI agents, globally, by 2030. In the US alone, the firm estimates up to $1 trillion in B2C retail revenue will be orchestrated through agentic channels. Morgan Stanley puts the figure more conservatively at $385 billion in direct impact by the same date.
The range in estimates is wide because the definition of "agentic commerce" is itself still crystallizing. At the narrow end, it means purchases completed autonomously by an AI agent on a consumer's behalf. At the broader end, it includes any transaction where an AI meaningfully shaped the discovery or decision process before checkout.
The narrower definition is futuristic. The broader one is happening now.
Shopping-related generative AI searches grew 4,700% between July 2024 and July 2025. AI referrals from ChatGPT and Perplexity to ecommerce brands increased 752% year-over-year during the 2025 holiday season. These are not projections. They are measurements of behavior that already occurred.
So what does this mean for a brand doing $500K-$5M a year?
The honest answer is: it depends on what you sell and who you sell it to.
Let's break it down by product category and consumer behavior.
Where agentic commerce is arriving fastest
AI agents are most active in categories where purchases are:
- Replenishment-driven: consumables, subscriptions, household staples. Consumers already want to delegate routine purchases. AI agents make this easy.
- Research-heavy: electronics, outdoor gear, supplements, specialty tools. These are the categories where consumers ask "what should I buy?" rather than navigating directly to a brand.
- Gift-driven: the query "what's a good gift for X" is one of the most common AI shopping prompts. Independent brands with strong narrative and clear positioning do well here.
If you sell in any of these categories, the shift is arriving for your customers sooner than the 2030 headline implies.
Where it is arriving more slowly
Categories where tactile experience drives the decision, think furniture and high-end apparel, are seeing slower AI-driven purchase completion. Consumers still want to see and feel. But even here, AI is increasingly shaping the shortlist. The discovery and research phases are moving into AI interfaces even when the final purchase does not.
The part the hype pieces leave out
Most coverage of agentic commerce frames it as a threat to small merchants. The standard narrative: large brands with dedicated data teams and platform partnerships will capture the AI recommendation layer, and independent sellers will be squeezed out.
That narrative misreads how AI recommendations actually work.
AI agents recommend based on clarity, completeness, and trust signals, not brand size. A well-documented independent brand with strong third-party reviews, specific use-case content, and complete product data is more recommendable to an AI than a large brand with vague, keyword-stuffed descriptions and scattered customer reviews.
In traditional search, brand size correlated strongly with domain authority and link profiles, which correlated with rankings. In AI discovery, the relevant inputs are more democratically distributed. A small cookware brand that publishes clear content about who their pans serve, why, and what customers say about them can appear in AI recommendations alongside much larger competitors.
This is not guaranteed. It requires deliberate preparation. But the preparation is available to any merchant willing to do it.
What "agent-ready" means at the $1M-$5M level
At the practical level, being agent-ready for an independent merchant means four things.
One: product data that speaks in complete sentences, not attribute strings. AI agents need to understand a product well enough to recommend it in response to a natural language question. "Ceramic mug, 12oz, navy blue" does not tell an agent enough to recommend your mug when a consumer asks for the best gift for a coworker who drinks too much coffee.
Two: trust signals that live outside your domain. Reviews on Google, Trustpilot, editorial coverage, and press mentions are the inputs AI systems use. Merchants who have cultivated these signals have an advantage that compounds over time.
Three: return policies, shipping terms, and availability that are clearly stated and machine-readable. AI agents increasingly surface policy information alongside product recommendations.
Four: presence on channels that AI systems actively index. This includes marketplaces and feeds that connect to AI shopping interfaces, schema-marked product pages, and platforms like Nohi that are specifically built for AI agent discovery.
