The $5 Trillion Agent Commerce Opportunity: Why Most Businesses Aren't Ready

McKinsey puts the global agent commerce opportunity at $3 trillion to $5 trillion by 2030. Morgan Stanley estimates agentic shoppers could account for $190 billion to $385 billion in US e-commerce spending alone, capturing up to 20% of the market. According to Mordor Intelligence, the agentic AI in retail and eCommerce market sits at $60.43 billion in 2026, heading to $218.37 billion by 2031 at a 29.29% CAGR. eMarketer projects AI platform-driven ecommerce sales will hit $144 billion by 2029 — 8.8% of total retail ecommerce.

The interesting part isn't the size of the opportunity — it's how badly prepared the industry is to capture it.

A Mirakl survey of technology partners who work directly with retailers rated AI commerce readiness at 4.4 out of 10. According to the Rithum 2026 Commerce Readiness Index, three out of four executives say AI is advancing faster than they can keep up. And SoftServe's AI Data Readiness Report found that 73% of business leaders say their data strategy needs major updates or a complete overhaul.

Meanwhile, the consumers have already moved. The IBM Institute for Business Value reports that 45% of consumers use AI for at least part of their buying journey. Morgan Stanley estimates that roughly 23% of Americans bought something via AI in the past month. During Cyber Week 2025, Salesforce put the contribution of AI assistants and digital agents at $14.2 billion of the $79 billion spent online globally on Black Friday alone. Surveys consistently show younger demographics are particularly open to AI agents handling shopping on their behalf — and that willingness is translating into actual purchasing behaviour, not just survey responses.

The market is forming. The money is real. And most businesses are nowhere close to ready.

The Data Problem Nobody Wants to Talk About

Every conversation about agent commerce eventually lands on protocols, standards, and platform partnerships. Those matter. But the unglamorous truth is that most businesses will fail at agent commerce for a much simpler reason: their product data is terrible.

AI agents need clean, structured, machine-readable data to make purchasing decisions. They need accurate pricing, current inventory status, complete product attributes, and consistent information across channels. Most retailers don't have this. Their catalogues were built for humans who can look at a photo and infer what the product is. An AI agent can't do that — it reads structured data, and if the structured data is thin, incomplete, or contradictory, the agent moves on to a competitor whose data it can parse.

SoftServe's research confirms the scale of the problem: the majority of business leaders acknowledge that outdated or inaccurate data is actively undermining their decisions. This was already a problem for analytics and personalisation. For agent commerce, where the AI makes autonomous purchasing decisions based entirely on available data, it becomes the bottleneck.

The Visibility Blindspot

Mirakl's technology partner survey found that AI search visibility received the lowest readiness scores of any category they measured. Most retailers haven't even started monitoring how they appear in AI-generated responses.

Think about what that means. Brands are spending millions on traditional SEO and paid search while an entirely new discovery channel — one that may represent a quarter of e-commerce by 2030 — operates without any measurement at all. They don't know which queries surface their products in LLM responses. They don't know where they rank relative to competitors. They're competing in a channel they literally cannot see.

Adobe emphasises that visibility in LLMs requires a fundamentally different content strategy, and that brands need to start building GEO (Generative Engine Optimisation) practices now. This isn't an incremental extension of existing SEO — it's a new discipline.

Who's Actually Positioned to Win

While most retailers scramble, one business model has been quietly building the right infrastructure for years.

McFadyen Digital predicts that marketplace operators will be the winners in agentic commerce, and the reasoning is structural rather than strategic. Running a marketplace forces you to solve the exact problems that agent commerce demands: automated data validation across thousands of sellers, real-time inventory syncing, API-first architecture, and consistent product information standards. These aren't features marketplace operators chose to build — they're operational requirements that come with managing a multi-seller platform.

Ulta Beauty's leadership highlighted this on their Q4 2024 earnings call, noting that expanding their marketplace assortment — and the rich product content that comes with it from sellers and their community — positions them well as agent commerce matures.

First-party retailers without marketplace infrastructure face a harder path. Their product data sits in systems designed for human browsing, their APIs weren't built for autonomous agent transactions, and their discovery signals are optimised for Google, not for ChatGPT.

The Protocol Fragmentation Problem

The technical plumbing of agent commerce is being built right now, and it's already fragmented.

Google CEO Sundar Pichai announced the Universal Commerce Protocol at the National Retail Federation conference — the first serious attempt at standardising how AI agents interact with commerce systems. But Microsoft has adopted competing standards, and OpenAI's Agentic Commerce Protocol takes a different approach again.

For retailers, this creates an awkward strategic question. According to PAZ.ai, businesses will likely need to support both UCP and ACP, because choosing one exclusively means cutting off an entire AI shopping ecosystem. But implementing multiple protocols is expensive, and the standards are still evolving.

The security dimension adds further complexity. Shoppers are increasingly willing to let agents act on their behalf, but only with confidence that their data is protected and decisions are reversible. Industry observers expect leading brands to converge on transparent consent flows, granular permissions, agent action logs, and secure payment authorisation — but none of this is standardised yet.

The Geographic Picture

According to Mordor Intelligence, North America held 37.35% of the agentic AI retail market in 2025, while Asia-Pacific is projected to grow fastest at a 34.88% CAGR. The split reflects maturity versus momentum.

In Asia, the pace is striking. SoftBank aims for 1 billion AI agents by 2026. In India, Mordor Intelligence reports that 48% of retailers are piloting generative AI, backed by government skilling funds. The scale of experimentation in Asia-Pacific markets suggests that some of the most important lessons about agent commerce may come from outside North America and Europe.

What This Means for Different Businesses

Traditional retailers face the sharpest disruption. When AI agents handle discovery and purchasing, they don't need to visit your website. That eliminates the assumptions behind cross-sell, on-site media, loyalty engagement, first-party data collection, and click-based attribution. The question becomes whether to optimise for AI discoverability or fight to maintain direct customer relationships — and increasingly, that's not a choice but a sequencing problem. You need both.

B2B commerce may see even larger effects. Forrester predicts that 1 in 5 sellers will need to respond to AI-powered buyer agents with dynamically generated counteroffers via their own seller-controlled agents. B2B procurement is repetitive, specification-driven, and high-volume — exactly the kind of workflow AI agents handle well.

Technology providers face growing demand for agent commerce infrastructure. CB Insights identifies three AI agent infrastructure markets primed for growth as enterprise adoption shifts from experimentation to production. The common thread: enterprises are struggling to measure the ROI of AI agents, creating opportunity for platforms focused on performance visibility, context management, and cost attribution.

What to Do About It

The specifics depend on your starting point, but four priorities apply broadly.

Audit your agent readiness. Can your product catalogue be read and understood by AI agents? Are your APIs ready for automated transactions? Do you know how your brand appears in AI-driven search results? If the answer to any of these is no — or "I'm not sure" — that's the place to start.

Take a position on protocols. UCP, ACP, or both? Wait for consolidation or implement now? This is a strategic decision with multi-year consequences, not a technical checkbox.

Treat data quality as a strategic priority. Not a cleanup project. Not an IT task. The ability to support autonomous, agent-completed transactions cleanly and reliably depends on data quality more than any other single factor.

Start monitoring AI visibility. Adobe is right that this requires a new kind of content strategy. If you can't see how AI agents perceive your brand and products today, you can't improve it.

The Real Question

McKinsey compares the potential impact of agent commerce to the web and mobile revolutions, arguing it could unfold even faster because AI systems operate along the same digital pathways as human users — they ride the rails of existing commerce infrastructure rather than requiring entirely new ones.

Whether that comparison holds up remains to be seen. But the direction is clear, the investment is real, and the businesses that treat this as a future problem are already falling behind the ones that started six months ago.

BCG warns that without action, retailers risk becoming background utilities in agent-controlled marketplaces. That's probably overstated for 2026. By 2028, it might not be.

---

Assess your current position with the Zeodyn™ Scanner — a free, comprehensive evaluation of whether AI agents can do business with you.