2025 established the protocols. 2026 is when merchant infrastructure determines who captures value. Here are the statistics and trends that will shape agentic commerce this year.

Key agentic commerce statistics for 2026

Market projections:

• $5 trillion in global agentic commerce volume by 2030 (McKinsey)
• $190-385 billion in US e-commerce spending through agents by 2030 (Morgan Stanley)
• 15-25% of e-commerce flowing through agentic channels by 2030 (Bain)

Current adoption:

• 39% of consumers, and over half of Gen Z, already use AI for product discovery (Salesforce)
• 23% of Americans bought something via AI in the past month (Morgan Stanley)
• 805% year-over-year increase in AI traffic to US retail sites on Black Friday 2025 (Adobe)

The performance gap:

• Less than 0.2% of e-commerce sessions currently come from ChatGPT referrals (Kaiser and Schulze)
• ChatGPT referrals convert 86% worse than affiliate links (Kaiser and Schulze)
• 4.4x higher conversion rates for AI-generated product recommendations vs traditional search (McKinsey)

Infrastructure reality:

• 42% of customers abandon purchases due to insufficient product information (Mirakl)
• $15 million annually lost on average per business due to poor data quality (Mirakl)
• 55% of US advertisers report inconsistent targeting and attribution from retail media networks (eMarketer)

The visibility gap driving 2026 investment

The core challenge is what we call the visibility gap. In traditional e-commerce, retailers see everything: impressions, clicks, dwell time, add-to-cart events, funnel drop-offs. In agent-mediated commerce, the behavioral data stream starts at the add-to-cart moment. The discovery, the browse, the consideration, the refined preferences all live inside ChatGPT. Attribution collapses. Personalization breaks. Retail media goes dark.

This explains the performance gap in the statistics above. Consumer demand for AI shopping is real (39% adoption, 805% traffic growth), but conversion lags (86% worse than affiliates) because merchant infrastructure was not built for agents. The 4.4x conversion potential exists, but only for merchants whose data infrastructure can support agent requirements.

The forcing mechanism of the next two years will be measurement. Not "is this cool," but "does this actually make us better?" Boards will demand rigor. Operators will need visibility. AI initiatives will live or die by outcomes, not experimentation theater.

2026 is when retailers either close that gap or accept permanent blindness to how customers actually shop.

Protocol evolution trends for 2026

The Agentic Commerce Protocol (ACP) launched in 2025 as a minimum viable standard: single-item purchases, basic checkout flows, straightforward transactions. 2026 will see significant protocol maturation.

Multi-item cart support becomes standard. Current ACP handles single products well, but consumers build carts. When a user tells ChatGPT to "order everything I need for taco night," the agent needs to build a cart with ground beef, tortillas, cheese, salsa, and limes, then execute that as a single transaction. Protocol extensions for multi-item transactions will roll out across major platforms.

The infrastructure implication: merchants need real-time inventory visibility across their entire catalog, not just individual SKUs. If the agent builds a cart and one item is out of stock, the entire experience degrades.

Subscription and recurring purchase support. Agents will move beyond one-time purchases to manage ongoing relationships. "Reorder my coffee beans every three weeks" requires protocol support for recurring billing and preference storage. PayPal's automatic ACP support for merchants in its network (launching 2026) will accelerate this capability across millions of merchants.

Cross-merchant orchestration emerges. When a customer asks for "a complete home office setup under $2,000," the ideal response includes products from multiple merchants. Protocol extensions for coordinated checkout will begin testing in 2026, though full production deployment may extend into 2027.

Agent-to-agent commerce develops. Beyond consumer-facing agents, B2B procurement agents will begin negotiating directly with supplier agents. A retailer's inventory management agent might automatically reorder stock from a manufacturer's sales agent when levels drop. These agent-to-agent transactions require even more structured data and API standardization than consumer-facing commerce.

The infrastructure implication across all these trends: merchants need standardized product data, real-time inventory visibility across entire catalogs, and APIs that support complex multi-item transactions. The same data quality foundations that enable consumer agent commerce will power B2B agent transactions.

Consumer behavior trends beyond checkout

Agents are not a new channel. They are a new consumer surface, on par with what Google Search was to discovery twenty years ago. The difference is that the conversation itself is the funnel.

2025's agents focused on product discovery and purchase. 2026's agents will handle the full customer lifecycle.

Post-purchase support becomes agent-mediated. Customers will ask their agents to track orders, initiate returns, request refunds, and resolve issues. "Where's my package?" and "I need to return these shoes, they don't fit" will become agent conversations rather than customer service tickets.

Agents will manage customer preferences across merchants. Instead of maintaining separate accounts with size preferences, payment methods, and shipping addresses at each retailer, customers will delegate this to their agent. Merchants who enable frictionless agent-mediated checkout will capture transactions that friction-heavy competitors lose.

Proactive replenishment becomes mainstream. Agents will monitor consumption patterns and initiate purchases before customers run out. "You usually order toothpaste every six weeks. Should I reorder?" This requires agents to access purchase history across merchants, a capability that will develop throughout 2026.

The infrastructure implication: merchants who share purchase history data with agents (with customer permission) will be included in replenishment recommendations. Those who do not will lose recurring revenue to competitors who do. Order management systems need agent-accessible APIs. Returns, exchanges, and refunds must be executable through structured requests, not just human-facing portals.

Retail media transformation trends

Retail media faces fundamental disruption as agents bypass traditional advertising surfaces.

Kantar's analysis predicts retail media will evolve from selling ad placements to licensing data access and influence. Traditional metrics (CPM, CPC, ROAS on sponsored placements) become less relevant when agents make decisions without seeing visual advertising.

The identity infrastructure gap compounds this challenge. According to eMarketer, 55% of US advertisers already report inconsistent targeting and attribution from RMNs. Agent commerce will accelerate this crisis for networks without cross-surface identity infrastructure.

Key shifts for 2026:

• Measurement moves from impressions to influence signals
• Brands pay for competitive positioning in agent decision-making, not banner placements
• Promotional APIs replace creative uploads for real-time counteroffer delivery
• First-party data loses exclusivity advantage as agents accumulate cross-merchant intelligence

RMNs that can prove influence on agent recommendations will command premium rates. Those that cannot will lose advertiser confidence. The infrastructure implication: RMNs need data infrastructure that can prove influence, requiring tracking not just final transactions but the data signals agents consumed when making recommendations.

Data infrastructure requirements for agent commerce

The Kaiser and Schulze research explains the performance gap: ChatGPT referrals convert poorly not because of consumer disinterest, but because merchant infrastructure cannot support agent commerce requirements.

Product data quality. Mirakl research shows 42% of customers abandon purchases due to insufficient product information, and over a quarter abandon due to poor image quality. Agents inherit these evaluation criteria and apply them at scale. Merchants with incomplete GTINs, thin descriptions, or inconsistent attributes get skipped entirely.

Real-time inventory sync. Feed updates must happen in minutes, not hours. Products showing as available in agent recommendations but out of stock at checkout create failed transactions and damage reliability scores.

Server-side data collection. AI agents do not trigger client-side JavaScript. They make API calls directly to merchant systems. Merchants relying on pixel-based tracking have measurement blind spots that grow as agent traffic scales.

Cross-surface identity resolution. Agent transactions arrive without the customer journey context merchants are used to. Connecting agent purchases to customer profiles enables personalization, attribution, and loyalty across surfaces.

Schema-compliant product markup. Schema.org Product markup on product pages allows agents to parse structured data directly. Products without proper markup force agents to guess, and agents do not guess in your favor.

These are not advanced capabilities. They are foundational data infrastructure that many enterprises have not prioritized because the ROI was not clear until now. The merchants who win in 2026 will be those who invested in these foundations in 2025, while competitors were still debating whether agentic commerce was real.

Agent authentication and fraud trends

78% of financial institutions expect fraud to spike from AI shopping agents. 2026 will see authentication frameworks formalize.

Know Your Agent protocols. Commerce will develop standards for distinguishing legitimate shopping agents from malicious bots. Cloudflare and others are pushing cryptographic verification via Signature-Input and Signature-Agent headers.

Fraud model recalibration. AI agents exhibit behavior patterns that traditional fraud detection flags as suspicious: rapid sequential orders, purchases across unrelated categories, unusual velocity patterns. Systems need recalibration to distinguish legitimate agent behavior from compromised accounts.

Liability framework development. The ambiguity about responsibility when agent transactions go wrong (merchant, customer, or agent provider) will begin resolving through industry standards and platform policies. Merchants need systems that capture the audit trail for agent-mediated transactions.

The infrastructure implication: server-side data collection must capture and route agent authentication signals. The first mile of data infrastructure becomes the point where legitimate agent traffic gets fast-tracked and suspicious traffic gets scrutinized. Transaction logging must capture agent identity, authorization context, and decision signals.

What these trends mean for retailers in 2026

The data points to a clear pattern: consumer demand for AI-assisted shopping is real and growing, but merchant infrastructure cannot yet capture it. The performance gap (4.4x higher conversion potential vs 86% worse actual conversion) represents the infrastructure opportunity.

The trajectory from 2025's sub-1% traffic share to the 15-25% projections requires rapid scaling. That scaling happens in 2026 and 2027. The protocols are stable. The consumer demand is accelerating. The question is whether your infrastructure can capture it.

Retailers who invest in data infrastructure now will capture disproportionate value as agent traffic scales. Those who wait for proven ROI will find themselves locked out: invisible to agents, unable to measure performance, disconnected from customers who have shifted to agent-mediated shopping.

The winners will not be the merchants with the biggest marketing budgets or the most recognizable brands. They will be the merchants with the cleanest data, the most responsive APIs, and the infrastructure that makes agents prefer them over competitors. That advantage builds now, compounds through 2026, and becomes increasingly difficult to replicate as the category matures.

‍

MetaRouter is first-mile data infrastructure for the agentic commerce era: server-side collection at the point where agent signals enter your ecosystem, identity resolution that connects transactions to customer profiles, and real-time routing across all commerce surfaces including agents. See how it works.

‍