Agentic Commerce Landscape (2026): Where Value Will Accrue

Agentic commerce is no longer a speculative edge case. In 2026, it is becoming a new operating layer for retail.

The most important shift is simple: discovery, evaluation, and recommendation are moving inside AI conversations, while checkout is becoming API-native. That changes who controls demand, what gets measured, and where margins go.

Market context: rapid adoption, large upside

Multiple datasets now point in the same direction:

• MetaRouter and BCG research suggests AI-agent mediated commerce could approach US$5 trillion globally by 2030.
• Morgan Stanley estimates US e-commerce via agents at US$190-385 billion by 2030.
• Bain projects 15-25% of e-commerce volume flowing through agentic channels by the end of the decade.
• Consumer behavior is already shifting: around 39% of consumers (and over half of Gen Z) use AI for product discovery, and 23% of Americans report buying through AI in the past month.

This is no longer about whether agentic shopping will matter. The question is who becomes discoverable, trusted, and executable inside agent workflows.

The visibility gap is now a strategic risk

Most merchants still measure performance with assumptions from web-era funnels: impressions, clicks, sessions, and on-site conversion.

That model breaks in agentic commerce.

Today, less than 0.2% of e-commerce sessions are attributed to ChatGPT referrals, and those referrals convert materially worse than traditional affiliate traffic. At first glance, this looks like weak channel quality. In reality, it reflects a measurement blind spot:

• Product discovery happens inside the agent's interface.
• Product evaluation and shortlisting happen before the merchant ever sees a request.
• The merchant often sees only the add-to-cart or checkout call.

This is the new visibility gap. If you cannot observe influence, you cannot optimize for it.

And the upside is real: AI-generated product recommendations are already showing significantly better conversion performance than traditional search in several datasets. But only merchants with agent-ready feeds, clean attributes, and reliable execution rails are capturing that lift.

Protocols are maturing, but still fragmented

Early agentic protocols were optimized for narrow flows (single item, straightforward intent, one merchant). In 2026, the requirement set is broader:

• Multi-item carts
• Recurring purchases and subscriptions
• Cross-merchant orchestration
• Real-time inventory and price guarantees
• Consent, liability, and auditability for delegated spend

The protocol stack is moving in that direction, but not uniformly. Discovery, identity, authorization, and checkout are still split across ecosystems.

Comparing the major initiatives

Stripe Agentic Commerce Suite + ACP

Stripe has built one of the strongest transaction rails for agentic checkout. ACP provides an open specification for agent-to-merchant transaction flows, while Stripe's suite handles catalog syndication, checkout paths, and fraud controls.

Strengths

• Mature payments infrastructure and merchant trust
• Flexible integration paths (embedded, headless, machine payment)
• Strong safety model (tokenization, risk controls)

Limitations

• Merchant coverage depends on Stripe integration and opt-in
• Long-tail merchant discovery remains incomplete
• Multi-merchant orchestration is still emerging

OpenAI Instant Checkout

OpenAI proved that conversational commerce can drive direct purchase behavior from the chat surface itself, with an explicit take-rate model.

Strengths

• Massive intent surface in ChatGPT
• Friction-reduced "buy in conversation" experience
• Clear merchant monetization model via transaction fee

Limitations

• Early merchant depth remains narrow
• Cross-merchant carts and advanced recurring flows are limited
• Ranking opacity means discoverability is highly platform-dependent

Google AP2 + UCP

Google's approach combines payment authorization standards (AP2) with a broader orchestration layer (UCP), linking discovery and execution in the same ecosystem.

Strengths

• Strong trust posture around consent and intent verification
• Deep distribution via Search and AI Mode
• Powerful end-to-end flow control from discovery to checkout

Limitations

• Requires merchant and ecosystem opt-in
• Raises centralization concerns around ranking and platform control
• Financial institutions must adapt policies for delegated agent actions

Shopify MCP/UCP + Agentic Plan

Shopify is lowering activation friction for SMBs by letting merchants syndicate once and appear across major AI surfaces without standing up custom infrastructure.

Strengths

• Fast path for merchant participation
• No monthly fee model encourages experimentation
• Merchant tooling for knowledge and answer quality

Limitations

• Coverage is primarily Shopify-native merchants
• Universal discovery outside platform boundaries remains unresolved
• Cross-merchant checkout remains constrained

MoltBillboard (majbbtech)

MoltBillboard introduces a different layer: an agent-native discovery and attribution surface built as machine-readable digital real estate. The public pixel canvas is human-visible, while manifests, offers, and action IDs create a bot-first path for discoverability and attributable action.

Strengths

• Novel discovery primitive not tied to a single commerce platform
• Attribution-friendly architecture through manifests and action-level metadata
• Reservation-backed flow that can bridge into Stripe human and machine checkout paths

Limitations

• Early-stage adoption and ecosystem participation
• Still dependent on external payment/protocol interoperability for full flow coverage
• Advanced capabilities (trading, messaging, deeper analytics) are roadmap-dependent

BRUZEN The Fold

BRUZEN focuses on semantic decisioning: mission-adaptive product reasoning rather than keyword matching or static vector similarity.

Strengths

• Explainable, mission-based ranking logic
• Better handling of trade-offs across price, quality, and reliability
• Potentially stronger upstream recommendation quality

Limitations

• Discovery/reasoning layer only; no native payment execution
• Must be paired with checkout and authorization protocols to complete value capture

Structural gaps that define the next wave

Despite fast progress, five systemic gaps still constrain the market:

1. Universal merchant coverage: most protocols require explicit opt-in, leaving the long tail invisible.
2. Data-to-checkout continuity: product data stacks and payment stacks are still weakly connected.
3. Attribution blind spots: merchants cannot reliably measure conversational influence before checkout.
4. Monetization realignment: take-rate economics are shifting value from ad inventory to agent gateways.
5. Reasoning quality: retrieval alone is insufficient; mission-aware ranking is becoming table stakes.

What matters strategically in 2026

For operators, the practical takeaway is not "pick one protocol and wait." It is to build for optionality across three layers:

• Discovery readiness: structured product data, high-integrity attributes, robust review signals, and machine-readable brand knowledge.
• Execution readiness: API-first checkout, inventory freshness, consentable payment flows, and auditable order state.
• Measurement readiness: influence-aware attribution models that track surfaced -> shortlisted -> selected -> purchased pathways.

The firms that win agentic commerce will not just process transactions. They will make their catalogs legible to agents, measurable to operators, and executable across heterogeneous protocol ecosystems.

Closing view

Agentic commerce is not one product launch or one protocol standard. It is the convergence of discovery infrastructure, reasoning systems, identity/consent rails, and checkout execution.

Stripe, OpenAI, Google, and Shopify are all building critical parts of the stack. New entrants like MoltBillboard and BRUZEN are exploring the open gaps in discovery, attribution, and semantic decisioning.

The next leg of value creation will come from closing the visibility gap and connecting reasoning to reliable action across the open merchant web. The prize is large, the market is moving fast, and the architecture decisions made in 2026 will likely determine who controls demand in 2030.