Salesforce Commerce Cloud and AI Agents: Platform Guide for Merchants

Salesforce Commerce Cloud (SFCC, formerly Demandware) is the enterprise e-commerce platform within the Salesforce ecosystem. It powers some of the world’s largest retail brands, from luxury fashion houses to consumer electronics giants. With Salesforce’s heavy investment in AI through Einstein and Agentforce, SFCC is positioned to become one of the more AI-ready enterprise platforms - but today, most of the AI investment has been directed inward (personalization, merchandising) rather than outward (AI agent discovery). This guide covers where SFCC stands and what enterprise merchants should prioritize.

Platform Overview

Salesforce Commerce Cloud is a cloud-hosted platform designed for large enterprises. It comes in two architectures: SFRA (Storefront Reference Architecture, the traditional server-rendered approach) and Composable Storefront (a headless, React-based frontend formerly called PWA Kit). Many SFCC merchants are transitioning from SFRA to Composable Storefront, which has significant implications for AI readiness.

SFCC is tightly integrated with the broader Salesforce ecosystem - CRM, Marketing Cloud, Service Cloud, and the Data Cloud. This integration means SFCC merchants often have rich customer and product data flowing across systems, which could be leveraged for AI agent interactions if the right interfaces were in place.

The platform is exclusively enterprise. There is no free tier or self-service signup. SFCC merchants typically have dedicated development teams and system integrators managing their storefronts. This means that technical AI readiness improvements are feasible - but they require prioritization within enterprise development roadmaps.

Salesforce’s AI strategy centers on Einstein (now evolving into Agentforce). Einstein powers on-site search, product recommendations, and personalization within SFCC storefronts. Agentforce extends this to AI agents that can handle customer service, order management, and commerce tasks. However, Agentforce is primarily focused on agents acting on behalf of the brand (internal agents), not on making the brand discoverable to external AI shopping agents.

AI Agent Compatibility

SFCC does not natively support ACP, MCP, or UCP. External AI shopping agents interact with SFCC stores through:

1. Structured data on product pages. SFCC storefronts can output JSON-LD, but the implementation depends on the storefront architecture and customization.
2. Open Commerce API (OCAPI) and Shopper APIs. SFCC exposes APIs that could theoretically be consumed by AI agents, but they require authentication and are designed for internal or partner use.
3. Product feeds. Enterprise merchants typically maintain product feeds for Google Shopping, marketplaces, and affiliate networks. These feeds are accessible to AI systems.
4. Content delivery networks. SFCC stores often use CDNs that can affect how AI crawlers access content.

The gap between Salesforce’s internal AI capabilities and external AI agent discovery is striking. A SFCC store might have sophisticated Einstein-powered recommendations on-site while being poorly discoverable by ChatGPT Shopping or Perplexity because the structured data is incomplete.

For enterprise merchants, the opportunity is to leverage their existing product data infrastructure - product information management (PIM) systems, rich attribute data, comprehensive catalog metadata - and expose it in formats that external AI agents can consume.

Structured Data Support

SFCC’s structured data output depends heavily on the storefront implementation:

SFRA (Server-Side Rendered): The traditional SFRA architecture renders HTML on the server, making structured data relatively straightforward to implement. However, the default SFRA templates include only basic product schema. Enterprise implementations typically customize the templates to add comprehensive JSON-LD.

Composable Storefront (Headless): The React-based Composable Storefront gives merchants complete control over structured data output, but it requires explicit implementation. JSON-LD must be added to the React components, and server-side rendering (SSR) must be configured properly for crawlers to access it.

Key considerations for SFCC structured data:

• PIM integration. Many SFCC merchants use external PIM systems (Akeneo, Salsify, inRiver) that contain rich product data. Ensuring this data flows through to structured data output is critical.
• Multi-site, multi-locale. SFCC supports complex multi-site configurations. Each site and locale should output appropriately localized structured data with correct currency, language, and availability.
• Product sets and bundles. SFCC supports complex product types (master/variation, product sets, bundles) that need careful structured data handling.
• Custom attributes. SFCC’s product model supports unlimited custom attributes. These should be mapped to Schema.org properties where applicable.

Enterprise merchants with SFCC typically have the product data needed for comprehensive structured data - the challenge is getting it from the PIM/catalog system into the frontend JSON-LD consistently.

Protocol Support

Protocol	Status	Notes
ACP (Agentic Commerce Protocol)	Not supported	No native integration. Enterprise custom builds possible.
MCP (Model Context Protocol)	Not supported	Could be built on OCAPI/Shopper APIs. No standard integration.
UCP (Universal Commerce Protocol)	Not supported	No integration available. Google partnership could change this.
JSON-LD / Schema.org	Implementation dependent	Full control available, but requires custom development.
robots.txt	Full control	Configurable through Business Manager or CDN.
llms.txt	Manual	Can be added to the static content or web root.
ai.txt	Manual	Can be added to the static content or web root.

Salesforce’s relationship with Google could be relevant for UCP support. If Google’s Universal Commerce Protocol gains traction, Salesforce might prioritize SFCC integration given the enterprise market’s Google dependency. Similarly, Salesforce’s own Agentforce platform could eventually create a bridge between internal and external AI agents.

Optimization Checklist

• Audit your structured data implementation. Whether you’re on SFRA or Composable Storefront, verify that your product pages output comprehensive JSON-LD. Test across product types (simple, variants, bundles, sets).
• Map PIM attributes to Schema.org. Work with your integration team to ensure product attributes from your PIM system (brand, materials, dimensions, identifiers, certifications) flow through to structured data output.
• Implement GTIN/EAN identifiers in JSON-LD. Enterprise catalogs usually have these identifiers. Ensure they appear in your structured data, not just in your internal systems.
• Ensure SSR for Composable Storefront. If you’re using the headless architecture, server-side rendering is essential for AI crawler access. Verify that your JSON-LD is present in the server-rendered HTML, not just added via client-side JavaScript.
• Optimize your product feeds. Enterprise merchants typically maintain feeds for multiple channels. Review these feeds for completeness and ensure they include all product attributes AI systems need.
• Add llms.txt and ai.txt. Place these files in your web root. For SFCC, this may require adding them through Business Manager’s static content configuration or your CDN.
• Review CDN caching and crawler access. Enterprise SFCC stores often use aggressive CDN caching and bot management. Verify that AI crawlers (OpenAI, Anthropic, Perplexity) are not blocked or rate-limited by your CDN or WAF.
• Coordinate with your Salesforce partnership. If you have an enterprise Salesforce agreement, ask about AI agent discovery features on the SFCC roadmap. Enterprise customers can influence product direction.
• Implement hreflang and localized structured data. For multi-locale SFCC stores, ensure each locale outputs correct language, currency, and availability in its structured data. AI agents in different markets need locally accurate data.
• Monitor AI shopping surfaces systematically. Enterprise brands should track their presence in ChatGPT Shopping, Perplexity, Google AI Overviews, and Microsoft Copilot Shopping. Build this into your commerce analytics alongside traditional SEO monitoring.

Related Terms

• Structured Data - Machine-readable information that helps AI agents understand product pages on SFCC storefronts.
• Product Feed - Standardized product data exports that AI systems increasingly consume for product discovery.
• Semantic Search - AI-powered search that understands product meaning, relevant to how AI agents query product catalogs.
• AI Visibility Score - A measure of how discoverable your products are to AI shopping agents.
• Agentic Commerce - The model where AI agents discover and transact on behalf of consumers.