B2B companies ready for agentic AI discovery

B2B ecommerce companies need to prepare for a future where artificial intelligence agents discover products for shoppers, a shift that requires significant changes to how product data is structured and presented.

Agentic AI operates on a spectrum ranging from simple tools to fully autonomous systems. Paul do Forno, global commerce practice lead at Deloitte, noted that these systems are currently used to solve specific problems within larger digital transformations. Rather than offering one broad solution for every issue, successful agents tend to target friction points across different business processes.

For example, an AI agent can now handle the “availability to promise” function. A shopper might ask an agent to check if a product is available for delivery in a specific timeframe. The agent can then query various internal systems to confirm stock levels. If the requested item is unavailable, the agent can search for alternatives that meet the user’s criteria. This capability extends to reordering, where buyers often use email, PDFs, or purchase orders to place orders. An agent can now convert these documents into formal orders, handling a process that was previously manual.

B2B companies must build a core commerce platform or cloud infrastructure before rolling out these features. This foundational layer is necessary to connect with marketplaces and other distribution channels. Without this base, adding agentic capabilities to the tech stack is difficult.

Related: Reformation files for IPO targeting billion dollar valuation

Businesses are transitioning from search engine optimization (SEO) to generative engine optimization (GEO) as LLMs become the primary discovery method. The fundamental difference lies in focus: SEO relies heavily on keywords, while GEO requires understanding user intent. A buyer looking for materials to build a house needs the system to understand the project context, not just match a specific term.

Deloitte research indicates that companies are still catching up. In a late 2025 workshop, less than 24% of suppliers reported using agentic AI in their selling processes. This suggests a bifurcated setting where B2B operations are generally behind B2C adoption. The complexity of B2B product data, including regulatory standards and specific permutations for fitment, makes this transition harder.

Content is the primary vehicle for this optimization. A product might be part of a complex assembly requiring 12 different components. If a company fails to clearly associate these components through content generation, the AI agent will not surface the product during a search. Providing detailed scenarios and use cases helps agents understand how products function in real-world applications. This approach allows for more contextual queries, such as identifying a product suitable for outdoor services and associating it with an expert recommendation.