SAP, Google Cloud unveil agentic commerce architecture

SAP and Google Cloud announced an agentic commerce architecture that connects SAP Commerce Cloud, Google BigQuery and Google Gemini to automate marketing, shopping and fulfilment workflows. The architecture standardises data exchange so software agents can execute search, transaction processing and post-sale tasks directly against enterprise backends.

The partners framed the effort as a response to data fragmentation in enterprise commerce. SAP research found 78% of businesses view AI as essential to retaining customers by 2026, while 37% share customer data across customer experience platforms and 39% across CRM systems.

At the centre of the architecture is the Universal Commerce Protocol, which SAP Commerce Cloud has adopted to standardise information exchange among retailers, payment gateways and autonomous agents. Engineering teams integrate the protocol so intelligent agents can interact directly with commerce platforms, reducing integration work and speeding onboarding into AI-driven channels. Merchants will not need to rebuild existing backend systems for the protocol to handle inventory checks, cart management and payment processing.

SAP Commerce Cloud integrates Google Gemini capabilities to power a Shopping Assistant for chat, voice and text engagements. The assistant retains session state throughout the shopping cycle and takes in live behavioural inputs, warehouse capacities and active marketing data to assemble product recommendations and event configurations. Before presenting an item, the assistant queries live warehouse records to verify availability.

Data flows route through SAP Business Data Cloud Connect into Google BigQuery via a bidirectional, zero-copy link secured by administrative controls. Keeping data in place rather than duplicating it reduces storage duplication and lowers network latency. BigQuery consumes live variables such as weather, location and advertising interaction rates, while SAP Customer Experience systems supply internal context including customer profiles, transaction histories, service interactions and consent records.

Autonomous agents running in SAP Engagement Cloud orchestrate personalised interactions across the customer lifecycle based on those inputs. Generative models from Google Gemini, including the Nano Banana 2 iteration, create localised marketing output: messaging, customised imagery and campaign variants that update in response to incoming engagement data. Messages can be delivered through enriched messaging channels such as Rich Communication Services.

Marketing teams set business goals and grant enterprise data access; the autonomous agents handle audience segmentation, content generation and campaign execution using analytics from BigQuery. Consented engagement and transaction data captured during interactions feed back into SAP Customer Experience solutions to update local customer profiles, providing fresh context for future model outputs and unified records for fulfilment or service resolution.

The deployment addresses operational failures common in digital commerce, including fragmented APIs, delayed fulfilment updates and disconnected support channels. By linking engagement, commerce and fulfilment, the system synchronises frontend experiences with backend warehouses and maintains a continuous data-linked chain from initial purchase intent through post-sale resolution.