SAP и Google Cloud развернули мультиагентную платформу для автоматизации ритейла

SAP and Google Cloud announced the joint deployment of an agent-based commerce architecture—a system in which multiple AI agents autonomously manage marketing and retail operations at the enterprise level.

AI as a condition for survival in retail

According to SAP research, 78% of companies identify artificial intelligence as an essential tool for customer retention in 2026. A few years ago, such consensus in the corporate environment was unthinkable—now it is a strategic priority for most. However, a significant gap remains between intentions and real infrastructure.

The study identifies a paradox: fewer than two in five companies systematically exchange customer data within their own systems. Only 37% of organizations share information across customer experience channels, and just 39%—across CRM platforms. This means that an AI agent analyzing customer behavior in an online store often has no knowledge of their offline purchase history—and vice versa. Personalization under these conditions inevitably remains superficial.

This situation resembles the early days of omnichannel retail, when sales data lived separately from loyalty programs, and those—separately from the warehouse. AI makes the fragmentation problem even more acute: the more accurate the algorithm, the more it suffers from incomplete data.

How agent architecture works

The partnership is built on the concept of multi-agent systems. Instead of a single monolithic tool, several specialized agents take on specific tasks:

• Demand forecasting agent analyzes sales, seasonality, and external market signals
• Personalization agent shapes offers for each customer segment
• Inventory management agent synchronizes logistics in real time
• Marketing campaign agent automatically launches and optimizes communications
• Customer service agent handles inquiries and escalates complex cases

Agents operate on Google Cloud infrastructure and are integrated with SAP's corporate systems—ERP, CRM, supply chain management platforms. The key goal is to eliminate information gaps: the customer experience agent receives CRM data in real time, not through manual integrations with delays.

Why this matters for business

Agent architectures change the very logic of automation. Classical RPA tools executed pre-written scenarios without the ability to adapt. Modern AI agents can adjust to new conditions, make decisions in non-standard situations, and coordinate actions with each other without operator involvement.

For retail and e-commerce, this means a fundamentally different scale of operational efficiency. Personalization stops being a task for analysts working with weekly reports: agents process data and adjust offers continuously. Early experience with agent systems shows conversion growth of 15–30% with proper configuration. The fundamental difference from traditional AI recommendations is speed of response: an agent notices a change in customer behavior within minutes, not days.

According to analysts' forecasts, the corporate AI software market in retail will exceed $20 billion by 2028. SAP, which controls a significant share of the global ERP market, and Google with its cloud AI tools are obvious contenders for leadership in this segment.

What this means

Multi-agent systems are transitioning from experimental phase to industrial application. The SAP-Google alliance is a market signal: business competitiveness is increasingly determined not simply by the presence of AI tools, but by their architecture. Fragmented point solutions are giving way to coordinated platforms where agents work as a unified operational team.

For competitors—Microsoft, Oracle, Salesforce—this alliance increases pressure to develop their own agent products. The battle for corporate AI in retail is just beginning.