Presented by SAP
Marketing sits at the intersection of generative content, real-time decision-making, and customer data.
It’s where the gap between signal and action shows up fastest — and most measurably — in spend, conversion, and whether customers feel understood or ignored. That makes it a natural proving ground for what enterprise AI actually looks like when it crosses from experimentation into operation.
The pressure to cross that line is real. According to SAP Engagement Cloud data, 80% of brands see AI as essential to retaining customers, and 34% identify real-time decisioning and automated content generation as AI's greatest value. Yet more than half still cannot access real-time or usable data, and fewer than 20% of decision-makers say their teams are truly coordinated.
The problem isn't the tools or the intent, says Jessica Keehn, CMO of SAP customer experience. It’s the architecture that connects them.
"Customer data, operational data, and content creation have always lived in separate systems, owned by different teams who are making decisions independently but losing the context that would make those decisions better," Keehn says. "They want to share the data, but the time it takes to do that pulls focus away from their key roles. The outcome is that the data gets lost."
Why AI-driven personalization keeps falling short
The persistent divide between what brands believe they are delivering and what customers actually experience comes down to that data architecture. That gap exists because solving it requires connecting two things that have almost never lived in the same place: the operational truth of a business and everything that business knows about its customers. Operational truth is the live state of the business — inventory levels, order history, fulfillment status, supply chain signals — and what the company can actually deliver in that moment, at what cost. Customer knowledge is the accumulated picture of a person — browsing behavior, purchase history, loyalty status, service interactions — and what they’ve told you (explicitly and implicitly) about what they want.
And when the systems that hold enterprise data are not connected to the systems responsible for acting on it, engagement decisions are made without complete context. Even organizations with sophisticated martech stacks are optimizing in fragments, and customers feel it, especially as expectations rise.
Keehn points to Molton Brown, the luxury fragrance retailer, as an example of a brand operating with complete context. When a customer uses the brand's online fragrance finder, that preference data flows into SAP and becomes part of a unified customer profile. When that same customer walks into a boutique and scans their loyalty profile, the store associate already knows what fragrances they have browsed and what they have purchased.
"They can have a tailored conversation in the store about what matters to that customer, because the online and in-store experiences become a single continuous journey rather than two interactions starting from scratch," Keehn says.
How enterprise AI moves from informing decisions to executing them
AI agents are changing the game, because they don't wait for a human to interpret data and decide what to do. Instead, they reason over governed enterprise data and carry out workflows autonomously, continuously adapting as conditions change. The implications for marketing are substantial.
"Instead of being a system of record, it becomes a system of execution," Keehn says. "Gen AI creates content and insights, operational data validates what's possible. Execution systems activate in real time. This creates a closed-loop system that's always learning and adapting."
It now becomes possible to act with certainty rather than speed, and that certainty comes from grounding execution in live, governed enterprise data rather than replicated or approximated signals.
When the inventory system, the fulfillment system, and the customer engagement system are all drawing from the same governed data layer, the decisions AI makes are based on what is actually true about the business in that moment. An agent can adjust a live campaign — removing or replacing depleted products based on real-time inventory signals — rather than relying on a delayed batch snapshot. A proactive service message about a shipment delay goes out before the customer thinks to check because the fulfillment signal is live, not lagged.
These are not incremental improvements on existing marketing workflows. They represent a different operating model, one where the boundary between the system that holds enterprise data and the system that acts on it has effectively dissolved.
How SAP and Google Cloud are building a real-time intelligence foundation
The expanded partnership that SAP and Google Cloud announced last month addresses the fragmentation problem at the architectural level. By deepening integration across SAP Customer Experience (CX) solutions and Google Cloud’s AI and data platforms, the two companies enable organizations to create, execute, and optimize marketing and engagement programs with AI, grounded in trusted enterprise data and delivered at scale.
At the foundation is SAP Business Data Cloud (BDC) Connect for Google BigQuery, announced in October 2025, which enables bidirectional, zero-copy data access between SAP systems and Google Cloud. Enterprise data can be activated by AI models in real time, without replication or latency, while remaining governed by SAP’s enterprise-grade security, privacy, and compliance controls.
Building on this shared data foundation, SAP and Google can support agentic interoperability and multi-step agentic orchestration across platforms. Through SAP’s Agent Gateway APIs, SAP Joule agents can securely coordinate with AI agents running on Google Cloud — exchanging context, triggering actions, and optimizing outcomes across systems.
This enables multiple AI agents to work together across data, content creation, and execution, while remaining aligned with enterprise governance, because the value of agentic AI depends directly on the quality and trustworthiness of the data it acts on. And in marketing, an agent making real-time decisions about marketing spend, inventory substitution, or customer messaging needs to be operating on data that is accurate, governed, and current. The zero-copy architecture ensures that the data AI agents access is the same data the business runs on, not a replica that may already be out of date, offering what Keehn calls complete context.
"We're creating a real-time understanding of both the customer and the business, in a single governed environment," Keehn says. "With complete context, you can pivot in real time when inventory on promotion is depleted, for instance, without manual intervention."
In SAP Engagement Cloud, this is already showing up in production. A marketer defines an objective: increase repeat purchases among high-value customers in the next 30 days. SAP Joule agents analyze live enterprise signals. Google AI agents, powered by Vertex AI and Gemini, generate tailored messaging, imagery, and creative variations. The agents deliver personalized experiences across email, messaging, and rich channels including RCS, and as customers engage, performance signals flow back into the system. Creative, timing, and targeting refine continuously. While marketing is the first use case, the architecture extends across the full SAP CX portfolio, including commerce, service, and industry-specific customer journeys.
For marketers, the shift is one of elevation rather than displacement. "We want to keep people doing the fun work and make sure the data analysis and insights and connectivity is happening in the back end," Keehn says. "I want to do the creative stuff. I want to have fun. I want to tell stories. I want to create things. But I want to do it better, and I can do it better when I have more data and insights."
Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.