AI hallucinations pose a significant barrier to the use of AI by companies. According to Salesforce, more than 80% of enterprise AI projects never progress beyond the demo phase because the systems are not based on the company’s data, definitions and processes. Without that context, agents misinterpret signals like “customer ID,” “order status” or “return request,” leading to poor insights or incorrect actions.
Dig deeper: 5 ways AI changed marketing strategy in just one year
“AI without context is just guessing or hallucinating,” said Rahul Auradkar, EVP and GM of Salesforce’s Data 360. “By combining Informatica’s metadata with MuleSoft’s real-time signals, we’re replacing guessing with reasoning. We’re giving AI the foundation it needs to operate securely – and ensuring that when an agent acts, it does so with the full weight of business truth behind it.”
Creating context across three layers
Salesforce’s new unified data engine is built on three pillars:
1. Business insight via metadata: Data 360 now includes Informatica’s Master Data Management (MDM) capabilities, which include entities such as products, suppliers, and assets. That gives agents a shared vocabulary and understanding, so when an AI model sees “SKU-123,” it knows it’s the same as “Part A” in another system. The integration also includes data lineage features, allowing AI to verify the freshness and reliability of data. Combined with a comprehensive data catalog that spans on-premise, cloud and legacy systems, this all provides an in-depth map of the organization’s data landscape.
2. Real-time awareness via MuleSoft: MuleSoft feeds live operational signals into the system – everything from shipping delays to inventory changes and customer actions. AI agents can then act based on the current state of affairs in the business, and not yesterday’s snapshot.
3. Unified context with zero-copy architecture: Data 360 becomes the unified memory layer, merging Informatica’s historical context and MuleSoft’s real-time signals. This context is shared between AI agents using a zero-copy approach, meaning data is accessible but not duplicated, reducing storage costs and latency.
This new data layer forms the foundation for Salesforce’s broader Agentforce 360 platform – a four-tiered architecture designed to support enterprise-grade AI agents. If everything works as promised, this AI allows agents to execute workflows like refunds, order adjustments, or personalization with the same clarity as seasoned employees.
What matters to marketers
For marketers, this is not just a back-end IT story. Unified, trusted data is the key to making AI agents truly useful for campaigns, journeys and customer interactions. With access to real-time status, clear definitions, and verified customer records, marketing agents can recommend content, adjust targeting, and personalize messages without straying off course.
Dig deeper: How to prevent AI decisions from replicating human biases
It also lays the foundation for more confident and transparent AI governance – a crucial step as compliance requirements tighten and customers demand clarity about how their data is used.
Salesforce’s new data engine is available today as part of Data 360, Informatica, MuleSoft and the Agentforce 360 platform. For companies looking to deploy AI at scale, this could be one of the most critical tech updates of the year – not because it’s trendy, but because it solves the most persistent problem in business AI: lack of context.
Energize yourself with free marketing insights.
MarTech is owned by Semrush. We remain committed to providing high-quality reporting on marketing topics. Unless otherwise stated, the content of this page was written by an employee or paid contractor of Semrush Inc.
#Salesforce #engine #aims #enterprise #reliable #MarTech