The Agentic Future Demands an Open Semantic Layer

The Agentic Future Demands an Open Semantic Layer

Salesforce is making a first-of-its-kind commitment with an alliance of industry leaders to build towards an open, interoperable future required to accelerate agentic AI.

Share article

The world of enterprise technology is experiencing a seismic shift. For decades, analytics revolved around outputs like dashboards and reports as the primary vehicle for insight. But the rise of agentic AI-in which agents can autonomously analyze, reason, and act-requires us to realign our focus toward something far more fundamental: the semantic data model.

That's why Salesforce is making a first-of-its-kind commitment with an alliance of industry leaders-to build towards an open, interoperable future that the AI era demands. And with a deeply integrated technology stack built to connect semantic definitions directly to tangible business outcomes, we are uniquely positioned to help lead this charge.

The 'data meaning disconnect' is a critical blocker to AI adoption

Accenture predicts that by 2030, AI agents will become the primary users of enterprise software, but these new intelligent workers are only as effective as the data they can access and understand. Agents need clear and trusted semantic definitions to translate intent and to produce accurate and relevant outputs-whether that's data visualizations, recommended next steps, or even taking autonomous actions. Without this robust core to ground its analysis in consistent, accurate data meaning, we risk our investments in AI producing noise rather than trusted insights.

As companies race to embrace AI, many are hitting a wall I call the "data meaning disconnect." While data may be unified, its meaning is not. A single business term like "customer churn" or "active lead" often has multiple, conflicting definitions across different systems and teams. A customer satisfaction score (CSAT) can be defined on widely different scales with varying definitions of "good." This problem is compounded by "semantic sprawl," where different dashboards have slightly different logic for the same metric-like one report defining a week as starting on Monday and another on Sunday.

These inconsistencies may seem small, but their impact is enormous. They erode the most critical asset for any data-driven organization: trust. When leaders see conflicting numbers, they lose confidence and hesitate to make decisions. This forces data teams to spend hundreds of hours manually reconciling business logic across siloed platforms-a major operational bottleneck that also adds enormous complexity. These redundant efforts stall innovation and undermine the reliability of the very AI models meant to accelerate business outcomes.

In the agentic era, the semantic model is the true center of gravity for AI and analytics; it is the foundational "source code of business understanding." However, this foundation is often locked into a single vendor's stack. This creates another data silo that forces customers to constantly rebuild business logic in new environments, and prevents AI from scaling across the best-of-breed tools that businesses rely on.

Introducing the Open Semantic Interchange (OSI)

To solve this foundational challenge, Salesforce is proud to co-lead the Open Semantic Interchange (OSI) alongside industry leaders like Snowflake, dbt Labs, and more. The aim of this initiative is to create a common, vendor-neutral specification that ensures context, meaning, and data results are preserved across every platform our customers use. Our goal is simple: business meaning should be defined once and understood everywhere.

The OSI will work to establish a shared standard so that all tools can "speak the same language," creating interoperability across tools and giving companies the flexibility to adopt best-of-breed technologies without losing consistency. In this effort, Tableau's initial focus is on a few critical capabilities:

• Bi-directional metadata exchange: We are standardizing how metrics, dimensions, hierarchies, and relationships are exchanged between platforms. This allows business logic to flow seamlessly without manual work or redundant recoding.
• Seamless governance propagation: We are ensuring that the access controls and data lineage defined in a source semantic layer are automatically enforced and respected across all participating systems. This provides a consistent layer of trust and security, everywhere.
• Native query logic: We are preserving query integrity by using the source platform's native runtime to generate results. This guarantees that business logic is executed as intended by the source system, ensuring consistency between platforms.

The Salesforce advantage: An open ecosystem built on trust

At Salesforce, we believe the future of data is open and interoperable, not locked into a single vendor's stack. A semantic layer that only works within one ecosystem simply creates another data silo, defeating the entire purpose. This initiative is a natural extension of our open philosophy, giving customers the freedom to use the best tools for their needs without being forced to constantly rebuild business logic in every new environment.

Our position in the market provides a unique advantage to help drive this change.

• With decades of domain expertise across sales, service, marketing, and commerce, we can provide customers with pre-built semantic models that serve as powerful starting points, accelerating their journey to agentic readiness.
• Our deep integration across the full technology stack-from Data Cloud, Agentforce, and Tableau to operational applications like Sales Cloud, Service Cloud, Marketing Cloud, and more-allows us to connect semantic definitions to tangible business outcomes. We can train AI not just on user clicks, but on whether a recommended action led to a successful sale or renewal, creating a powerful feedback loop that attributes real value to analytical work.
• Our commitment is to flexibility. Salesforce Data Cloud is designed to act as a data orchestration layer, allowing customers with mature investments in platforms like Snowflake, Databricks, Google, AWS, and others to leave their data at rest while still leveraging our semantic capabilities through zero-copy integration.

The future is a collaboration-continue the journey with us

Successful businesses will be powered by a partnership between humans and their AI counterparts. For this collaboration to succeed, it requires the robust and open semantic foundation we are building today. The Open Semantic Interchange is our commitment, alongside our partners, to helping every customer build that foundation and confidently harness the power of trusted, scalable AI.

Share article

Just For You

Explore related content by topic

• Agentic AI
• Analytics
• Data Analyst

As Chief Product Officer, Southard sets product strategy and leads the product management teams for Analytics products at Salesforce, including Tableau, CRM Analytics, Operational Reporting, and Data Cloud Reporting.