Artificial intelligence is changing how organizations build software, analyze data, and make decisions. Large language models can generate SQL, analytics platforms can answer questions in natural language, and business users increasingly expect AI to deliver immediate, trustworthy insights. Yet as AI adoption accelerates, many organizations are discovering that the biggest obstacle is not the technology itself. It is the lack of consistent business meaning across the enterprise.
ER/Studio 21.1, available today, addresses this challenge by helping organizations transform Enterprise Logical Data Models into reusable semantic assets that support governance, analytics, and AI. The release also introduces automated semantic layer generation for Microsoft Power BI, Open Semantic Interchange (OSI), and dbt, while expanding platform support to SQL Server 2025 and IBM Db2 z/OS 13 and LUW 12. Together, these capabilities help organizations establish a semantic backbone that preserves business meaning from architecture through implementation, analytics, and AI.
Every organization depends on a common understanding of its data.
Products.
Assets.
Suppliers.
Facilities.
Employees.
These concepts appear throughout operational systems, analytics platforms, governance catalogs, and now AI applications. Although the technology surrounding enterprise data has changed dramatically over the past decade, one requirement has remained the same: everyone needs to agree on what these concepts actually mean.
Unfortunately, maintaining that shared understanding becomes increasingly difficult as organizations grow.
New business applications are introduced. Cloud platforms are adopted. Departments build their own analytics environments. Business units develop reporting tailored to their specific needs. Governance teams document definitions in one platform while data engineers implement technical structures somewhere else.
Over time, these efforts begin to diverge.
A business term documented in a governance catalog no longer matches the implementation in a data warehouse. A metric calculated in one dashboard differs slightly from the same metric in another. Analytics teams recreate semantic models because no governed version exists to build upon. AI applications inherit whatever definitions happen to exist in the datasets they receive.
Individually, these inconsistencies may appear minor.
Collectively, they undermine trust.
Business users begin questioning reports instead of acting on them. Data teams spend valuable time reconciling conflicting definitions instead of delivering new capabilities. AI systems produce answers that sound convincing but are based on inconsistent interpretations of enterprise information.
This challenge has become increasingly common as organizations invest in modern data platforms. While technologies for storing, moving, and analyzing data continue to improve, preserving business meaning across those technologies has become significantly more difficult.
That is where enterprise data architecture becomes essential.
Organizations often search for new technologies to improve AI readiness, but many already possess one of their most valuable assets.
The Enterprise Logical Data Model.
An Enterprise Logical Data Model does much more than describe database structures. It captures the business concepts that define an organization, the relationships between those concepts, the rules that govern them, and the terminology that allows people across departments to communicate consistently.
In many ways, it represents the organization’s shared understanding of its own business.
For decades, these models have served as blueprints for database design. Architects used them to guide implementation, communicate requirements, and ensure consistency across systems. Once the database was built, however, much of that business knowledge remained inside the modeling environment.
Meanwhile, governance teams maintained business glossaries elsewhere. Analytics teams recreated business definitions inside reporting platforms. Data engineers implemented metrics inside transformation frameworks. AI applications attempted to infer business meaning from operational data alone.
The information existed.
It simply wasn’t connected.
ER/Studio 21.1 changes that by extending Enterprise Logical Data Models beyond architecture and transforming them into operational semantic assets that can be shared across the enterprise.
The central theme of ER/Studio 21.1 is simple.
Business meaning should be defined once and reused everywhere.
Rather than allowing every governance platform, analytics tool, semantic layer, and AI application to establish its own interpretation of enterprise concepts, organizations can use their Enterprise Logical Data Model as the authoritative source of business meaning.
This is the foundation of the Semantic Backbone strategy.
Instead of treating enterprise architecture as documentation that guides development, ER/Studio enables organizations to operationalize their architecture by distributing governed business meaning throughout the technologies that depend on it.
This approach creates several important advantages.
Architecture teams remain the stewards of enterprise definitions while governance platforms inherit approved terminology directly from the source. Analytics teams begin with standardized business concepts instead of recreating them from scratch. AI applications receive consistent semantic context that improves the quality and reliability of generated responses.
Most importantly, organizations reduce semantic entropy by ensuring that enterprise meaning evolves from a single governed foundation rather than multiple disconnected implementations.
At the heart of ER/Studio 21.1 is the new Semantic Generator, a capability designed to unlock the value already contained within Enterprise Logical Data Models.
Logical models have always described far more than entities and attributes. They contain business definitions, relationships, identifiers, constraints, and contextual metadata that explain how an organization understands its information. That context is exactly what governance platforms, knowledge graphs, semantic layers, and AI systems need to produce consistent, trustworthy results.
The challenge has been making that information available in a format those technologies can consume. The Semantic Generator solves this by automatically transforming Enterprise Logical Data Models into standards-based semantic assets, including RDF (Resource Description Framework), SHACL constraints, and SKOS mappings.
RDF is a W3C standard for representing information as interconnected concepts and relationships in a machine-readable format. In practical terms, the Semantic Generator takes the business knowledge captured visually in an ER/Studio data model and translates it into a semantic representation that other technologies can consume. A model may define concepts such as Customer, Product, Order, and Supplier, what those concepts mean, and how they relate to one another. RDF makes that same knowledge available in a structured form that AI and LLM-based applications can use as part of their grounding and context.
This is particularly important for AI because access to enterprise data alone does not necessarily provide an understanding of what that data means. Rather than forcing each AI initiative to infer or recreate business context, organizations can provide governed definitions and relationships directly from the Enterprise Logical Data Model. Because everyone is working from the same modeled definitions, that context remains more consistent across AI, analytics, governance, and other downstream applications.
This represents a significant evolution in the role of enterprise data modeling.
Instead of ending with database design, Enterprise Logical Data Models now become active participants in governance, analytics, and AI initiatives. Business meaning no longer remains confined to architecture diagrams. It becomes an operational asset that can be shared across the enterprise while maintaining consistency with the original governed design.
For organizations investing in AI, this capability is particularly significant. AI systems cannot reliably distinguish between competing business definitions unless those definitions are explicitly provided. By generating semantic assets directly from governed enterprise models, ER/Studio helps ensure that AI applications are grounded in the same business understanding used by architects, governance teams, and analysts.
That creates a stronger foundation for trustworthy analytics, more reliable AI responses, and enterprise-wide semantic consistency.
Creating machine-readable semantic assets is only part of the story. Their real value comes from extending that governed business meaning into the platforms where employees analyze data, build reports, and increasingly interact with AI.
For many organizations, this is where semantic consistency begins to break down.
Analytics teams often create semantic models independently for each reporting platform. Business definitions are rewritten to meet project deadlines. Metrics evolve differently across departments. New dashboards inherit local interpretations of enterprise concepts instead of governed definitions established by data architects.
The result is duplicated effort and inconsistent business meaning.
Over time, reports that should answer the same question begin producing different results. Business users lose confidence in dashboards because key metrics appear to change depending on which report they open. AI assistants generate conflicting responses because they are grounded in different semantic models.
ER/Studio 21.1 helps eliminate these challenges by extending governed enterprise semantics directly into the analytics platforms organizations use every day. Rather than asking each analytics team to recreate business meaning from scratch, ER/Studio enables them to build upon the same Enterprise Logical Data Models that already define the business.
Microsoft Power BI has become one of the most widely adopted analytics platforms in the enterprise, particularly as organizations embrace Microsoft Fabric and AI-powered experiences such as Copilot and Power BI Q&A.
These capabilities make analytics more accessible than ever before, but they also increase the importance of a well-designed semantic layer.
Natural language queries depend on consistent business definitions. If different workspaces define the same metric differently, AI cannot determine which interpretation is correct. It simply answers using the semantic model it has been given.
ER/Studio 21.1 addresses this challenge by introducing the Microsoft Power BI Semantic Layer Generator.
Instead of manually recreating semantic models inside Power BI, organizations can generate complete semantic layer artifacts directly from analytics-ready physical star schemas. Business names, descriptions, and metadata inherited from the Enterprise Logical Data Model are carried forward automatically, helping ensure that reports, dashboards, and AI experiences all operate from approved enterprise terminology.
This capability delivers benefits that extend well beyond development productivity.
Analytics teams spend less time building semantic models manually.
Governance teams gain greater confidence that enterprise-approved terminology is being used consistently.
Business users encounter the same definitions regardless of which report or dashboard they access.
AI-powered analytics become more trustworthy because they are grounded in governed enterprise meaning instead of locally interpreted business concepts.
Most importantly, enterprise architecture becomes an active part of the analytics lifecycle instead of ending once the database has been designed.
Enterprise analytics rarely consists of a single platform.
Organizations often use Power BI alongside cloud data warehouses, data transformation frameworks, governance catalogs, and other business intelligence tools. Every platform introduces another opportunity for business definitions to drift if semantic layers are created independently.
ER/Studio 21.1 addresses this challenge by expanding semantic generation beyond a single analytics ecosystem.
Open Semantic Interchange, or OSI, represents an emerging industry effort to improve semantic interoperability between analytics platforms.
Rather than maintaining proprietary semantic models for every reporting environment, organizations can generate portable semantic assets that support multiple downstream technologies.
The new Open Semantic Interchange Generator enables ER/Studio to generate OSI semantic layer definitions directly from governed enterprise models while preserving business names, descriptions, and metadata inherited from the Enterprise Logical Data Model.
This provides several long-term advantages.
Organizations can establish business meaning once and distribute it across multiple analytics initiatives without recreating metrics or dimensions for every platform. As new technologies emerge, semantic assets remain portable, reducing vendor lock-in while helping preserve enterprise consistency.
It also supports one of the central goals of the Semantic Backbone strategy.
Business meaning should travel with the enterprise, not remain tied to individual technologies.
The growth of modern data engineering has made dbt a key component of many enterprise analytics environments. Data teams rely on dbt to transform warehouse data into trusted business metrics, but semantic definitions are frequently recreated during that process.
Descriptions are rewritten.
Relationships are reinterpreted.
Business terminology gradually diverges from enterprise architecture.
ER/Studio 21.1 introduces the dbt Semantic Layer Generator to help eliminate this duplication.
Analytics teams can automatically generate dbt Semantic Layer artifacts from ER/Studio models while preserving governed business definitions, relationships, and metadata inherited from the Enterprise Logical Data Model.
Rather than rebuilding enterprise knowledge inside every dbt project, organizations can extend approved business meaning directly into modern analytics engineering workflows.
This creates greater consistency between operational systems, data warehouses, semantic layers, dashboards, and AI applications while reducing the manual effort required to maintain them.
It also reinforces a broader principle behind ER/Studio 21.1.
Enterprise architecture should not stop at implementation.
It should continue delivering value throughout the entire data lifecycle.
While the semantic capabilities introduced in ER/Studio 21.1 represent the strategic direction of the release, enterprise customers also expect their modeling platform to evolve alongside the databases that power mission-critical applications.
Modernization projects often stall when modeling tools fail to support the latest platform capabilities. Architects are forced to document new database features outside the model, manually modify generated DDL, or postpone upgrades until tooling catches up.
ER/Studio 21.1 helps remove those obstacles.
The release introduces support for Microsoft SQL Server 2025, allowing organizations to continue modeling, reverse engineering, and generating DDL while upgrading to Microsoft’s latest database platform. Existing modeling investments remain protected, reducing migration risk and helping architecture teams remain aligned with production implementations.
ER/Studio 21.1 also adds support for IBM Db2 z/OS 13 and IBM Db2 LUW 12, including new platform capabilities that improve deployment accuracy and documentation of production environments. Support for advanced indexing features and security-related trigger behavior enables organizations to keep enterprise data models synchronized with modern Db2 implementations while reducing manual post-generation changes.
These enhancements continue ER/Studio’s long-standing commitment to helping enterprise customers modernize confidently while maintaining governance, consistency, and architectural integrity across evolving technology platforms.
The Semantic Generator transforms Enterprise Logical Data Models into machine-readable semantic assets that can be consumed by governance platforms, knowledge graphs, semantic layers, and AI systems. By generating standards-based outputs such as Turtle RDF, SHACL constraints, and SKOS mappings, organizations can extend governed business meaning beyond data modeling and establish a trusted semantic foundation across the enterprise.
ER/Studio 21.1 helps improve AI readiness by transforming governed business knowledge from Enterprise Logical Data Models into machine-readable semantic assets, including RDF. These assets provide AI and LLM-based applications with consistent business definitions, relationships, and context rather than relying solely on technical database structures. By grounding AI in the same business meaning used across the organization, ER/Studio helps support more consistent, accurate, and trustworthy AI responses.
ER/Studio 21.1 introduces a Microsoft Power BI Semantic Layer Generator that automatically creates Power BI semantic models from analytics-ready physical star schemas. Business names, descriptions, and metadata inherited from the Enterprise Logical Data Model are carried into Power BI, helping organizations maintain consistent business definitions across reports, dashboards, Microsoft Fabric, Copilot, and Power BI Q&A.
ER/Studio 21.1 adds support for Microsoft SQL Server 2025 as well as IBM Db2 z/OS 13 and IBM Db2 LUW 12. These updates ensure organizations can continue modeling, reverse engineering, and generating DDL while modernizing their database environments and taking advantage of the latest platform capabilities.
The Enterprise Logical Data Model serves as the authoritative source of business meaning by defining enterprise concepts, relationships, business rules, and terminology independently of any specific technology platform. In ER/Studio 21.1, these models can now be transformed into reusable semantic assets that support governance, analytics, AI, and semantic interoperability, helping organizations establish a consistent semantic backbone across their entire data ecosystem.
