ai/ml
Data intelligence catalog producer Alation has launched its Alation Intelligence Operating System (AIOS) to combining data, context, and agents into a unified operating system that is open, governed, and self-improving.
Alation has been primarily known for its software catalogs or maps of customer’s data, reading metadata from sources such as file systems and databases via a host of connectors. It provides intelligence about the types of data, its location and properties, helps feed it to upstack suppliers such as Databricks and Snowflake, and also provides data governance facilities. Alation was founded in 2012 and has pulled in $340 million in funding. HPE contributed to its $123 million E-round in 2022, when Alation’s valuation was over $1.7 billion.
The company acquired Numbers Station AI in May last year, and gained software to build AI agents for data workflows. It said it would be able to build “a new class of AI-native analytics applications featuring agentic workflows that operate with enterprise-grade governance and context.”
For organizations to achieve what is truly possible with AI, they need systems that are trustworthy and deeply integrated with diverse enterprise data ecosystems. Alation said “Structured data—such as customer records, supply chain records, and financial transactions—represents an organization’s most critical assets. AI agents struggle to understand and take action on that data due to incomplete semantics and data definitions, unclear governance policies, and missing context of the lineage and data quality of the underlying data sets. Without this foundation, agents risk producing inaccurate outputs and violating privacy and compliance requirements.” By “combining Numbers Station’s agents with Alation’s rich metadata foundation, customers will be able to build intelligent applications that reason over structured data, understand business context, and automate real-time decision-making—all while maintaining rigorous governance and compliance standards.”
Alation CEO and co-founder Satyen Sangani said: “Scaling AI to tackle high-stakes business processes requires a foundation of trusted metadata, governed data products, and deep contextual understanding.”
As we understand it, AIOS is that foundation. It builds upon Alation’s previous developments, such as AI Governance giving enterprises a system of record for AI compliance, registering every AI model, agent, and tool in a single inventory, mapping them to applicable regulations, generating evidence-backed model cards, routing approvals through regulation-aware workflows, and producing a live compliance posture for admins when requested.
AIOS is designed to help enterprises govern AI as they build and deploy it across their most critical business operations. Alation says that, when software breaks, it produces an error. When an agent breaks, it produces a confident, incorrect answer, where most organizations have no system of record to catch it before that answer drives a decision. AIOS catches it, tackling multiple failure points by combining data, context, and agents into a unified operating system that is open, governed, and self-improving.
It is a system of agents, context, and data.
Data is the foundation and there is a catalog; one system of record for every data asset across the enterprise. An asset’s lineage and quality (trust) are managed. Curation produces data descriptions.
The Context layer turns cataloged data into business intelligence that agents can reason over. Definitions, rules, and relationships tell agents how your organization actually works. The Data Product builder helps you create governed data packaged into certified, reusable products, with lineage, contracts, and policy baked in before anything publishes. Products are found and consumed in the marketplace. Ontologies ensure the relationships between a cjustomer’s business concepts are governed so "customer" or "active account" means one thing across the whole company.
The semantics component is where Alation ingests semantic models from your existing BI tools and platforms, rather than forcing a rewrite.
The Agents layer has an Agent Studio for building agents that take action and proving they’re acting correctly with evaluations before they reach production.
Alation says AIOS is self-improving. Corrections, evaluations, and decision traces flow back into data, context, and agents constantly, across apps. When an agent gets something wrong, the correction routes to the right layer. Fixes improve every agent that touches that context, not just one output. All feedback, approvals, and issues go to one place, to help you prioritize what needs attention and direct AI to do better work on your behalf.
Data sources are linked through more than 120 connectors, Alation saying this means open interfaces and standards (MCP, OSI, OpenLineage, OWL, RDF, SHACL, SCIM, SAML, OAuth, OAS and ODPS and more) giving access to any model, any warehouse, any semantic layer, and any storage.
Customers can bring their own AI models to bear, GPT, Claude, open source or custom.
All of Alation’s previous developments have led up to AIOS. It’s been developing it for its hundreds of enterprise customers, 40 percent of the Fortune 500 for example. Learn more here.
Comment
Suppliers such as VAST Data and the Modern Data Company have also launched upper stack, AI-focused OS software products.
VAST Data’s AI OS is built on top of VAST’s existing AI Data Platform and provides services for distributed agentic computing and AI agents. It consists of:
Kernel to run platform services on private and public clouds
Runtime to deploy AI agents – AgentEngine
Eventing infrastructure for real-time event processing – Data Engine Data Engine
Messaging infrastructure
Distributed file and database storage system that can be used for real-time data capture and analytics – DataStore, DataBase, and Data Space
Modern’s DataOS is an operating system for data providing a set of tools and services to help Fortune 500-class enterprise data professionals manage and analyze data more effectively.
Alation’s AIOS, VAST’s AI OS, and Modern’s DataOS each provide a services-based platform for analytics and AI agents to use in their work without having to be concerned with lower-level software plumbing. Think of a server operating system, like Linux or Windows, having device drivers; software providing a standard way to read and write from disk drives, or send and receive IO from network ports. Without these operating system drivers application developers would each have to write their own code to access disk drives or send data across a network link. The server OS provides standard, low-level device plumbing to do this, freeing the developer to focus on higher-level work.
These AIOS, AI OS and DataOS products take the same approach, but higher up the application stack.