Snowflake introduces new Horizon Catalog capabilities to manage and secure artificial intelligence across enterprise systems. This update aims to provide a single control framework for managing AI agents, applications, and data.
The changes include a new context layer called Horizon Context, added security features for AI systems, and tighter integration with Adaptive Compute, which Snowflake says will allow customers to manage fluctuating AI workloads without manual adjustments.
Snowflake is tackling a problem that grows as companies move AI projects from pilot stages to daily operations. Many companies struggle with maintaining consistency in data definitions across databases, analytical tools, and software agents while applying security controls designed for human users rather than autonomous systems.
Horizon Catalog is designed to serve as a central catalog for enterprise data and AI governance. With these additions, Snowflake aims to give data teams, business users, and AI agents access to the same business definitions and usage rules, reducing the risk of systems operating based on inconsistent interpretations of key metrics.
Christian Kleinerman, vice president of products at Snowflake, said the shift to more autonomous AI systems has changed governance and security requirements.
“When intelligence becomes autonomous, trust is no longer an afterthought but becomes fundamental,” said Christian Kleinerman, vice president of products at Snowflake.
“Organizations need AI that operates in a trusted business context with governance and security built in from the start. New advancements across the Snowflake Horizon Catalog provide every agent, app, and team with the trusted context and security controls they need to move AI from experimentation to real business operations,” Kleinerman said.
shared context
At the heart of this announcement is Horizon Context, designed to create a common layer of business meaning across your data assets. Business logic is often spread across SQL queries, dashboards, and AI tools, making it difficult to maintain a single agreed-upon definition for metrics such as revenue, Snowflake said.
The new approach integrates context from databases, data lakes, and business intelligence tools so internal teams and software agents can work from the same underlying definitions. It also includes Semantic Studio, which aims to help analysts define business logic without deep knowledge of SQL, and Semantic View Autopilot, which can automatically generate and adjust semantic views.
Snowflake said these semantic views and associated data agents can also be created for shared datasets, including datasets distributed through the Snowflake Marketplace. The company also supports the Open Semantic Interchange standard, which aims to enable business definitions to be used between external AI agents and third-party analytics tools.
According to Snowflake, BlackRock is one of its customers using Horizon Context to support common data models for AI and analytics.
“In the financial industry, trusted data and consistent business context are essential to providing accurate insights and managing risk across global markets,” said Jeff Miller, managing director and global head of data factories and enterprise data platforms at BlackRock.
“As AI becomes increasingly embedded across the enterprise, it is essential that applications, analytics, and agents operate from the same trusted understanding of the business. Snowflake’s Horizon Context helps us extend a consistent business definition across our broader data ecosystem, supporting more trusted and managed AI and analytics experiences at scale,” said Miller.
Focus on security
Snowflake is also adding security tools designed for AI agents. This reflects concerns that existing access controls do not scale well to systems that can capture, reason about, and act on data with limited human intervention.
The new agent identity feature allows software agents to obtain a verified identity before being allowed to access data or perform actions, Snowflake said. The company says this supports role-based permissions and generates an audit trail of agent activity.
Further additions to Snowflake Trust Center are aimed at enabling customers to monitor the security posture of their AI systems, investigate policy violations, and respond to risks with assistance shaped by the context of their AI workloads. Snowflake also said it uses centralized policy controls and machine learning-based detection to address threats such as data leaks, ransomware, and prompt injections.
Several customers are working with Snowflake to evaluate new security features, including Acxiom, NewDay, and Thomson Reuters.
said Ankur Jain, Chief Cloud and Data Modernization Officer at Acxiom. “As AI becomes increasingly embedded across the marketing industry, having the right security foundation in place is critical to helping our business responsibly scale innovation.”
“Snowflake’s new AI security capabilities have the potential to increase visibility and control over how AI systems access and interact with personally identifiable data, helping us scale AI adoption responsibly while maintaining the trust our clients expect,” Jain said.
Thomson Reuters also viewed the issue as one of surveillance as AI becomes more involved in professional work and customer-facing services.
“At Thomson Reuters, responsible AI adoption depends on strong security, visibility, and governance around how AI systems interact with corporate data,” said Caitlin Halferty, head of data and analytics at Thomson Reuters.
“As AI becomes more deeply embedded in professional workflows and customer experiences, it’s important to protect sensitive information while enabling innovation. Snowflake’s new AI security capabilities give us greater control and visibility, helping us scale AI with the trust, compliance and accountability that our customers expect.” Halferty said.
computing layer
Snowflake also says it can link Horizon catalog changes to adaptive computing to automatically allocate compute and software resources to AI and application workloads in real time. Just as companies seek to run increasingly unpredictable AI jobs at scale, the company argues that governance and security tools often create friction.
By marrying governance, security, and compute management, Snowflake aims to make its platform the center of enterprise AI operations, rather than just serving as a data warehouse or analytics environment. The announcement also reflects a broader shift among enterprise software suppliers to build control over so-called agent AI, where software systems are expected to operate with greater autonomy.
Snowflake says more than 13,900 customers around the world use its platform.
