Snowflake Helps Enterprises Deploy AI Agents with Trusted Business Context and Built-In Controls

• Definitions Agents Can't Argue With: Business Ontology gives organizations one governed place to record what their business terms officially mean across their data estate, and serves only approved definitions to agents
• Every Agent Under One Centralized Control: Cortex AI Gateway extends Snowflake's security and governance capabilities to all agents to help teams govern and use right-sized models for managing budgets
• Proof Across the Enterprise: Industry leaders including Glean and Kargo rely on Cortex AI Gateway to govern every AI agent across their business, while companies like dentsu and Merkle use Business Ontology to ground those agents in definitions the business has approved.

SNOWFLAKE EXPEDITION 2026 – Nov. 3, 2026 – Snowflake (NYSE: SNOW), the AI Data Cloud company, today at Snowflake Expedition 2026 announced new innovations that put context, security, and governance at the foundation of every AI agent, not bolted on after the fact. Snowflake is interoperable by design, connecting an organization's entire data estate across catalogs and clouds, providing the universal governance controls and context that agents and analytics depend on. To bring that same governance and context to autonomous agents, Snowflake is bringing new capabilities to Horizon Context, the home for governed business context inside Snowflake Horizon Catalog.

Snowflake’s new Business Ontology³ allows these agents to operate with a shared, pre-approved understanding of core business definitions to enable reliable answers. While Cortex AI Gateway¹ governs how the agents access models, invoke tools, and take action across enterprise systems. As a result, every decision and operation remains secure, fully auditable, and bound to a single set of policies, on a single data copy, regardless of which cloud, catalog, or engine the data sits in.

Trusted and reliable agents are hard to build. Not because of model limitations, but because enterprise context is inherently fractured. According to a recent MIT Technology Review report sponsored by Snowflake, 55% of executives cite a lack of business context as the primary cause of low-quality AI agent insights, with fragmented or inaccessible data ranking second at 45%.

As these agents spread across the business, connecting to models, MCP servers, and enterprise systems, they’re multiplying faster than IT and security can keep up. Companies are left with an impossible choice between scaling AI quickly while absorbing inaccuracies, vulnerabilities, and runaway costs, and slamming the brakes on the very momentum they are trying to build.

Snowflake eliminates that tradeoff by giving organizations the confidence to move fast without losing control. With shared business context grounded in an organization’s entire data estate, active execution guardrails, and unified policies governed across every cloud and engine, teams no longer have to choose between agent velocity and enterprise security. The agentic enterprise runs on more than data access, it runs on data companies can govern, agents they can control, and answers they can trust.

"For decades, governance gave organizations the confidence to make decisions with correctness, security and compliance. In the agentic era, governance is no different, it promotes trust and accelerates AI adoption." said Christian Kleinerman, EVP of Product, Snowflake. "The enterprises that win won't be the ones with the most agents. They'll be the ones whose agents know exactly what the business means by accessing the right data, abide by governance and guardrails specified, and can account for actions taken by agents. By extending Snowflake Horizon Catalog with Business Ontology and Cortex AI Gateway, context and governance stop being afterthoughts and become the foundation every agent is built on for the agentic enterprise."

Business Ontology Grounds AI Reasoning in Governed Business Context
A dashboard gives one answer, an analyst gives another, and an AI agent gives a third. People stop knowing which number to trust. Every business definition needs an official version and an accountable owner, whether the number is going to a team meeting or to a regulator. With Snowflake's interoperable architecture, Business Ontology can define that version once and apply it across an organization's data inside and outside of Snowflake.

"The value of a data product depends on everyone agreeing what each attribute means, whether it's a household income band or a B2B firmographic segment," said Joe Tobey, Head of Data Product Engineering, dentsu. "As agents become how people query that data, those definitions can't live in a static data dictionary. Business Ontology gives every definition an accountable owner, links it to the data that implements it, and makes it the answer every agent returns. As early product adopters, this is what lets us put agents in front of our data with confidence.”

AI agents need three kinds of context to answer reliably, including what a business term officially means, how it is calculated, and the operational knowledge nobody has written down. A Business Ontology governs the official meaning, a Semantic View calculates it, and Cortex Sense³ infers what isn’t already defined. Together, this unified framework establishes one single, authoritative business context for every agent, app, and person across the enterprise.

Snowflake is introducing Business Ontology as the newest capability within Horizon Context, the home for governed business context inside Horizon Catalog. Business Ontology allows teams to approve official definitions for core concepts like net revenue and link them directly to underlying tables and views, ensuring definitions persist across every tool.

This new capability builds on how Horizon Context collects, enriches, and activates intelligence across the entire data estate. Out of the box connectors collect schemas, query logs, dashboard definitions, and usage signals from sources like Tableau, Power BI, dbt, SQL Server, and PostgreSQL. Horizon Context enriches this metadata with end to end lineage, automated tags, AI descriptions, popularity signals, and Apache Ossie (incubating) open standard Semantic Views, creating the foundational graph that Business Ontology governs.

From there, Snowflake CoCo, the coding agent for data teams, and Snowflake CoWork, the personal work agent for knowledge workers, use these approved business concepts to ground their responses. Powered by Cortex Sense, agents synthesize this governed context alongside real time usage signals, ensuring human approved ontology definitions always take precedence over competing inferences.

Organizations that already keep definitions in a third-party catalog or spreadsheet can migrate them to Snowflake in a single guided CoCo session, which pulls out domains, candidate terms, relationships, and linked data assets and marks everything as a draft for owners to approve. Teams can then maintain these definitions and mappings centrally as their business evolves, giving people and AI agents reusable, governed context without recreating it for each application. Every change is logged, so when an auditor asks what net revenue meant in Q2 and who signed off, the answer is a simple lookup.

"Semantic drift is one of the most underestimated risks in multi-partner data environments — and at Merkle, dentsu's data and technology agency, we operate inside some of the most complex ones in the industry," said Morten Lileng, Global Head of Engineering, dentsu. "Snowflake's Business Ontology closes that gap at the infrastructure level, giving every definition an owner and every downstream system a single governed answer. That's what makes the clean room and RMN architectures we build at Merkle, on Snowflake, reliable enough to act on."

Cortex AI Gateway Brings Every Agent Under One Set of Controls
Cortex AI Gateway is the single front door for agents in an organization. As a part of Horizon Catalog, Cortex AI Gateway applies one set of policies to agent connections across clouds, frameworks, and AI providers. Agents take action on enterprise data through MCP servers, and Cortex AI Gateway applies the company's access policies, authentication, and permissions to those calls. Agents inherit the controls already established for human users, removing the overhead of creating and managing a different set of operations. 

Governance has to be in place before agents reach production, not after an incident forces the issue. A single ungoverned connection or unexpected bill is enough to pause an entire AI program. With Cortex AI Gateway, users can:

• Run agent code securely without managing infrastructure: Cortex Sandboxes⁴ provides agents a secure, private space to run custom code right alongside company data in Snowflake, automatically following the exact permissions of the person running the agent. While Cortex AI Gateway controls the models and tools an agent can call, Sandboxes controls the code that agent actually executes. These environments launch in seconds with no servers to manage and shut down automatically upon completion, giving organizations complete control over autonomous agent workflows without exposing credentials or moving data off Snowflake.
• Detect agent overreach and insider misuse before it becomes an incident: Cortex AI Gateway's Agent Abuse Detection⁴ analyzes complete agent trajectories, not individual prompts, to catch agents operating beyond their permissions and employees using agents to bypass approvals or exfiltrate sensitive data. A two-stage engine pairs efficient screening with context-aware reasoning, so security teams receive explainable findings instead of alert noise. Snowflake also served as a technical design partner with OpenAI on Private Safety Processing (PSP)⁵ which will enrich dangerous agent activity detection analysis with privacy-preserving safety signals from OpenAI and improve its effectiveness.
• Keep a record they own and unlock deeper insights: Observability¹ records what every agent did, from the models it called to the tools it used and the decisions it made, in the customer's own Snowflake account rather than a vendor's. Agent Trace Intelligence⁴ allows teams to query traces against their business data to find agents wasting money in loops, tools that are failing, and the real cost of each completed task.
• Inherit protections already in place: Cortex Inference¹ runs inference on Snowflake's infrastructure across a broad and growing set of frontier and open source models, giving teams the right model for each task without separate vendor contracts or dedicated infrastructure. Cortex AI Guardrails³ applies account-wide runtime protection against prompt injection, jailbreaks, and novel attacks to every agent routed through the gateway, including third-party agents.
• Know what each agent costs, and cap it: Cost governance¹ attributes AI spend to any user, team, department, agent, or app, and enforces hard limits that stop runaway agents before they exceed their budgets. Dynamic Model Routing⁴ matches each request to an organization-approved model based on cost, accuracy, and latency, keeping routine work off the most expensive models.
• Check every tool call before it runs: With MCP governance through Cortex AI Gateway Tools², every agent tool call across cloud, Saas & private VPCs is authenticated and validated against fine-grained authorization policies before it executes autonomously or on-behalf-of an user with auditability of tool calls and identity propagation. Unauthorized actions, unapproved API calls, and compliance violations are blocked before an agent touches a sensitive system.

Enterprises including Glean and Kargo are already using Cortex AI Gateway to scale AI agents across their business without losing control of security, compliance, or spend. Partners including Uniphore extend Cortex AI Gateway controls to the agents they orchestrate for joint customers.

“Enterprise AI needs more than access to data. It needs an understanding of how people, information, and workflows connect across the business. At Glean, we bring that context together across enterprise systems to make AI agents more intelligent and effective,” said Emrecan Dogan, Chief Product Officer, Glean. “As agents become more sophisticated and operate across more systems, that context must be matched with equally strong governance. Snowflake’s Cortex AI Gateway helps us extend visibility and control across AI interactions, supporting our ability to scale AI securely and responsibly.”

"We run a lot of AI against a lot of data, and the question we couldn't answer was which teams and which agents were driving the spend," said Andy Owens, VP of Data and AI, Kargo. "Cortex AI Gateway gave us that breakdown, plus hard limits so a runaway process doesn't quietly burn through a budget. Routing each request to the right model means our routine work isn't running on our most expensive one. We're running more agents than ever and we know what each of them costs."

“Enterprises don’t just need more agents. They need to industrialize AI across the business. Snowflake gives our joint customers a governed data foundation,” said Frank Petterson, Chief Product and Technology Officer, Uniphore. "Uniphore's zero-copy Data & AI Orchestration Platform gives the customer the ability to autonomously create the data ontology, fine-tune SLMs and build agents while data stays in-place, and governance carries through to every agent we deploy. Cortex AI Gateway keeps that model choice accountable as customers scale from a handful of use cases to dozens.”

With these innovations, Snowflake is raising the bar for what enterprises should expect from an AI agent. Most organizations run their data across multiple platforms, clouds, and environments. Snowflake is interoperable by design, connecting that entire data estate and providing the universal governance capabilities and context that agents and analytics depend on. The data beneath it is governed wherever it lives, across clouds, catalogs, or engines, the definitions behind its answers are ones the business has already approved, and every action it takes is recorded and accounted for. That's what turns an agent from a pilot into something the business can run on, and it's the foundation of the agentic enterprise.

Learn More:

• Learn more about Snowflake’s innovations to create shared meaning, governed actions, and controlled spend for trustworthy AI agents in this blog post.
• Double-click into Business Ontology in this blog post.
• Learn more about what’s new across Cortex AI Gateway in this blog post.
• Read the full MIT Technology Review and Snowflake research report, “Raising the quality of insights in agentic analytics”.
• Stay on top of the latest news and announcements from Snowflake on LinkedIn and X, and follow along at #SnowflakeExpedition.

¹ Snowflake product is now generally available.
² Snowflake product is now in public preview.
³ Snowflake product is now in public preview soon.
⁴ Snowflake product is now in private preview.
⁵ Snowflake product is in development.

Methodology
In July and August 2026, MIT Technology Review Insights conducted a global survey of 400 chief data and analytics officers, chief information officers, chief technology officers, chief AI officers, and other senior data and technology executives. The survey respondents work in organizations across eight industries. All of their organizations earn $500 million or more in annual revenue, with 36% earning over $10 billion annually.

In addition to the survey, in-depth interviews with senior technology executives offer firsthand insights into how organizations are striving to improve the quality of AI agent insights.

Forward Looking Statements
This press release contains express and implied forward-looking statements, including statements regarding (i) Snowflake’s business strategy, plans, opportunities, or priorities (ii) Snowflake’s products, services, and technology offerings, including those that are under development or not generally available, (iii) market growth, trends, and competitive considerations, (iv) Snowflake’s vision, strategy, and expected benefits relating to artificial intelligence and other emerging product areas, and (v) the integration, interoperability, and availability of Snowflake’s products, services, and technology offerings with and on third-party platforms. These forward-looking statements are subject to a number of risks, uncertainties and assumptions, including those described under the heading “Risk Factors” and elsewhere in the Quarterly Reports on Form 10-Q and the Annual Reports on Form 10-K that Snowflake files with the Securities and Exchange Commission. In light of these risks, uncertainties, and assumptions, actual results could differ materially and adversely from those anticipated or implied in the forward-looking statements. As a result, you should not rely on any forward-looking statements as predictions of future events.

© 2026 Snowflake Inc. All rights reserved. Snowflake, the Snowflake logo, and all other Snowflake product, feature and service names mentioned herein are registered trademarks or trademarks of Snowflake Inc. in the United States and other countries. All other brand names or logos mentioned or used herein are for identification purposes only and may be the trademarks of their respective holder(s). Snowflake may not be associated with, or be sponsored or endorsed by, any such holder(s).