TrustLogix has launched its TrustAI integration for Snowflake Cortex AI, aimed at enterprise teams using AI agents in Snowflake environments.
The integration adds a central policy layer for data and tool access used by Snowflake Cortex AI agents. It applies fine-grained, attribute-based access controls and carries authorisation rules through multi-agent workflows, from the user to the agent, then to tools and the underlying data source.
The launch comes as companies test broader use of AI agents for data tasks across large internal systems. That shift is putting pressure on security teams to manage machine-driven access requests at a scale that older, human-centred permission models were not designed to handle.
TrustLogix said its system can make access decisions based on the end user invoking an agent, the user's business role, the agent's purpose and other enterprise attributes. It added that data owners can define policies in business terms rather than through more technical rule structures.
Alongside the Cortex AI integration, TrustLogix introduced TrustAI Guardian, which it described as an autonomous security layer for monitoring and addressing data security risks across human users, service accounts and AI agents. The product includes a natural-language interface intended for Chief Information Security Officers and Chief Data Officers.
Guardian is designed to identify, explain, prioritise and remediate data access risks continuously, according to TrustLogix. Snowflake users can use it to monitor how AI agents access data and how data moves through their environments, the company said.
Ron Longo, Chief Executive Officer of TrustLogix, linked the release to the pace of AI-driven activity inside data platforms.
"AI agents are accessing Snowflake at machine speed. The TrustAI Cortex Integration extends TrustLogix's policy control plane directly into AI agent workflows, so businesses can deploy agentic AI at scale while ensuring security, governance, and compliance," said Longo.
Marketplace listing
TrustLogix also said its Data Security Scanner for TrustCentre is now available on Snowflake Marketplace. The free tool is intended to give joint customers more visibility into data usage and help them identify access risks quickly.
As a Snowflake Native App, the scanner can be installed and run inside customers' Snowflake accounts without moving data outside the platform, according to TrustLogix. The company said this reduces silos and simplifies deployment.
The rollout adds to a growing market focus on governance controls for AI agents as enterprises move beyond isolated pilot projects. The issue has gained prominence as agents are expected to work across different data platforms, cloud services and internal systems.
In comments provided alongside the launch, TrustLogix Chief Executive Officer Ganesh Kirti said governing agents within a single vendor environment is only part of the challenge for large organisations.
"Snowflake is doing important work making governed AI real inside its own platform, and the Natoma acquisition signals they understand that connectivity without accountability is a liability. But most enterprises aren't running a single platform, and that's the architectural reality that gets underweighted in conversations like this. They're running Snowflake alongside Databricks, Microsoft Fabric, Dremio, and whatever else the last acquisition or business unit decision introduced. When agents start operating across that entire estate, reading from an Iceberg lakehouse that spans multiple engines and clouds, the governance question doesn't stay neatly inside one vendor's perimeter. The trust model Snowflake is building is valuable. The harder problem is extending that same model consistently, same policy enforcement, same identity propagation, same audit trail, across every platform where agents are touching data. That's exactly the problem TrustLogix was built to solve: taking the same trust model enterprises are building inside Snowflake and extending it across the full lakehouse estate, so security teams can govern agents and data access consistently whether the query lands in Snowflake, Databricks, Fabric, or Dremio. A security team that can govern agents in Snowflake but has no consistent visibility into what those same agents are doing everywhere else hasn't solved the problem. They've just contained it," said Kirti.
Control plane
TrustLogix Chief Product Officer Ron Zaparaniuk said the broader shift in enterprise AI architecture points to the need for a distinct governance layer.
"The clearest signal from Snowflake CEO Sridhar Ramaswamy's keynote at Snowflake Summit 2026 was architectural: the agentic enterprise is no longer theoretical, and it requires a fourth component beyond data, models, and applications. It needs a control plane. Sridhar's framing was precise. Agents operating without coordination create compounding failures at machine speed. Snowflake's acquisition of Natoma, an enterprise MCP gateway company, reinforced the point: connectivity between agents and enterprise systems is necessary, but not sufficient. Identity-aware authorization, context-aware policy enforcement, and full auditability have to be built into how those connections are made. An agentic control plane is emerging as a distinct infrastructure category, and the enterprises investing in it now will have a structural advantage over those cleaning up ungoverned agent activity later. The second takeaway came from Anthropic CEO Dario Amodei, who joined Sridhar on stage and delivered what may have been the most underrated line of the entire keynote: trust is an accelerant. In enterprise AI, the organizations that build governed data foundations, auditable agent behavior, and predictable outputs earn the right to move faster, not despite their governance investments, but because of them. Sanofi's story made this concrete: five years of data foundation work before generative AI existed is what made their live AI concierge viable on that stage. As a CEO watching this unfold, I keep coming back to the same question every enterprise leader should be asking right now: the data is the fuel, but who controls access to it, and can you prove it? The organizations that answer that question with real infrastructure, full traceability, and enforced least-privilege will outpace those still treating governance as a compliance checkbox," said Zaparaniuk.
Snowflake also endorsed the integration as part of its partner ecosystem. "TrustLogix's TrustAI Cortex integration brings a centralized policy control plane that governs AI agent data access across Snowflake and external platforms, enforcing entitlements end-to-end from user to agent to data source," said Seth Youssef, Head of Data and AI Sovereignty EMEA at Snowflake.
A customer reference included in the launch focused on operational monitoring. "TrustAI Guardian gives us an autonomous assistant that continuously monitors our Snowflake data platform, proactively identifying data access risks before they become issues and bringing out-of-the-box enterprise-grade security posture policies based on industry standards. As we expand our use of Snowflake Cortex agents, having an independent, always-on security layer gives us peace of mind that our data platforms are secure and confidence that they are being used safely and responsibly," said Murali Balakrishnan, Director of Snowflake & Data Governance Platform Tools at BestEgg Inc.