Inference and authorization are not the same thing. Our latest resource explores the difference between inference and authorization, and why the distinction matters for governed AI data access. Read more: ow.ly/3GTs51012AL

Delegation, data-plane enforcement, and just-in-time access. Our latest whitepaper explores how these principles can help organizations govern AI agents as distinct identities while preserving the context of the users they act for. Read the whitepaper: ow.ly/R6q15100yK4

Manual governance hits a ceiling. And it can happen sooner than you think. Too little access can leave people waiting for the data they need to do their jobs. Too much access can create unnecessary risk for the organization.

How should AI agents access enterprise data? There are a few ways to approach it. Immuta CTO Steve Touw walks through all three approaches and explains why just-in-time access vending provides a stronger foundation for agentic data access.

Paris was a good one. It was great to connect with the Snowflake community and hear how organizations are approaching AI agents and governed data access.

Join Immuta at Teradata World Tour London to explore how the data authorization layer for the agentic enterprise can help teams apply policy-driven, temporary, and task-bounded access for humans and AI agents. Book a meeting here: ow.ly/tWp550ZYvBQ

AI agents shouldn’t have access by default. When an AI agent is compromised, any standing permissions it holds could give an attacker access to data they shouldn’t have. A better approach is zero standing privileges.

In our latest article, Immuta CTO Steve Touw explores why human-centric access controls will never work for AI at scale, and what a new authorization model needs to look like. Read the full article: ow.ly/CM6G50ZXGE0

Can your team explain every AI agent data request - who made it, what they asked for, and which policy allowed or denied access? If not, you’re missing an authorization layer. ow.ly/UU4450ZXg1T

For CDOs, the challenge is no longer simply how to make data available. It’s how to make it usable for AI while maintaining the governance, security, and accountability the business needs. That’s what we’ll be exploring at the CDO Summit. ow.ly/F9a350ZXfWJ

Sam Carroll shares his thoughts on AI’s growing intelligence. Effective access controls on data platforms need to be implemented to prevent AI agents from impersonating users or gaining autonomy over data you never thought they could access.

A great evening of conversations at AI + Data Afterhours. It was great to bring the @ServiceNow and Immuta communities together in San Francisco for an evening focused on what’s actually happening across AI, data, security, and governance.

Meet with us at Snowflake World Tour Paris on October 15th. We’ll be talking about what comes next, policy-driven, temporary, and task-bounded access that complements Snowflake’s native security foundation. Book time here: ow.ly/BxVN50ZVq5n

Every request for data needs permission. Now there are millions of requests. At Teradata London 2026, the Immuta team will be exploring how organizations can enable AI innovation while maintaining control, accountability, and governance. ow.ly/x79U50ZUxUz

Your fastest-growing data consumer isn’t human. The challenge is that data access models were built for people, not autonomous systems. Read our CEO, Matthew Carroll’s, blog to explore this: ow.ly/ZhZw50ZTyYb

It was great to spend time with the Snowflake community in Berlin discussing how organizations are approaching this shift and the role governed authorization plays in building trust in AI.

We’re hiring. Come build what’s next in data security. Explore opportunities at Immuta: immuta.com/careers?utm_sourc…

OpenAI and Anthropic’s agent security guidance is a useful starting points but as agents work across systems and on behalf of people, authorization has to account for identity, policy, task, and context, not just authentication. ow.ly/Bs7950ZQVhz

What a week at Oktane 2026. It was great to share more about Immuta and introduce the Data Authorization Layer. Thank you to everyone who stopped by, joined the conversation, and helped shape the discussion around the future of enterprise data.

The Data Authorization Layer is the decision point between identity & enterprise data. It evaluates the request against policy and context before the data platform enforces the outcome, determining what the request may access, under which conditions, and for what task.