40% of your agents are a critical risk. Read the CISO Playbook for Securing AI Agents

Announcing our Series D to fuel agentic productivity, securely

CONTINUOUS GOVERNANCE

Write each control once. See which apps fail it.

Every application defines security differently. Obsidian brings supported configurations, identities, permissions, and activity into one control model, so you catch gaps and get them fixed.

Request a Demo

Trusted By

Set your standards. Know when an application stops meeting them.

Your security team sets the rules, but each application holds the settings that decide whether they’re met.

Governing each application on its own

With Continuous Governance

Each application implements your rules differently

The same requirement gets translated by hand into each application’s roles and settings.

One set of controls, assessed the same way in each supported application.

Every change can open a gap

A new permission, integration, or configuration change can take an application out of policy between reviews.

Supported controls are checked continuously, and each failure shows what changed.

Gaps wait for the next assessment

Point-in-time reviews find problems late, and someone still has to chase down each owner.

Each finding carries the context to route the fix to its owner.

Check one control across your apps, and route each failure to its owner

Standardize
Assess
Remediate

See each app pass or fail the same rule

Write each requirement once instead of translating it into every app’s settings. Obsidian maps supported identity, permission, configuration, and activity data to one control model.
  • Mappings to NIST, CIS, ISO 27001, and GDPR
  • Each tenant’s pass or fail, violations, and exceptions for a rule, in one view

Know when an application falls out of policy

Someone grants an external user admin access. A setting changes. An old credential stays active. Obsidian continuously checks supported controls and shows what changed, so you’re not waiting for the next assessment.
  • Control status against your threshold, with available change history

Route each failed control to the owner who can fix it

Each finding carries the available identity, permission, and ownership context, so your team can find the cause and route the fix through the tools it already uses.
  • Accepted exceptions kept apart from real gaps
  • Fixes routed through Jira, ServiceNow, and other supported tools

One set of controls, checked in each connected app

A privileged account looks different in Salesforce, Workday, and Microsoft 365. You write each requirement once, and Obsidian checks it in each connected app that exposes the data it needs.

1 Connect your applications

Prebuilt integrations connect major third-party applications, with extension options for custom environments.

2The Knowledge Graph maps them

The Knowledge Graph connects available identities, configurations, entitlements, and activity across those applications.

3 Controls are evaluated through one model

Continuous Governance checks supported controls the same way in each application, mapped to your frameworks.

4 Failures reach their owner

Each finding shows what changed and who owns it, and the fix goes through Jira, ServiceNow, or other supported tools.

The same controls for the agents inside your apps

Salesforce Agentforce, Databricks, GitHub, Microsoft 365, Snowflake, ServiceNow, and Workday now ship their own agents. Many arrive switched on, and users can build more inside the app, often with no security review.

Identify built-in agents

Find the in-app agents in your applications, including ones that arrived switched on and ones users built themselves.

Map what they can reach

See the identities, tools, and data each agent can reach.

Flag risky configurations

Catch agent configurations that fall outside your standards.

Hold agents to the app’s standards

Agent findings show up alongside the application’s other control failures, so you fix them the same way.

Obsidian now discovers and governs the AI agents built into your third-party apps

Those agents sit closest to your most sensitive data: business data in Snowflake and Databricks, people data in Workday, and operational workflows in ServiceNow.
Read the blog →

Additional resources

Control thousands of AI agents with one set of policies

Visit the Runtime Policy Decision Engine page

Obsidian now discovers and governs the AI agents built into your third-party apps

Read blog

How the Knowledge Graph connects your third-party applications

Read blog

Automate compliance across third-party applications

Visit the Audit and Compliance page

Frequently asked questions

How is Continuous Governance different from the Runtime Policy Decision Engine?

Continuous Governance assesses how your applications, and the agents built into them, are configured, and helps your team fix what fails. The Runtime Policy Decision Engine checks individual AI agent actions at runtime, before they run.

Which frameworks can controls map to?

NIST, CIS, ISO 27001, GDPR, and applicable regional frameworks.

What happens when an application doesn’t expose enough data?

Obsidian marks the control as partial or unavailable. Missing evidence never counts as a passing control.

Do we have to change how we track remediation?

No. Route fixes through your existing workflows, including Jira, ServiceNow, and other supported tools.

Is Continuous Governance only useful when an audit is coming?

No. Your team finds supported control gaps as part of everyday security operations. When an audit does arrive, you already have a process for reviewing control status, investigating exceptions, and gathering available evidence.

Know where your applications stand today.

Request a Demo