Classie launches Supervise to track, control and account for enterprise AI agents in real time

Classie today unveiled its AI supervision platform, giving CIOs and CISOs a unified view of AI agent activity across the enterprise. With Classie Supervise, now generally available, organizations can monitor sanctioned and unsanctioned agents as they run and enforce policy in real time.

AI agents are moving rapidly into enterprise applications and workflows. Gartner predicts that 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from less than 5% in 2025. Coding agents can already write and deploy code, while knowledge-worker agents connect to business applications and act on behalf of employees, creating a new operating challenge as software begins to make decisions and take actions at runtime.

Governance has not kept up. Gartner reports that only 13% of organizations believe they have the right AI agent governance in place, while warning that growing agent sprawl is increasing IT complexity and exposing enterprises to risks including oversharing and data loss. Gartner recommends that organizations establish centralized agent inventories and maintain ongoing visibility into agent usage and behavior.

Runtime Security and Supervision for Agents Already at Work

Supervise provides runtime security for sanctioned and unsanctioned agents, combining AI posture assessment with inline monitoring and control. Supervise follows agent activity across endpoints, browsers and enterprise compute environments, creating a runtime transcript as the workflow happens. That transcript connects agent and user identity with the context, intent, data accessed and actions taken during the session. 

Because Classie can tie activity and token consumption back to the people, agents and environments involved, organizations gain real-time attributable spend alongside a chain-of-custody record of agent activity.

Organizations define rules through an Open Policy Agent (OPA) policy engine. Lightweight sensors enforce those rules inline. Signals from sensitive-data detectors and Classie’s Supervisor agents allow the platform to stop or redirect an action before it becomes an incident.

Classie can be installed in 15 minutes and begins discovery immediately. The platform is designed to deliver a monitored enterprise posture within five days.

Key capabilities of Classie Supervise include:

  • Real-time intervention. Alert, restrict, reroute or stop an agent action as it happens.
  • Cross-surface visibility. Follow sanctioned and unsanctioned agent activity across endpoints, browsers and corporate compute environments.
  • Real-time attributable spend. Connect AI activity and token consumption to the users, agents, sessions and environments that generated it, giving enterprises visibility into where AI spend originates.
  • Policy enforcement. Apply consistent operating rules through an OPA-based engine and lightweight sensors.
  • Agentic chain of custody. Create a traceable and auditable record connecting agent activity with identity, context and intent, providing tamper-evident provenance for investigation, compliance and governance.
  • Private deployment. Keep enterprise data inside the customer’s environment.

One Platform, from Visibility to Control

The general availability of Supervise follows the release of Classie Discover in April 2026 and Classie Analyze in July 2026. Together, Discover, Analyze and Supervise take organizations from discovering AI activity and understanding its behavior, to applying controls in real time:

  • Discover. Find sanctioned and unsanctioned AI tools, agents and models, then map the people and data connected to them.
  • Analyze. Examine agent behavior and intent, replay interactions and identify activity that needs attention.
  • Supervise. Apply enterprise rules in real time and intervene when an agent moves outside approved boundaries.

The sequence is critical because controls are only reliable when teams first know what is running and understand how it behaves. Classie creates that context before enforcement begins. The platform works across AI vendors, allowing companies to carry one set of operating rules across a heterogeneous environment.

Classie runs its continual learning infrastructure in customers’ private cloud, allowing them to retain control of the data used to adapt supervision models and the resulting model weights. This approach keeps institutional knowledge inside the enterprise and helps automate the development of new guardrails over time.

Customer Use Case Examples

Healthcare company: Classie was deployed in 15 minutes and, within 24 hours, identified the full range of known and previously unknown AI tools being used by employees. The security team could see complete sessions and files being accessed, including previously unseen use of personal AI applications and instances where sensitive information was being shared with AI tools the company had neither licensed nor sanctioned. The findings gave the security team a concrete view of AI activity it could take to leadership and use to inform its governance approach.

High-growth AI company: Classie surfaced every live AI session within five minutes of installation, including several agent harnesses the company had not previously tracked. The company now uses Classie to bring agent transcripts and spend into one place, giving leadership a consolidated view of AI activity as it works to supervise agents and understand how effectively they are being used.

Availability

Classie Supervise and the complete Classie platform are generally available today. Observability pricing starts at $99 per user. For deployment and pricing information, contact Classie at letstalk@classie.ai.

Leave a Reply

Discover more from The IT Nerd

Subscribe now to keep reading and get access to the full archive.

Continue reading