Skip to main content
Managing your AI workforce requires knowing precisely what your agents are capable of, what tools they have access to, and how they behave in production.

​
Agent Credentials

Agents act on behalf of your organization. Depending on your configuration, Corgtex provides mechanisms to manage their specific credentials.
  • Platform Integrity: The UI allows administrators to securely configure the API keys required to hit external model providers (OpenRouter, OpenAI) if you are self-hosting.
  • Tool Access: Agents can be granted or denied access to specific platform tools based on the credentials assigned to their runtime profile.

​
The Workforce Graph

The Workforce Graph provides visibility into every agent running within your environment. It maps out:
  • Which AI agent is assigned to which tasks.
  • Which specific roles “own” or manage those agents.
  • What data sets (Sources, Data Connectors) the agents are authorized to access.

​
Reviewing Agent Runs

As agents execute tasks asynchronously (handled securely by apps/worker), they leave a dense audit trail. You can view the execution logs of an Agent Run:
  • See the initial prompt or trigger.
  • Review the specific AgentSteps taken by the LLM.
  • Inspect the raw JSON payloads of any AgentToolCall (e.g., when the agent triggered a search against the Organization Brain or requested a status check).
Human-on-the-loop: For high-impact actions (like modifying a database record or submitting a proposal on behalf of a user), Corgtex agents can be constrained to require human approval before finalizing the execution.

​
The Daily Briefing

Your window into this visibility is the The Daily Surface. Corgtex synthesizes the activity of the AI workforce, highlighting only what is relevant to your assigned roles. The briefing surfaces anomalies, summarizations of completed agent runs, and alerts for governance proposals requiring your attention, ensuring you are never blinded by the speed of automated execution.