Scenarios

a business example of porc consider a case where an operations focused agent wants to restart a production service principal an operations engineer, through an approved agent operation restore service availability resource a production service or server context change window, ticket status, business criticality, backup state, and approval state the result might be to allow the action, require approval, or deny it altogether, depending on policy scenarios building a secure ai agent a developer creates an ai agent for incident diagnosis and checks the code into the approved repository, triggering the standard build pipeline during this process, agent visor validates compliance with security and governance requirements, checking for approved base images and scanning for vulnerabilities if critical issues or policy violations are found, the build is halted and the developer is notified if all checks pass, a secure container image is generated and published to the organization's approved image repository business value security is embedded into the development lifecycle developers do not need to perform additional governance steps only compliant agents are allowed into production environments upgrading an existing agent with agent visor an organization with existing agents wants to implement governance and security controls without starting from scratch the agent owner rebuilds the agent using the agent visor enabled process, updates the agent catalog with the new image version, and the deployment begins automatically, retaining existing configurations and schedules this allows the agent to benefit from agent visor's governance, policy enforcement, audit logging, and observability features business value existing agents can be secured quickly no need for re registration or redeployment governance is applied consistently across all agents deploying the agent a patch compliance agent is registered once in the agent catalog and used by multiple customers with varying operational needs operations teams create separate deployments of the same agent—customer a deploys it weekly with one llm model, while customer b uses a different llm for daily deployments each deployment can have distinct schedules, permissions, policy sets, and configurations, ensuring that each customer receives a tailored solution business value reuse agents across multiple customers reduce development and maintenance effort support customer specific governance requirements enforcing policies during agent execution an agent requests access to an external service while performing a task agent visor evaluates the request against policy rules if compliant, the request is approved; if not, it's blocked and the reason is logged this decision is part of the execution audit trail for future review by administrators or compliance teams business value provides real time governance prevents unsafe agent behaviour ensures adherence to corporate policies visualizing agent decisions and policy outcomes an account team seeks to understand how governance influenced an agent's actions through a policy evaluation view, which visually outlines the agent's activities users can track the request's journey through policy evaluations, decisions made, and their impact on the final outcome, all in a business oriented workflow that links agent actions, policy decisions, and recommendations, instead of sifting through complex logs business value makes governance understandable to non technical users improves user confidence in agent decisions facilitates operational reviews and compliance discussions