Help Desk Agent
Answers employee questions from policy and the system of record.
75%1 (modeled outcome) of HR case volume deflected
Glossary
Definition
A governed agent is an AI agent confined to one workflow, permitted data, and logged actions, with human approval required for consequential decisions.
A governed agent is distinguished from an open-ended assistant by its constraints. It is scoped to one workflow. It reads only data it has been granted, under the source system's security model. Every action it takes is logged with the identity it acted for. Actions with consequences, such as moving money, changing pay, or sending documents outside the organization, are held for a named person's approval.
The constraints are not a limitation on usefulness. They are what allow an organization to let the agent act at all. An audit committee can accept an agent that prepares journal entries for a human to post; it cannot accept one that posts on its own judgment.
All twelve Meridian agents are governed agents, and the same constraints apply to agents built in Studio or connected through the Gateway.
Related agents
Meridian agents whose work depends on governed agent. Each is scoped to one workflow and logs every action.
Read next
Long-form explainers and frameworks that use this term, dated and signed.
Governance
A seven-part framework for governing AI agents in HR: scope statements, least-privilege data permissions, approval tiers, audit trail, model choice, employee communication, and works council and GDPR alignment.
Marcus Feld7 min read
Buying and ROI
How credit-based pricing works for AI agents: what counts as a billable action, Meridian's published credit rates, three worked examples, budgeting, and guardrails against overage.
Aisha Bello6 min read
Get started
Start with one workflow, one approver, and one number to move. Most design partners were live in five weeks.