A Guth Labs publication ↗
Back to News

Una AI releases planning agents and governance controls

Three green glass robotic nodes arranging cream tiles beneath a gold approval arch
AI-generated editorial illustration

Una AI made Orchestrator and AI Governance generally available to its customers on November 18, combining specialized planning agents with controls over their access and actions.

One example workflow takes an Excel budget, builds a planning model, imports actual results from an enterprise resource-planning system and prepares a review process.

A modeling agent, data agent and workflow agent handle different stages. Una says their proposed changes are applied only after the finance team approves them.

Administrators can restrict AI to read-only operation, permit or block individual tools, and prevent access to sensitive data even when the initiating employee can see it. Action logs record the user, outcome and reasons for denied requests.

The controls also cover external AI applications connected through the Model Context Protocol, including Claude and ChatGPT. According to Una, AI cannot expand its own permissions or access more than its initiating user.

The announcement does not provide independent performance or security results. Its significance is pairing broader automation with explicit approval and access boundaries inside financial planning, rather than presenting autonomous changes as the default.

Sources and citations

Revision history

Published .