Policy & Compliance

AI governance —
applied automatically

Stop relying on users to remember the rules. UCM makes governance happen — every conversation, every agent, every time.

The Reality

Policies only work if they're followed

Most organisations have governance policies — acceptable use documents, data handling rules, tone guidelines. The problem is not the policies themselves. It is getting AI users and agents to apply them consistently.

Manual reminders fail. Shadow AI ignores them entirely. UCM delivers your governance documents automatically to every relevant AI conversation — no user action required.

How It Works

Write the policy once. It applies everywhere

📋

Policy Document

Your governance rules, written in plain markdown

🗄️

UCM Policy Store

Stored with a priority order and applied to the right targets — org-wide, by team, or per agent

🤖

AI Agent

Receives your guidance before every conversation begins — automatically

Governed Conversation

Every interaction receives your guidance — automatically, every time

A policy is a document stored in UCM — your acceptable use rules, data handling guidelines, tone of voice, or anything else your AI should follow. Assign it once and it is applied automatically to every relevant conversation, in the order you choose.

Key Capabilities

Governance that runs itself

Policies in UCM are documents — written in plain language, stored and versioned like any other file. Assign them to the right targets — your entire organisation, a specific team, or an individual AI assistant — and UCM delivers them automatically before every conversation.

No user intervention. No clipboard paste. No hoping someone read the memo.

  • Digitise existing governance docs into versioned, automatically applied policies
  • Assign to 5 target types: organisation, team, repository, agent, recent activity
  • Delivered automatically before every conversation — in the priority order you set
  • Required Reading policies — always delivered, regardless of context
  • Enable/disable policies without deleting assignments
  • Assignment history: who assigned each policy, and when — recorded in the database
  • Compliance reporting dashboard* — visibility into policy coverage across agents and teams

Active policies

⚖️
Acceptable Use Policyorg-wide · priority 1
injected
🔒
Data Handling Rulesteam/engineering · priority 2
injected
✍️
Tone & Brand Voiceagent/marketing-bot · priority 1
injected
📋
GDPR Guidancerepo/customer-data · priority 3
injected

What You Can Govern

Any AI interaction. Every governance layer

For every team, every use case, and every AI tool in your organisation — UCM governance applies consistently.

🏢

Organisation-Wide Rules

Apply a policy to your entire organisation and every AI interaction automatically follows it — no exceptions.

👥

Team-Specific Guidelines

Assign policies to specific teams or workspaces. Engineering standards, legal constraints, and customer service scripts each apply to the right group.

📁

Repository Standards

Tie policies to specific repositories. Coding standards, data classification rules, and security checklists apply when AI agents work in that context.

🤖

Agent Instructions

Assign policies directly to individual agents. Customer support bots, onboarding assistants, and document processors each get their own governance layer.

📌

Required Reading

Mark a policy as Required Reading and it is always injected — regardless of context. Your non-negotiables are truly non-negotiable.

📊

Conversation Logging

Chat agent conversations are retained for 30 days. Compliance reporting and audit dashboards* are on the roadmap.

Get Started

Governance that works — without asking users to

Publish your first policy in minutes. Assign it to your organisation. Every AI conversation is governed from day one.

* on the roadmap