AI governance

People set the boundaries for AI

Serene and AI Employees work inside your permissions. Changes wait for a person’s approval, actions are recorded, budgets cap spending and admins can switch AI off.

  • Approve before anything changes
  • Monthly credit cap
  • One switch to turn AI off
Serene, HyoNova’s AI teammate, beside a lock, permission switches and a manager’s approval.

At a glance

How does HyoNova keep AI under control?

AI works with the permissions of the person or AI Employee, risky actions wait for approval, actions are recorded and admins can switch AI off.

What can AI never do?

Make payments or refunds, change permissions or security settings, or delete data. A person does those.

Who decides AI actions?

The member who asked decides Serene’s proposals. Managers, admins and owners decide AI Employee proposals when they hold the permission the action needs.

Which roles can use AI?

Owner, Admin, Manager and Employee. Restricted members cannot. Custom roles use ai.read, ai.approve and ai.admin.

Controls built into every AI action

The same rules apply in chat, in voice and to AI Employees.

Permissions

Serene acts as you. AI Employees use only the permissions, tools and knowledge you grant.

Approvals

Edits, reassignments and customer replies wait for Approve or Reject.

Runs once, then expires

An approved action runs exactly once, and undecided proposals expire after 7 days.

Audit

AI actions are recorded for review, with who approved them.

Budgets

Owners set a monthly Serene credit cap, and each AI Employee can have a daily spending cap.

Off switch

Admins can turn Serene, voice conversations or AI Employees off for the company. Nothing is deleted.

Autonomy levels

AI Employees can draft only, work under supervision, or handle low-risk actions with your consent.

Human handover

AI Employees hand customers to your team for sensitive topics, VIP customers and attachments.

Clearly AI

Serene and AI Employees are always marked as AI in the office and in conversations.

Limits on every run

Each question has limits on credits, steps, tool use and time.

How it works

Decide who uses AI

Give roles ai.read, ai.approve or ai.admin.

Choose sources

Switch on the company knowledge Serene may search.

Set budgets

Set a monthly credit cap and AI Employee budgets.

Review

Decide approvals and check AI usage and audit records.

Governance that is real

Same rules as people

AI passes the same permission checks as members, on every tool, every time.

Nothing simulated

If the AI service is unavailable, Serene says so instead of guessing.

Platform staff see metadata only

HyoNova’s team sees duration, cost, which tools ran and whether they succeeded, never your questions, answers or tool results.

Questions and answers

Can Serene message a customer on her own?

No. A customer reply proposed by Serene waits for a person to approve it. AI Employees reply only within the autonomy you choose.

Who can turn AI off?

Owners and admins, in Company administration › AI & Serene.

How do we control cost?

Set a monthly Serene credit cap. When it is reached, Serene stops and says so.

Is governance included in every plan?

Yes. Permissions, approvals, audit and the off switch come with Serene on every plan.

Use AI with clear boundaries

Start free. AI follows your rules from day one.