Human in the loop: your central dashboard for every human approval of AI actions
AI speeds things up, but it takes no responsibility. So ZiyaOS pauses every action you have set to need approval and puts it in front of the right person: from agents, workflows and apps, from external agents on the gateway and from the firewall.
Sales agentSend quote to Müller GmbHEmail with 2 attachmentsRejectApproveApproved
Invoice workflowApprove payment of €4,380Supplier Becker KGRejectApproveApproved
Claude CodeDelete customer recordCRM, contact 2817RejectApproveRejected
FirewallSend data to api.neuer-dienst.ioResearch agentRejectApproveApproved
How it works
1
Pause
If an action is set to ask first, ZiyaOS pauses it and puts it before a person with everything needed to decide.
2
Route
The request goes to the people who may approve it by rule. Management sees everything, accounting sees its payments.
3
Decide
Approve, edit or reject. You can change a draft before approving it, such as the greeting of an email.
4
Continue
After the decision the agent or workflow continues where it paused. The decision is in the log.
AI speeds things up. Responsibility stays with people.
Responsibility matters more than ever today. When AI sends quotes, changes data or moves money in seconds, it has to be clear who answers for it. In ZiyaOS the AI prepares as fast as it can. Whatever has consequences is approved by a person, and every decision is logged with a name.
AI preparesin 2 minutes47cases prepared44 run automatically
Needs your approval
Cancel contractSupplier Becker KG · by 31 MarchA. Weber
Change pricesShop · 214 items · +4%A. Weber
Send reminderKraemer Bau · final noticeA. Weber
Every approval logged with a name
Human in the loop in action
In these solutions a person decides what your AI may do.
Approvals for AI agents
Every request in one place, with the person in charge.
Questions from agents, workflows, apps, the gateway and the firewall land in the same place.
Everyone sees their own
Roles & permissions decide who may approve what. Everyone sees exactly the decisions that concern them.
Edit before approving
Drafts can be changed before approval. The log shows whether and what was changed.
Notes for the agent
Add notes with your approval, for example about the greeting. The agent takes them into account next time.
Nothing runs without you
As long as nobody decides, the action waits. Nothing happens in the background.
Fully logged
Every run, every call and every decision is in the log, internet requests and external agents included.
Frequently asked questions
Who decides when the AI asks for approval?
When the AI asks for approval, a person you choose decides. All requests from agents, workflows, apps, the Gateway and the Firewall land in one place. There you approve, edit or reject, and every decision is logged.
Who is liable if the AI makes a wrong decision?
Responsibility stays with people: AI output has to be checked by people, and AI permissions and approvals are granted by people. With ZiyaOS you can trace exactly who did what, how and when, based on which approval and from whom.
How do you make sure the AI doesn't do anything wrong?
We make sure the AI doesn’t do anything wrong with the central Guard: every action an AI wants to take runs through it. Without a rule or explicit approval, there is no access. Every action is logged: who, when, what, why, from where and to where. The same applies to people, apps and workflows.