AI that works inside your system
We are building an AI assistant that knows your PolyOmi data, and AI agents that carry out routine tasks alongside your automations, with approval steps and a full log. It is not available yet. Join the waiting list to get notified.
Planned
What the AI will help with
The aim is fewer clicks on routine work, not a chatbot bolted on the side.
-
Ask your data Planned
Ask in plain language: which invoices are overdue, how many hours went into a project, which clients have open tickets.
-
Draft replies Planned
Suggested answers to tickets and client messages, in the client's language, for you to edit and send.
-
Summarise tickets Planned
A short summary of long ticket threads, so a colleague can take over quickly.
-
Categorise vouchers Planned
Suggested account, VAT code and supplier for receipts in the voucher inbox. You approve before anything is posted.
-
AI agents in Auto work Planned
Agents that carry out routine tasks next to your automation rules, such as preparing weekly reports or sorting new tickets.
-
Approval steps and audit log Planned
Agents propose, people approve. Every action is logged with what was done and why.
-
Choose your model provider Planned
Pick the AI provider and model you trust, and use your own API key.
-
You control the data Planned
Decide per module what the AI may read. Nothing is switched on until you turn it on.
Fits what you have
Built on Auto work and your audit log
PolyOmi already automates routine work. Auto work creates tasks on its own for recurring jobs, expiring domains and SSL certificates, unanswered tickets and approaching deadlines, and logs every run.
AI agents will be the next step: instead of only creating a task, an agent can do the first part of the work and hand it to a person for approval. They will follow the same roles and permissions as everyone else, and their actions will land in the same audit log.
- Available today: Auto work rules with a run log
- Available today: roles, permissions and a full audit log
- Coming: AI assistant and AI agents with approval steps
Questions about AI in PolyOmi
Will I have to use AI?
No. Our principle is that AI is optional and switched off until you turn it on. PolyOmi works fully without it.
Which AI models will it use?
The plan is that you choose the provider and model yourself and connect it with your own key. We will list the supported providers when the feature is released.
Can an AI agent post to my books?
Not without a person approving it. Agents will propose, and a person with the right permissions approves. The ledger stays tamper-proof, with every posting numbered and logged.
When will it be ready?
It is in progress now. We do not give dates. Follow the roadmap or join the waiting list to hear first.
Tell us which routine tasks to take off your plate
Join the waiting list and let us know where AI would save you the most time. We will notify you when the first features are ready to try.