The price of an AI employee is made of a one-off build, monthly support and the AI model cost. It depends on the number of processes, the tools it connects to, the volume and where the AI runs. That is why quotes are custom: after a free audit you get a fixed price and a forecast of the hours and money you'll save.
What the price is made of
- Build (one-off): process audit, setup, connections to your tools, testing, team training.
- Support (monthly): monitoring, fixes, new cases, a monthly report.
- AI model: pay-per-use for cloud models. With local AI on your own server this cost disappears, but there is a hardware cost.
What drives the price
- How many processes the AI takes over and how complex their rules are.
- How many tools it connects to and whether they have an API.
- Volume: enquiries, documents or videos per month.
- Where the AI runs: contracted cloud, private EU cloud or your own server.
How to work out whether it pays off
Take the hours a month the task takes and the hourly cost of the person doing it (salary plus employer costs, divided by working hours). Multiply by the share the AI can take over, usually between half and nearly all of the repetitive part. Compare with the monthly price in the quote. The calculator on our home page does exactly this.
When it isn't worth it
When the task takes little time, when every case is unique, or when the data is so scattered it has to be sorted first. Then we say so plainly instead of selling.
Guarantee
If within 60 days of launch your first AI employee isn't saving at least 20 hours of work a month (against what we measured in the audit), we keep tuning it at no extra cost. If it still isn't by day 90, we refund the build fee. The exact terms are in the contract.