Why banning AI or buying a few licences is not enough
Banning AI doesn't work in practice: people open ChatGPT on a private account and paste in a contract, a client e-mail or a price list. The company then doesn't know what left, where it went or who is responsible.
Buying individual licences means several subscriptions from different vendors, no common rules and no overall view of spend.
How FutuGen AI Gateway works
Gateway sits between your employees and AI models. Every request passes the rules you set before it reaches a model.
- Personal data de-identification – names, national ID numbers, contacts, addresses or account numbers are replaced with placeholders before anything goes to an external model.
- Data-class and region rules – decide which data may go to which model and where it may be processed. Sensitive requests can be routed to a local model.
- Budgets and caps – limits for the company, teams and individuals, with real-time usage.
- Audit log – for every request you know who used which model, when, and what it cost.
- EU hosting – the hosted instance runs on Google Cloud in Frankfurt (europe-west3).
What employees see
Employees work in the chat at chat.futugen.com – much like ChatGPT, but with company rules. They can choose between GPT-5, Claude, Gemini, Mistral and local models, as allowed by the company. Developers reach the same models through one API.
What it costs
You can start free: after sign-up you get CZK 100 to try it out, then you pay only for actual usage from prepaid credit in CZK – no monthly fee. Larger teams can choose plans with their own (dedicated) instance from CZK 4,900 per month excl. VAT. Invoices are Czech, with company ID and VAT ID.