An environment where data does not leak
We detect what is sensitive before it leaves and mask it, route it to a model hosted in the EU, or block it. By your policy, not by luck.
SecurityView →Acotia is an AI governance platform for companies. It gives your team the best models inside a perimeter you define: sensitive data never leaves raw, every request is measured and spend has a ceiling.
What is AI governance?
AI governance is the set of rules, controls and records that let a company use models such as ChatGPT, Claude or Gemini without exposing sensitive data, knowing who uses them and keeping spend under control.
Four principles we hold to on every deployment, from the first pilot. Not slogans: architectural decisions you can verify.
View Principles →Sensitive data is routed to models hosted in the EU. Whatever travels to a provider abroad travels sanitised.
02Your team conversations do not feed third-party models. And it does not depend on each employee remembering to switch it off.
03What was masked, what was blocked and why. Exportable for audit, for your DPO and for the client who asks.
04One login with Google or Microsoft. No more loose accounts or shared cards.
The environment where your team works with AI, and where the company takes back control of its data, its usage and its bill.
View Platform →We detect what is sensitive before it leaves and mask it, route it to a model hosted in the EU, or block it. By your policy, not by luck.
SecurityView →Who, which model, how much and for which team. Every request is classified on the fly and the server stores the label, not the text: leadership sees real adoption and real cost without reading a single conversation.
MetricsView →You pay for real usage, with caps per person and team. No more unused licences or tokens burned on tasks that never needed them.
SavingsView →Nothing gets cut off on day one. We start by looking at what already leaves your company, and policies arrive once the numbers are on the table.
Companies of 50 to 300 people where AI is already used daily and where a leak is not a scare but a case file.
Your clients' tax and employment data is what your team works with every day. Including when it uses AI.
Professional privilege does not distinguish between an email and a chat window. The duty is the same.
Health data is a special category under GDPR. The margin for error is the narrowest there is.
Brokerages, asset managers and advisers with regulation on top and client data in every email.
Where it runs, what it keeps, which model accounts it uses and who can change what. What they usually ask, at a glance.
See the trust page →A desktop app for each person and a gateway on the server. Accounts, policies and the log live in the EU.
Metadata for each request: who, team, model, categories, action and cost. Not the text: conversations are stored encrypted on the device.
The company connects its own OpenAI, Anthropic, Google or Mistral accounts. Keys are encrypted and never return to the browser.
Sign-in with Google or Microsoft, three roles (super admin, admin and employee) and a log of every policy or permission change.
An AI governance platform for companies of 50 to 300 people. It gives the whole team a single access point to models, detects and masks sensitive data before it leaves, measures every request and puts a ceiling on spend. More about who we are.
With two barriers. First, on the device, tax IDs, IBANs, social security numbers, cards, secrets, emails, phone numbers, IPs and the custom terms the company defines are detected and replaced with labels (for example [IBAN_1]); the real values do not leave. Then the server repeats the detection and applies the stricter action. No detection is perfect: details in security.
The server stores metadata for each request (who, team, model, detected categories, action and cost), not the text. Conversations are stored encrypted on each person's device and the company decides how long the log is kept. More in data in the EU.
A fee per company plus an amount per person: Essential, €79 a month + €9 per person (up to 25), and Professional, €199 + €12 per person (up to 50); for more than 50 people, a tailored proposal. Excluding VAT. Each company pays model usage directly to its AI provider, with no markup from us. The 15-day pilot costs €149 up to 25 people, €299 for 26 to 50, and for more than 50 the equivalent of the first month (about €800 for 50 people on Professional) or a tailored quote. It is deducted in full from your first invoice if you sign up within 30 days. Details in pricing.
With a 15-day pilot: connect on days 1 and 2, measure in audit mode (blocking nothing) from day 3 to day 12 and decide policies on days 13 to 15. The full detail is in rollout and audit mode.
Companies of 50 to 300 people that handle other people's data and have no dedicated security team: accounting and advisory firms, law firms, private clinics and financial services.
Tell us how your team uses AI today. One call and we will tell you whether a pilot makes sense.
Request a pilot→ or write to hola@acotia.com