Making AI use in a company explainable
Who we are
Acotia is an AI governance platform for companies of 50 to 300 people. We are based in Barcelona and currently work with our first companies, in pilot phase.
Updated
What we do
Acotia is the single point of access to a company's AI models. It detects sensitive data before it leaves (tax IDs, IBANs, cards, secrets, emails, phone numbers and custom terms), masks or blocks it according to the company's policy, logs every request without storing its content and puts a ceiling on spend per company, team and person. We explain it in detail in the platform.
Why it exists
AI is already used inside companies, often with personal accounts and without anyone having decided it. Leadership and legal can neither ban it successfully nor allow it blindly. We believe the answer is neither a ban nor yet another chat, but a layer between people and models that makes use visible, bounded and explainable. The principles we follow are written down and can be checked.
How we work
We start with a pilot of 15 days in audit mode: the system records what it would have done without blocking anything, and with that report the company decides its policies. We say what the product does and what it does not. No detection is perfect (for example, pattern-based detection does not reliably recognise people's names): we reduce risk measurably, we do not eliminate it.
What we keep and what we do not
For each request we keep who made it, which team, which model was used, which data categories appeared, which action was applied and what it cost. The text of messages is not stored on the server: conversations are stored encrypted on each person's device. Accounts and policies live in the EU (see data in the EU) and models are not trained on your data.
Who we work with
Companies of 50 to 300 people that handle other people's data: accounting and advisory firms, law firms, private clinics and financial services. We are in pilot phase, so we do not yet publish clients or usage figures we cannot back up.
How to talk to us
Write to hola@acotia.com and tell us how your team uses AI today. In one call we tell you whether a pilot makes sense. We reply in Spanish, English and Catalan.
At a glance
- What it is
- AI governance platform for companies
- Based in
- Barcelona, Spain
- Founded
- 2026
- For
- Companies of 50 to 300 people
- Status
- In pilots
- Apps
- macOS and Windows
- Languages
- Spanish, English and Catalan
- Contact
- hola@acotia.com
What we believe
- 01
AI is already inside
It is used daily, with or without permission. Banning it does not work; making it visible does.
- 02
Measure before you ban
That is why everything starts in audit mode: first you know what leaves, then you decide.
- 03
The sensitive value is not needed
The model almost never needs the real tax ID to help. With a label it works just as well.
- 04
What cannot be explained cannot be defended
Every request leaves a trace, without storing its content. That is what you show an auditor.
Frequently asked questions
Who is behind Acotia?
A Barcelona company founded in 2026 that builds an AI governance platform for companies of 50 to 300 people. It provides a single access point to models, detects and masks sensitive data before it leaves, measures every request and puts a ceiling on spend.
Where is Acotia based and how do I contact it?
Acotia is based in Barcelona, Spain. Contact is hola@acotia.com, and we reply in Spanish, English and Catalan.
How does Acotia treat company data?
It stores metadata for each request (who, team, model, detected categories, action and cost), not the text of messages. Conversations are stored encrypted on each person's device, and accounts and policies are in the EU.
Does Acotia have customers yet?
We are in pilot phase with companies of 50 to 300 people in Spain. We do not publish customer names yet.