How we work with you
Our services:
secure AI, tailored to you
We help with any artificial intelligence project, with data protection as our speciality. These are the four ways of working together: you can start with whichever presses hardest today.
Any AI case
We take on any AI project.
Our speciality is that it does not cost you your data.
If your case does not quite fit any of the four doors below, it is still our case. These are the requests that reach us most often.
Start and decide
- Where to start: which AI uses make sense in your organisation and in what order.
- Which tools and which models to choose — and which to rule out.
- Training for the team, in Basque or Spanish, with cases from your sector.
- A second opinion on an AI proposal or supplier.
Build and protect
- Building chatbots and internal assistants connected to your documents, ERP or CRM.
- Automating repetitive tasks and workflows with AI.
- Integrating AI with your ERP, your CRM or your documents.
- Reading and extracting data from documents (invoices, files, reports).
- Data protection, audit and regulatory compliance.
We work from Azpeitia (Urola Kosta, Gipuzkoa) with companies, schools and public bodies across the whole Basque Country, in Basque, Spanish and English. If you do not know where you stand, that is precisely the conversation we like to have.
For those already using external AI
LLMProtect:
preventing data leaks in AI
Nothing leaves your company unless you decide it.
We analyse in real time what your team sends to external AI chats and block sensitive information before it leaves.
How it works
When someone on your team copies a document into an AI chat (ChatGPT, for instance), that text leaves your company, often without anyone stopping to think. LLMProtect kicks in at that exact millisecond: it analyses the message, the document or the image, and stops whatever must not leave.
Where it fits best
Especially useful in industry (where your own know-how is your greatest asset), and in education or professional practices, where every day you handle data that is not yours but belongs to students, families or clients.
What it includes
- Real-time analysis of text, documents and images.
- Automatic blocking of sensitive or confidential data.
- Four service levels (including a free tier to start trying it out).
For those who want total control
Chatbots and assistants · local AI:
a private corporate chat
Your own Artificial Intelligence, without the data leaving home.
We implement a private chat with models specialised for your company, connected to your documents, ERP or database. Nothing leaves your servers.
How it works
It is the direct route to real governance of your AI. Instead of sending your data to a third party's cloud, the AI lives inside your company. You can connect it to your internal information so the team can run queries, look up data or generate content in complete privacy.
Where it fits best
Powerful for industry, education and public institutions. Ideal for settings where confidentiality, industrial secrecy and data sovereignty weigh more than the convenience of an external public cloud.
What it includes
- Language models (LLMs) specialised for your sector.
- Secure connection to your documents, ERP or databases.
- A private chat interface exclusively for your team.
For those who need legal and technical certainty
Bidelagun:
AI audit and regulatory compliance
Know where you are, before deciding where to go.
We audit the cybersecurity of your AI systems, assess how well you are adapted to the EU AI Act and the GDPR, and map out a step-by-step action plan.
How it works
Bidelagun adapts to your situation with two clear approaches:
- Security audit: we review how well protected your system is against leaks or unauthorised access (essential if you work with public administration).
- Regulatory diagnosis: we translate legal requirements into clear, practical protocols for your team, with no unnecessary bureaucracy.
Where it fits best
The security audit is a must in public institutions (where transparency is mandatory). The regulatory diagnosis fits perfectly in industry, education and the public sector, for the peace of mind of knowing things are being done properly.
What it includes
- Security testing of AI / LLM systems.
- Compliance diagnosis (EU AI Act, GDPR, NIS2).
- Creation of documentation and internal protocols.
- A final review session to check progress.
For those who want to empower their team
HazIA Akademia:
AI training for companies
So the whole team understands AI, not just whoever implemented it.
Practical training for the safe everyday use of AI, with a specific compliance block if you need it.
How it works
No tool works if people do not understand it or are afraid of it. That is why we split our training into two adaptable blocks: you can take one or both.
Where it fits best
Block 1 transforms the way of working in industry and professional services. Block 2 is critical for education and public institutions, where knowing how to use AI while respecting privacy matters as much as the tool itself.
What it includes
- Block 1 (Practical use): how to apply AI in everyday work, with real examples from your sector, steering clear of generic theory.
- Block 2 (Regulation): what the EU AI Act requires and how to comply without slowing down productivity or reading 200-page manuals.
- Delivered in Basque and Spanish.
The first step
The first step is to talk
Almost nobody arrives knowing whether they need an audit, a local AI system or simply training for their team. Let us start by understanding where you are right now; we will decide the rest together.