An AI assistant with your company knowledge.
An internal assistant that answers questions from your own documents: manuals, policies, product data, contracts. With source references, with access rights and, if you wish, entirely on your own server.
Like ChatGPT, but with your knowledge and under your control.
In every company, enormous knowledge sits in documents that nobody can search quickly. An AI assistant makes this knowledge accessible: employees ask a question in plain language and get an answer from your own documents, with a reference to the passage it comes from.
Unlike public AI services, your documents stay under your control. You decide which content the assistant knows, who may use it and where it runs.
- Answers only from your approved documents
- Cites the source with every answer
- Respects who may see which documents
- Runs with EU providers or on your own server
- Says honestly when it does not know something
What companies use the assistant for
Manuals and maintenance
How is error 47 on plant B fixed? The assistant finds the passage in the manual and summarises it.
Internal policies
How many days of leave can I carry over to next year? Answers from works agreements and policies, with references.
Product knowledge for sales
Technical data, variants and arguments at your fingertips, including for new employees.
Customer service
Draft replies from your knowledge base, which a person reviews and sends.
Finding contracts and documents
Where does the framework agreement with supplier X state the notice period? The assistant shows the passage. It does not replace a legal assessment.
Onboarding
New employees ask the assistant instead of five colleagues. The knowledge is equally available to everyone.
From document to answer
Select documents
Together we decide which documents the assistant should know, and we start with a clearly defined area.
Prepare
The documents are processed on the server and made searchable. Access rights are carried over.
Test with real questions
Your team asks the questions that actually come up in everyday work. We refine until the answers are right.
Roll out and maintain
Training for your team, clear rules for use. New documents are added automatically.
What an AI assistant cannot do.
An assistant is only as good as the documents it answers from. Outdated or contradictory documents lead to poor answers. And like any AI, it can make mistakes. That is why every rollout includes tidying up the sources, being open about the limits and training the team in how to use it, for example with our AI training and an AI policy.
Matching guides
Using ChatGPT in a company in compliance with data protection
Which versions are suitable for business use, which data is off limits and what your data protection officer should check.
ReadGuide · 6 minAI agents: what is behind them and what can they do today?
A sober explanation without exaggeration: what agents are, where they help and where they do not.
ReadGuide · 8 min.AI policy for companies: template and explanation
Why every organisation needs clear rules for AI, what belongs in them, and a template to adapt.
ReadFAQ
Which AI is behind it?
We choose the model to suit your requirements: providers with data centres in the EU, or models that run entirely on your own server. Your documents are not used to train third-party models.
Can the assistant give wrong answers?
Yes, that can happen with any AI. That is why the assistant answers only from your documents, always names the source and says when it does not know something. Important answers should be checked against the source.
Does everyone see all documents?
No. The assistant respects access rights. Anyone who is not allowed to open a document gets no answers from it either.
How long does the rollout take?
A pilot with a clearly defined set of documents is often up and running within a few weeks. It is then expanded step by step.
