Do you know what company data your colleagues paste into public AI tools?
Your employees already use AI, just free, public tools, and they paste company data into them too.
In use in the first area within 6-10 weeks · a source in every answer · on your own servers or in an EU cloud
This is how HR, knowledge management and IT leaders tend to put it.
“Management asks where AI readiness stands, and I have nothing to show.”
Head of HR and training, large enterprise
“I want to go to the director with numbers, not complaints.”
Head of knowledge management, large enterprise
“We knew employees were pasting into public AI, and we did nothing.”
Head of HR and training, large enterprise
“It turned out everyone had been working from an outdated policy.”
Head of knowledge management, large enterprise
“I need to know where the model runs, what happens to the questions, and who sees what.”
CIO, large enterprise
“My work only shows when it’s missing.”
Head of knowledge management, large enterprise
Our expert will tell you whether it fits your situation.
From then on, they are responsible for it. If there is nothing to show at the next management meeting, or it turns out that someone has meanwhile pasted company data into a public AI tool, it is they who are held to account, not IT. To have something to show, a single area is enough where the assistant already works, answers with sources, and its use can be measured. That says more than a training plan.
59%
of US employees used an unapproved AI tool, according to a 2025 survey, and 75% of them shared sensitive data.
Source: Cybernews, TechRadar, 2025
We index the sources (file shares, email, intranet, DOCMAN) together with the existing permissions.
Colleagues ask in their own words, and see only what they have access to.
The system answers from the hits and tells you which document it found the answer in.
If the sources do not contain the answer, it does not answer. The monthly measurement shows these questions separately.
A corporate knowledge base works if colleagues find the answer in it, and the CIO can say where the model runs, what happens to the questions and who sees what. The assistant is not tied to any office suite, and we have measured how it performs in Hungarian.
At the next management meeting, you can talk in numbers about where the company stands with AI.
Who it is for: the COO or another operations leader who launches it, and the CIO who approves it. In regulated sectors (finance, iGaming, energy) and wherever there is a lot of technical documentation or policy material.
An area where the documents are accessible and have an owner. If the permissions are messy, they need to be put in order first; that is what AI-Ready Data is for.
Introducing the first area takes 6-10 weeks. The price depends on the size of the area and on whether it runs in the cloud or on your own servers.
Our intelligent search, which also works as a chatbot, document processing, extracting the knowledge held in email, AI consulting and training, and our experience in measuring Hungarian language models.
Where an enterprise AI assistant is already in place and the data is in order, we do not replace it. It does not answer from general internet knowledge, and it does not make decisions.
What we did not cover on this page: the steps from a small first step to the complete solution, why it pays to go this way, and how we work along the way. Feel free to forward it to your colleagues.
In 2-4 weeks, the Situation Assessment shows where most time goes on manual work and how many hours a week that adds up to. It costs HUF 199,000 + VAT, which we deduct from the price of your first development. If we do not find at least 20 hours a week of replaceable manual work, we refund the fee.
It answers only from the company’s sources, and gives the source with every answer. If something is not in the sources, it does not invent an answer.
On your own servers or in an EU cloud, as the company decides. The system inherits the existing permissions, so everyone sees only what they have access to.
An area where the documents are accessible and have an owner. If the permissions are messy, we put them in order first.
A policy and training too. We write the company AI use policy, train managers and users, and measure usage every month.
The price consists of a one-off implementation fee for the first area and a monthly operating fee. The monthly fee depends on whether the system runs in the cloud or on your own servers. We give you an exact quote after the Situation Assessment or a conversation with one of our experts.
How does AI find the answer in the company’s own documents, and what happens when colleagues paste internal data into public online tools?
Read what AI readiness means beyond training, or subscribe to the Signal over Hype newsletter.
Describe the situation in a sentence or two, and an expert in the field will get in touch with you.
Or call: +36 30 235 0100
