AI/ML 
Software Development 
Faster Access to Knowledge in a Regulated Environment with AI
Client
GE Healthcare is a global leader in medical technologies, operating in more than 100 countries. The organization operates in a highly regulated environment where processed data is confidential and subject to strict security and compliance requirements.
GE Healthcare’s facility in Norway faced the challenge of improving access to operational knowledge — without the risks associated with using public AI tools and without violating internal security policies.
Challenge
In industrial organizations, knowledge is accumulated over many years in the form of:
- PDF and Word documents,
- Excel spreadsheets,
- Email correspondence,
- Technical and pharmaceutical documentation,
- Data from ERP, CMMS, or MES systems.
These resources are often fragmented, inconsistently labeled, and difficult to search.
Finding an answer to a specific question — for example, regarding a machine’s maintenance history, replaced components, or technicians’ recommendations — may require many hours of manual work.
At the same time:
- The use of public AI models was not permitted due to security concerns,
- Data was not allowed to leave the company’s infrastructure,
- Full source attribution for each response was required,
- The system had to operate in an offline environment (without Internet access).
The objective was to create a solution that provides fast access to knowledge without compromising security or regulatory compliance.
Solution
We designed and implemented a local AI knowledge assistant operating in an on-premises model.
1. LLM-Based Intelligent Knowledge Base
The system leverages Large Language Models (LLMs) to analyze and process internal documents.
The solution:
- Indexes documents in multiple formats (Word, Excel, PDF, TXT, emails, web content),
- Enables full-text search,
- Generates responses within seconds,
- Clearly indicates the exact sources used to generate each answer.
Users interact with the system through a convenient web-based conversational chat interface.
2. Multilingual Support
The assistant supports both English and Norwegian, allowing users to ask questions in natural language while maintaining conversational context.
3. Full Data Security
The system:
- Operates exclusively on internal data,
- Runs within the company’s local network infrastructure (on-premises),
- Has no Internet access,
- Does not share documentation with third parties — not even the solution provider.
The client independently manages document uploads and indexing.
4. Future-Ready Architecture (Tool Calling)
The system architecture allows for future integration with additional enterprise tools, such as:
- ERP,
- CMMS,
- MES,
- Microsoft Teams, SharePoint, OneDrive,
- Courier and order management systems.
Through a tool-calling mechanism, the system can evolve from answering questions to executing specific operations within external systems.
Results
The implementation of the secure AI assistant delivered measurable operational benefits for GE Healthcare:
- Instant access to critical knowledge stored across distributed sources,
- Significant reduction in time spent searching for information,
- Faster and more informed decision-making,
- Reduced costs associated with manual document review,
- Full control over sensitive data thanks to the on-premises architecture,
- Compliance with the strict security requirements of the medical industry.
The system transformed the way organizational knowledge is accessed, turning fragmented archives into an intelligent, interactive source of information available within seconds.
Design, Development, DevOps or Cloud – which team do you need to speed up work on your projects? Chat with your consultation partners to see if we are a good match.
Jakub Orczyk
Member of the Management Board/ Sales Director VM.PL
Book a free consultation