Our AI statement

How the assistant works

Our AI assistant answers questions by retrieving information from our own published website content e.g. articles, service descriptions, FAQs and technical resources. It does not generate answers from general internet data or third-party training sets. Every response is grounded in content we have authored and reviewed.

The system uses retrieval-augmented generation (RAG), meaning it searches our content library for relevant passages and uses a large language model to formulate a helpful response based on those passages.

The retrieval of information is executed through a mixture of graph, keyword and semantic retrieval which are then re-ranked temporal and by relevance.

Hosting infrastructure

Unlike many AI-powered services, our assistant runs entirely on infrastructure we own and operate. There are no third-party AI cloud services involved in processing your queries

• Local inference:
  The language model runs on our own hardware. Your questions are not sent to OpenAI, Google, Microsoft or any other external AI provider.
• Local storage:
  The content library and operational logs -- (all our 'AI server data') is stored on our own servers.
• No external dependencies:
  The complete AI stack operates within our private network with no outbound data transfer to third-party AI services.

This architecture gives us full control over how our knowledge is processed and ensures the interactions remain within our infrastructure.

What data we collect and what we do not

We do not collect or store:

• Personal names, email addresses or contact details through the assistant
• IP addresses or device fingerprints linked to conversations
• Any data that could identify you as an individual

We do collect or store:

• Anonymous session metadata (session count, timestamp of last activity, number of messages exchanged) for operational monitoring and service improvement
• Aggregated, non-identifiable usage analytics to understand which topics are most helpful
• Feedback comments provided by the user are stored for a 90 day period

Our system follows a privacy-first design where session analytics contain no message text and no personally identifiable information.

How we handle your message history

The AI assistant uses browser local storage to store the following information:

• Your communication preference [edx_emiel]
• Your session chat history [edx_emiel_chat_history< guid >]

This information can be deleted at any point by clicking the recycle bin icon within the chat interface, or by emptying your browsers local-storage.

GDPR compliance

As an EU-based organization, we are committed to full compliance with the General Data Protection Regulation (GDPR).

• No personal data processing:

  The assistant does not process personal data as defined under GDPR Article 4. No user-provided information is stored, profiled or used for any purpose beyond generating an immediate response.

• No profiling or automated decision-making:

  The assistant provides informational responses only. It does not make decisions about individuals or create user profiles.

• Data minimization:

  We collect only the minimum operational metadata necessary to maintain the AI service reliability. This data cannot be linked to any individual.

The assistant is not a substitute for professional consultation. We urge you to be privacy conscious and aware while sharing information with our AI service.

Content accuracy

The assistant responds based on our published content. While we strive for accuracy, AI-generated responses may occasionally be incomplete or lack nuance. For decisions related to platform selection, migration planning or technical architecture, we recommend speaking with our team directly.

If you have questions about our use of AI or how we handle data, please contact us at [email protected].