AI Functionalities Overview

1. Intelligent Answers (In Search)

This feature instantly transforms any short skills-related question (up to 200 characters) into a structured micro-lesson.

Each response includes:

Practical explanations.
Real-world examples.
Actionable steps.
Checklists.
Coaching questions.
Personalized recommendations.

It provides fast, practical guidance directly connected to on-the-job learning.

2. Intelligent Chat (In Content Page)

This interactive chat appears inside any learning video and allows users to:

Clarify key concepts.
Apply learning to their own context.
Reinforce knowledge through guided reflection.
Complete personalized quizzes.

It transforms passive video consumption into an active, supported learning experience.

Both AI features operate with strict guardrails to ensure they respond only to skills-development topics, maintaining focus, security and relevance.

Privacy, Data Protection & Security

1. How does the system protect user privacy and personal data?

We prioritise Privacy by Design.

No user-identifiable information is ever sent to Azure GPT services. Only the text of the question is processed—without usernames, email addresses, user IDs or metadata.

All responses are generated securely within Microsoft's Azure enterprise environment.

2. What information is sent to Azure GPT services?

Only the content of the user's question.

We do not transmit:

Personal data.
Navigation details.
Behavioural information.

Azure receives only the minimum amount of information required to generate a response.

3. Are interactions used to train the GPT model?

No.

Azure OpenAI does not use your organisation's data to train, fine-tune or improve its models.

All content remains isolated and is used exclusively to generate responses.

4. What information is tracked internally for reporting?

We record that:

A question was submitted.
A response was generated.
The interaction is linked to the authenticated user, similarly to any other learning activity.

However, we do not log or store:

The actual question.
The generated answer.

Reporting is designed to support Learning & Development teams by measuring engagement and learning time—not by monitoring individual conversations.

5. How does this integrate with IT and security requirements?

All AI processing takes place within the Azure enterprise cloud hosted in European Data Centers.

No data leaves the European Union.
Azure provides enterprise-grade security.
Compliance and certifications align with standard IT governance and data protection frameworks.

AI Models, Guardrails & Technical Behaviour

6. Which model powers the assistants?

Both assistants are powered by the GPT-5 model deployed through Azure OpenAI Services, providing high-quality responses while maintaining enterprise-level privacy and security.

7. What guardrails prevent misuse?

The AI is strictly limited to skills-development topics.

Guardrails automatically block requests related to:

Personal advice.
Medical topics.
Legal matters.
Financial advice.
Sensitive content.
Any subject unrelated to learning.

8. Why are questions limited to 200 characters?

Keeping questions short helps to:

Improve clarity.
Increase response accuracy.
Support data-minimisation principles.

This approach enables the rapid generation of focused micro-lessons.

9. Can the AI make mistakes?

Yes.

Like any large language model, the AI may occasionally generate incomplete or inaccurate information.

Users should always verify important or business-critical information before relying on it.

Learning Impact & User Experience

10. How do these assistants support skills development?

The assistants generate:

Structured micro-lessons.
Practical examples.
Learning frameworks.
Checklists.
Coaching prompts.
Personalised recommendations.

These features help users apply learning immediately in their daily work while improving knowledge transfer and long-term retention.

11. Can employees use the AI in any language?