The AI assistant generates personalized answers by combining information about the employee, the organization, the conversation, and relevant company content.
This approach helps ensure responses are:
- Relevant to the employee
- Consistent with company policy
- Aligned with the organization’s tone of voice
How personalization works
User profile context
When an employee signs in, the AI assistant retrieves profile information such as:
- Name
- Country or location
- Job title
- Other HR or account details
The system converts this information into user context and provides it to the AI.
The AI is instructed to:
- Use this information only to personalize responses.
- Avoid asking for information that is already known.
- Avoid repeating personal details unnecessarily.
This approach improves relevance while protecting employee privacy.
Organization voice and branding
Each organization can define a theme that includes:
- Company or brand name
- Preferred tone of voice (for example, formal or friendly)
The AI assistant applies this theme so responses reflect the organization’s voice.
Responses follow the configured tone while remaining compliant with company policies and legal requirements.
Conversation context
To maintain continuity, the AI assistant includes a summary of the conversation with each request.
This allows the assistant to:
- Understand previous questions and responses.
- Respond naturally to follow-up questions.
- Avoid asking users to repeat information.
For example, if an employee asks about leave entitlement and then asks a follow-up question, the assistant uses the existing context to respond appropriately.
Knowledge grounding
Before generating a response, the AI assistant searches the organization’s knowledge base for relevant information.
Only relevant content is included in the request to the AI. The AI is instructed to:
- Base responses on the retrieved content.
- Avoid generating unsupported information.
- Include citations where applicable.
This ensures responses are both personalized and grounded in approved company information.
Channel-specific responses
The AI assistant adapts responses based on the channel where the conversation occurs, such as:
- Web chat
- Microsoft Teams
- Mobile apps
Each channel uses a different prompt template so responses match the context.
For example:
- Web responses may be more detailed and formal.
- Messaging channels may prioritize shorter, clearer replies
Time awareness
The AI assistant provides the current date and time as part of the request context.
This allows responses to include time-aware references such as:
- “As of today”
- “This month”
- “For the current policy year”