This article describes the key capabilities of the AI assistant.
Conversational interactions
The AI assistant lets employees communicate with enterprise services through natural language.
They can access the assistant through multiple channels, including:
- Web portals
- Mobile apps
- Collaboration tools such as Microsoft Teams
The assistant maintains conversation context so employees can continue discussions without repeating information.
Employees can:
- Start new conversations
- Continue existing conversations
- Rename conversations
- View previous messages and responses
Responses stream in real time as the assistant generates them.
Knowledge retrieval
The AI assistant retrieves answers from organizational knowledge sources.
When an employee asks a question, the assistant:
- Searches approved knowledge sources.
- Identifies relevant information.
- Generates a response grounded in that information.
- Includes citations that link to the original sources.
Knowledge retrieval follows organizational access policies.
Supported controls include:
- Role-based access control
- Geographic restrictions
- Language preferences
- Effective date filtering
These controls ensure that employees receive only the information they are authorized to access.
Multi-agent task orchestration
The AI assistant coordinates multiple specialized agents to complete complex requests.
A primary orchestration agent interprets the employee’s request and assigns tasks to specialist agents.
Specialist agents can:
- Retrieve knowledge from organizational sources
- Query enterprise systems
- Validate business rules
- Trigger workflows or approvals
- Update records in connected systems
The orchestration agent combines the results and returns a single response.
This allows employees to complete multi-step processes within a single conversation.
Workflow automation
The AI assistant can start and manage workflows directly from a conversation.
Employees can:
- Start workflows
- Continue or resume in-progress workflows
- Cancel workflows
- Receive workflow updates in chat
Workflow events can appear in the conversation so employees can track progress without leaving the chat interface.
Enterprise system integration
The AI assistant interacts with enterprise platforms and APIs.
Depending on configuration, the assistant can:
- Retrieve information from enterprise systems
- Submit requests or transactions
- Execute API calls
- Trigger automated processes
These integrations allow employees to complete tasks through conversation instead of navigating multiple systems.
File and attachment support
Employees can upload files within a conversation.
Supported capabilities include:
- Uploading documents, images, spreadsheets, and media
- Allowing the assistant to read and analyze file content
- Searching within uploaded files
- Extracting text, tables, and structured data where supported
Security protections include:
- Virus scanning of uploaded files
- Blocking unsafe file types
- secure file storage
- Automatic file expiration (typically after 90 days)
These protections help maintain the security of uploaded content.
Feedback and response evaluation
Employees can provide feedback on AI responses.
Feedback options can include:
- Positive or negative ratings
- Identifying responses as harmful, inaccurate, or unhelpful
- Optional written feedback
This feedback helps administrators monitor performance and improve response quality.
Cross-channel access
The AI assistant is available across multiple communication channels.
Supported channels can include:
- Web apps
- Mobile apps
- Microsoft Teams
Conversations remain consistent across channels so employees can continue interactions wherever they work.
Conversation context and history
The AI assistant maintains context across messages within a conversation.
Capabilities include:
- Persistent conversation sessions
- Viewing message history
- Restoring recent conversations
- Maintaining context across interactions
This allows employees to continue conversations without repeating information.
Security and access controls
The AI assistant includes security controls to protect organizational data.
Key protections include:
- Role-based access control
- Conversation isolation
- Input validation and sanitization
- File scanning and security checks
- Data lifecycle and retention management
These safeguards help ensure AI interactions remain secure and compliant with organizational policies.