Virtual Assistant Configuration Guide
Updated at April 16th, 2026
Table of Contents
The Virtual Assistant Module provides intelligent call handling by recognising speech input and routing interactions accordingly. Configuration settings define how the assistant responds, directs conversations, and manages user queries.
1. General Configuration
| Setting | Description |
|---|---|
| Name | The name of the Virtual Assistant instance. |
| Play personal greeting | Greets the caller using their first name or full name when available. |
| Intro prompt | The message played when a call enters the module. For example “You are connected with a Virtual Assistant” followed by “How can we help you?” |
| Multiple language support | Supports up to four languages based on the first four set in the Management Console. The assistant replies in the same language used by the caller. |
| Enable directory assistance | Allows callers to ask for people or departments and be connected or receive availability information. |
| Enable listing of alternatives | If multiple matches are found, the caller is asked to choose. Be aware this will read out names which may conflict with local legislation. |
| Enable missed calls assistance | Enables support for missed call scenarios. |
| Use model | Uses an LLM to recognise caller intent. If no model is set, keywords must be defined. Models require ongoing maintenance as inputs change over time. Azure Conversational Language API key and endpoint must be configured. |
| Azure LLM model or deployment name | Defines the LLM used to identify caller intent. Example gpt 4o mini. Must match the name in the Azure tenant environment. |
| Enable content chat | Allows automatic responses using FAQ content and ChatGPT. |
| Content domain | Defines which FAQ content is used. Topics can be applied to guide responses more tightly. |
| Azure embedding model or deployment | Model used to convert text into vectors for semantic search and retrieval. If AI cannot answer, it can return [Help] and route the caller to the default router. Must match Azure tenant configuration. |
| Azure chat model or deployment name | Model used for live chat responses. Can be the same as the LLM model or a different one for improved output. Must match Azure tenant configuration. |
| Azure chat instruction | Instruction prompt used to guide ChatGPT responses. |
| Allow follow up interaction | Enables handling of follow up questions to create a more natural conversation. May introduce less relevant answers if content goes beyond the defined domain. |
| Try count | Number of attempts allowed before routing to the Exit module. |
| Answered text | Text displayed after the assistant responds. Translation is automatic when using LLM. Use punctuation like “...” to add pauses. |
| Default router | Determines open or closed status, backup availability, and routing when help is needed. |
| Exit module | Used when no input is received or an error occurs. |
2. Topic Actions
Topics allow the Virtual Assistant to trigger predefined responses or call flow actions.
| Topic Action | Description |
|---|---|
| Textual/Audio Response | Provides a direct response (e.g., answering a question about business hours). |
| Menu Follow-Up | Collects additional information (e.g., address or city) via a speech-enabled menu. |
| Transfer to Destination | Transfers the caller to a specific endpoint. |
| Call Flow Action | Links the topic to another Router or Transfer Module, forwarding interactions to internal or external agents. |
| FAQ-Based Response | Uses ChatGPT to generate an answer based on a linked FAQ article (requires Microsoft Azure OpenAI key). |
| Work Item Creation | Creates a work item, collecting details step by step. (Feature to be developed). |
| Workflow Action | Triggers a workflow process for handling a request. (Feature to be developed). |
3. Topic Configuration
| Setting | Description |
|---|---|
| Enabled | Defines if the topic is active and available. |
| Name | The topic name. |
| Suggestion | Offers the topic as a recommended option (e.g., WebChat favourites). |
| Tag | Links the topic to an intent from a model. |
| Inputs | Defines keywords or phrases used to select the topic. |
| Output | The response text if the topic is selected (optional). |
| Action | Defines the action to trigger (optional). |
| FAQ Article | The linked FAQ article, used with ChatGPT for generating responses. |
| Work Item Type | Specifies the work item type created by this topic. |
| Transfer | Defines the transfer destination. |
| Menu | Specifies the follow-up menu. |
| Next Module | Determines the next execution module in the call flow. |
