- 28 Nov 2024
- Print
- PDF
Configure AI Chatbot for Agents
- Updated on 28 Nov 2024
- Print
- PDF
Customize your AI Chatbot for Agents to align its characteristics and responses with your company-wide strategic and business goals. This includes your company’s internal voice and tone, value proposition, and approach to customer satisfaction.
Requirements:
SysAid Spaces
SysAid Copilot license
SysAid Copilot enabled
This document describes how to customize AI Chatbot for Agents (under General Settings), including:
Organization Details
To configure and customize the AI Chatbot for Agents, go to Copilot’s General Settings area:
In SysAid Spaces
Go to Settings > SysAid Copilot > General Settings
In SysAid Classic
Go to Settings > Service Desk > General
Scroll all the way down and click Configure SysAid Copilot Settings.
You can modify the following:
Setting | Where it’s displayed | Default Value |
---|---|---|
Organization Name | Identifies your organization in the AI Chatbot interface | IT Department |
Assistant Name | Specifies the role of the AI Chatbot in its Chat prompts | AI IT Assistant |
Chatbot User Role | Where AI Chatbot’s User Role is identified (Monitor & Fine-tune, Data Pool) | IT Agent |
AI Chatbot Name | AI Chatbot interface | AI Chatbot for Agents |
Prompts
The AI Chatbot for Agents curates its responses contextually based on your organization’s Prompt and Datasets configured only for AI Chatbot for Agents.
Dataset Access
AI Chatbot can access and refer to all Dataset-types available to AI Chatbot for Agents, such as Service Record Datasets, SharePoint Datasets, etc.
AI Chatbot for Agent's contextual responses are also based on the screen the AI Admin has open while conversing with the AI Chatbot:
Active Service Records
Queue (displayed Queue content, according to active Filters)
Variables
You can use different variables as dynamic text that changes according to the user or the use case. This can help personalize the chatbot experience and improve work processes for both your agents and end users.
Pleae note:
Using an unsupported variable in a field will break the experience in the chatbot and cause errors.
List of variables:
Variable | Description |
---|---|
<%= assistantName %> | The name of the assistant |
<%= orgName %> | The name of the organization |
<%= userRole %> | The role of the user (e.g., employee, admin) |
<%= userDetails %> | Details about the user |
<%= relatedText %> | Related answers or context for the current issue |
<%= pageContext %> | The context of the page on which the user is currently operating |
<%= firstName %> | The first name of the user |
<%= orgName %> | The name of the organization |
<%= violatedRules %> | Specific rules that were violated by the user's input |
You can use these variables in all the auto-messages listed below and in the AI Chatbot for End Users:
Guardrails Message
The AI Chatbot’s Guardrails Message increases your organization’s AI Chatbot’s compliance and security of – so that Agents can't misuse or violate its functionalities, permissions, security protocols, and ethical standards.
Your organization’s default Guardrails Message for the AI Chatbot for Agents is the same as the default Guardrails Message for AI Chatbot for End Users, but can be edited.
Chat System Message
Modifying the Chat System Message allows you to construct and modify the Assistant’s ‘personality’ and provide specific instructions about its behavior throughout the conversation.
Task Advisor
Task Advisor uses a prioritization formula (the Task Advisor prompt in the Chat System Message) to identify and recommend the next Service Record the AI Admin should work on.
This recommendation is based on the highest priority ticket in their Queue, as determined by these Fields:
Priority
Due Date
AI Emotion (Request User Sentiment) – coming soon
Learn more about AI Chatbot for Agent’s Task Advisor
Automatic Messages
Your organization’s default Automatic Messages for the AI Chatbot for Agents allow you to craft and customize:
Welcome Message (triggered when the AI Chatbot window first opens)
Starter Message (triggered after the end user clicks the “New Chat” button
Error Message