AI in ServiceDesk Plus

Service desks, across organizations, manage high volumes of requests and operational data. As request volumes increase, manual processes can slow down resolution, reduce consistency, and limit visibility into recurring issues. These challenges are addressed in ServiceDesk Plus by Zia, a robust AI engine that helps automate routine service desk tasks and provide operational insights.

Overview of AI Technologies in ServiceDesk Plus

Zia, the AI assistant used by ServiceDesk Plus On-Premises, enhances your service desk experience by bringing AI-powered capabilities through Generative AI, Predictive AI, AI Agents, and a Virtual Bot. It analyzes service desk data and assists technicians and administrators in managing requests more efficiently and consistently. Additionally, Zia reduces the volume of service desk requests by providing automated responses and guidance to users.


In this document, we will discuss how to configure ServiceDesk Plus AI features that are powered by AI and how best to use them to enhance the service desk experience for your organization.

The following features, categorized by modules, use predictive, generative, or agent-based AI to perform specific service desk operations.
 

Configuring AI in ServiceDesk Plus  

Ask Zia (Virtual AI Assistant): Allows users to query service desk data by using natural, simple language. This virtual bot helps users quickly retrieve information and insights without manually searching through the application.

Go to Admin > Zia > Zia Bot > Zia Chat Bot Settings > Ask Zia.

For Ask Zia to work within the application, you must also configure a large language model (LLM), such as ChatGPT. Currently, only ChatGPT is the supported LLM. Check for ChatGPT integration under Admin > Apps & Add-ons > Integrations > Third Party > ChatGPT.

Furthermore, configure MCP tools to secure inadvertent exposure of sensitive data, restrict unauthorized actions, and define access scope for users. MCP tools help an AI agent securely call real functions in the system such as querying data, creating tickets, or triggering workflows. It uses a standard protocol with structured inputs and outputs. They bridge the AI model and the application so the agent can take actions, not just generate text. MCP tools, within ServiceDesk Plus, allow you to perform over 65 operations.

Classic Zia Bot: Interacts with users and answers common questions and guides them through service desk processes. It resolves routine queries automatically.

Configure under Admin > Zia > Zia Bot > Zia Chat Bot Settings > Classic Bot.

Zia Agents: Assists technicians by automating routine actions and providing contextual recommendations. They help improve efficiency by reducing manual effort during request resolution.

Configure under Admin > Zia > Zia Agents.

Also, ensure that ChatGPT is integrated under Admin > Apps & Add-ons > Integrations Third Party > ChatGPT.

Note: Zia Agents can be executed only via Request Triggers.

Data Handling in ServiceDesk Plus  

ServiceDesk Plus shares data with LLMs in a minimal, controlled, and privacy-aware manner. Only data that's necessary for the task is shared, masking or anonymizing sensitive information wherever possible.

  

AI-Powered Request Intelligence   

Validate Request Reopening    

What is this action?

A check that determines whether a closed request is eligible to be reopened. Zia evaluates end-user replies to completed requests and intelligently reopens only those that require further action.

Why does this matter?

Prevents unnecessary reopening, preventing clutter from thank-you notes, out-of-office messages, and acknowledgments, and ensures that only valid issues re-enter the workflow.

How does this work?

Zia evaluates the reply content and predicts whether to reopen the request. When Zia (predictive AI) cannot confidently determine intent, ChatGPT (generative AI) is used as a fallback to analyze the response.

 

Learn more:Request Reopening Validation | ChatGPT Integration - Reopen Prediction

 


Execute User's Approval Action     

What is this action?

An action handler that interprets and executes user approval or rejection responses. Zia scans approval emails and automatically applies approvers' decisions

Why does this matter?

Eliminates delays caused by unclear approval responses and speeds up request processing.

How does it work?

Zia analyzes user responses (Approve/Reject/Need Clarification) and maps them to the appropriate approval action automatically. It scans approval emails and automatically applies approvers' decisions on:

ChatGPT serves as a fallback when Zia fails to arrive at a conclusive/accurate response.
 


Learn more:Email-based Approvals | ChatGPT Integration - Approval Prediction


Category Prediction    

What is this action? 

An AI-powered action that automatically predicts the most appropriate category for an incoming requests.

Why does this matter?

Improves classification accuracy, reduces manual effort, and enables faster routing and resolution.

How does it work?

Zia learns from historical request data and applies pattern recognition to categorize new requests. Zia predicts and suggests the most appropriate request category based on the subject and description.

The minimum historical data required for the training is 500 requests. Suggestions are shown during request creation and can be auto-applied for API and email-created requests.
 


An illustration of category prediction in request form

Learn more:Category Prediction


Template Prediction    

What is this action? 

 An AI-powered action that recommends the most relevant request template even as users raise requests.

Why does this matter?

Ensures consistent data capture, helps users submit complete and accurate requests, and reduces technician follow-ups.

How does it work?

Zia analyzes request context and past submissions to predict the best-matching template. It follows two methods:


An illustration of template prediction in request form

 


An illustration of incident template prediction in Zia Bot

Learn more:Template Prediction - Zia | Template Prediction - ChatGPT


Generate Post-Incident Review       

What is this action? 

An AI-powered action that generates detailed post-incident review reports by using request details, impact, resolution, and recent conversations or notes.

Why does this matter?

Reduces manual documentation effort and supports continuous improvement and compliance needs.

How does it work?

ChatGPT reviews incident timelines, actions taken, and resolution data to generate a structured review. 

AI analyzes the request's subject, requester name, created time, resolution, impact description, and the last 10 conversations and notes to create a comprehensive review including incident summary, root cause analysis, resolution steps, impact assessment, and recommendations.


Learn more:Generate Post-Incident Review


Request Summarization       

What is this action? 

An AI-powered summarizer that provides the synopsis of a request and its conversation history.

Why does this matter?

Helps technicians quickly understand the issue without reading long threads.

How does it work?

ChatGPT extracts key information from request descriptions, notes, and conversations. Provides instant summaries of request details under Show Summary on the request details page.

ChatGPT analyzes the latest 30,000 characters in the description, along with the last 10 conversations and 10 notes to generate a concise summary. For requesters, only visible notes and conversations are included.


Learn more:Request Summarization

Reply Assistant       

What is this action? 

An AI-powered action that helps technicians draft accurate and contextual responses. Helps technicians generate replies, rephrase content, and perform grammar checks while replying to or forwarding requests.

Why does this matter?

Saves time, improves communication quality, and ensures consistent responses.

How does it work? 

AI uses request context and historical replies to generate suggested responses.The ChatGPT icon appears in the text editor for Reply, Reply All, and Forward actions.

Technicians can:


Learn more:Reply Assistant 

 
Email Automation     

Email-Based Automation with Zia Parser       

What is this action? 

An AI-powered email parsing that extracts structured information from incoming emails. Zia Parser analyzes email content and maps key details to request fields using learned patterns.

Why does this matter?

Automates ticket creation and field population, reducing manual data entry.

How does it work?

The Zia Parser processes incoming emails with configured subject phrases (default: @ZIA) and performs automated actions.

When an email containing the configured phrase is received, Zia Parser analyzes the email content by using natural language processing and executes various actions. Multiple actions can be performed through a single email.

The following automated actions are performed through Zia Parser:

Zia Parser is trained daily with the latest 50,000 records to improve accuracy.
 


Zia Parser Configuration
 


Accessing Parsed Properties and Actions via Zia Notifications

Learn more:Zia Parser

AI Agents

L1 Support Agent   

What is this action? 

An AI agent-based action that handles common, repetitive support requests. It acts as a first-level support agent, identifying the most relevant solutions and communicating them to the requester via email.

Why does this matter?

Reduces the technician workload and resolves routine requests quickly.

How does it work?

The L1 agent uses predefined workflows, knowledge base articles, and contextual understanding to respond or resolve requests.

The agent analyzes incoming requests, searches the knowledge base for relevant solutions, and automatically sends solution recommendations to requesters, reducing the load on technicians for common issues.
 


Learn more: L1 Support Agent


Request Resolution Agent  

What is this action? 

An AI agent designed to assist in resolving requests end-to-end.  Generates a concise request summary and adds resolution details directly to the request, helping technicians close requests faster.

Why does this matter?

Accelerates resolution times and improves consistency in handling complex issues.

How does it work?

The agent analyzes request context and suggests a resolution.

The agent analyzes the request conversation and history, creates a summary of the issue and resolution steps, and automatically populates the resolution field to streamline request closure.

 


Learn more: Request Resolution Agent 


Incident Review Summariser  

What is this action? 

An AI-powered summarizing action that generates a concise summary of the incident’s life cycle and resolution.

Why does this matter?

Makes incident insights easily accessible for audits, reviews, and reporting.

How does it work?

The agent condenses incident logs, communications, and resolution steps into a clear review summary. The review is added as a note to the request, helping administrators maintain consistent incident documentation with minimal manual effort.

This agent is configured to trigger after an incident request is closed or resolved.
 


Context-Aware Solution Suggestion     

What is this action? 

An AI-powered action that recommends relevant solutions based on the request context. AI matches request details with historical resolutions and knowledge base content.

Why does this matter?

Increases first-contact resolution and reduces dependency on senior technicians.

How does it work?

When ChatGPT is enabled, Zia Bot analyzes the user's query and generates a summary of relevant solutions from your Solutions module. If internal solutions aren't sufficient, users can explore external sources via ChatGPT.

Learn more: Context-Aware Solution Suggestion

 


Explore Solutions (External Search)    

What is this action? 

An AI-powered capability that searches external sources for possible solutions.

Why does this matter?

Expands troubleshooting options beyond internal knowledge bases.

How does it work?

 AI performs contextual searches on trusted external resources and presents relevant findings.

When searching for solutions in Classic Zia Bot, users can explore external knowledge sources by using ChatGPT through the I'd like to try ChatGPT option.

ChatGPT searches the web to generate relevant solutions based on the user's query. Users can provide additional prompts to refine results.
 


Learn more: Explore Solutions

 

Explore Solutions While Creating a Request    

What is this action? 

An AI feature that suggests solutions to users while they are submitting a request.

Why does this matter?

Deflects tickets and enables users to resolve issues independently.

How does it work?

AI analyzes the request being typed and recommends relevant solutions in real time.

While creating a request using Zia Bot, users can ask ChatGPT to search for solutions from the web. If no solution resolves the issue, users can exit ChatGPT and proceed to create a new request.

While submitting a request, users can use the I would like to try ChatGPT option to search for external solutions that might resolve their issue.
 


Learn more:Explore Solutions While Creating Request


Ask Zia – Conversational Bot      

What is this action? 

Ask Zia, a conversational AI bot assists users with common service desk interactions. AI enables users and technicians to interact with ServiceDesk Plus by using natural language. Users can ask questions and execute actions through simple conversation.

Why does this matter?

Provides instant support and reduces dependency on human agents.

How does it work?

The Ask Zia bot uses predefined intents, workflows, and AI-based understanding to respond to user queries.

Ask Zia leverages ChatGPT as the underlying LLM to understand queries and performs actions. It uses MCP (Model Context Protocol) Tools to access ServiceDesk Plus data and execute 65+ operations including creating requests, assigning technicians, adding tasks, and more.

Note that this available only in the Enterprise Edition.

 You can perform the following operations via MCP tools. 

Module

Subentity

Supported Operations

Requests

Request

Create, update, assign, pick up, close, summarize, and review requests

 

Templates

Retrieve all incident and service templates

 

Tasks

Add, update, assign, comment, close tasks; fetch tasks and task comments

 

Notes

Add, update, and retrieve notes

 

Approvals

Approve or reject request approvals

 

Notifications

Retrieve notifications linked to requests

 

Resolution

Retrieve resolution details

Changes

Change

Create changes; fetch all changes and their stages

 

Tasks

Add, assign, and retrieve change tasks

 

Notes

Add and retrieve change notes

 

Approvals

Approve or reject change approvals

Approvals

Pending Approvals

Retrieve all pending approvals for requests and changes

Announcements

Announcements

Fetch announcements

Solutions

Solutions

Add solutions, search solutions by using RAG; retrieve solutions, impacts, topics, comments

 

Learn more:Ask 

 

Learn more: Ask Zia Conversational Bot | Zia Bot Overview


Classic Zia Bot      

What is this action? 

A conversational AI bot that assists users with common service desk interactions. Zia Bot uses predefined intents, workflows, and AI-based understanding to respond to user queries.

Why does this matter?

Provides instant support and reduces dependency on human agents.

How does it work?

A guided chatbot helps users perform 20 predefined operations through menu-driven interactions, including raising tickets, searching solutions, viewing requests, and managing tasks.

Users select from available actions or type commands. The bot guides them through step-by-step workflows.
 


Learn more: Zia Bot


Ask ChatGPT      

What is this action? 

A generative AI feature that allows users to ask natural-language questions within ServiceDesk Plus.

Why does this matter?

Enables faster access to information, explanations, and guidance without leaving the platform.

How does it work?

ChatGPT processes user queries using contextual information and generates relevant, human-like responses.

Users can directly interact with ChatGPT within Zia Bot by selecting "Ask GPT," entering prompts, and continuing the conversation for general assistance.

Users submit queries or prompts up to 2,000 characters and receive AI-generated responses. This is useful for general questions that may not be directly related to ServiceDesk Plus operations.
 


Learn more:Ask ChatGPT in Zia Bot