Generative AI

Generative AI, a cutting-edge subset of Artificial Intelligence (AI), is revolutionizing the way we interact with technology through its ability to create content autonomously. Copilot powered by Generative AI empowers users to focus on higher-level decision-making and innovation.

By leveraging advanced Machine Learning models, Enterprise IT's Copilot understands context, anticipates user needs, and generates relevant suggestions or content. This intelligent support system:

• Enhances productivity and creativity
• Streamlines workflows
• Reduces cognitive load

Incident Clustering

Enterprise IT's Copilot leverages advanced machine learning techniques, algorithms, and large language models to categorize incident data into clusters based on similarities in symptoms, descriptions, priority levels, and other parameters. This proactive approach helps you identify potential problems, potential major incidents, major incidents and other priority 1 incidents.

To view the Incident Cluster, perform the following steps:

1. Navigate to Incidents > Manage Incidents from the left menu. The list of Incidents in displayed.
2. Clickon the top of list. Refer the below screenshot:
   Figure: Incident Clustering button
   A cluster of all the Incidents is displayed in the form of bubbles. Refer the below screenshot:Figure: Incident Cluster

Details about Incident Cluster

• Each bubble in the Incident Cluster contains Incidents with similar symptom.
• The outer circle of a bubble represents the common incident symptom. When you hover over the outer bubble, the tooltip displays the common Incident symptom of the Incidents contained in that bubble. Refer the below screenshot:Figure: Outer circle tooltip in bubble
• Each of the inner circle in a bubble represents an Incident. When you hover over any of the inner bubble, the tooltip displays details about that Incident (Ticket ID, Symptom, Description). Refer the below screenshot:
  Figure: Inner circle tooltip in bubble
• Clicking the inner circle of a bubble will open that particular Incident record in a new tab.
• Clicking the outer circle of a bubble will open the list page of all Incidents with the same Incident symptom.

The Incident Cluster screen lists out all the Incident symptoms with the color codes on the left side of the screen. Refer the below screenshot:Figure: Incident symptoms list

Filter Incident Cluster

Filter the Incident Cluster by clicking on a particular Incident symptom on the left side list. This will strike off the Incident symptom from the list on the left side. Subsequently, the bubble of that particular Incident symptom is removed from the cluster. Refer the below sample screenshot:

Figure: Filter Incident cluster

Click to view different options to download the Incident Cluster diagram. You can click the relevant option to view the Incident cluster in full screen, print it in chart format, or download it in different formats (PNG, JPG, PDF document or SVG vector image). Refer the below screenshot:

Figure: Image actions

Priority 1 (P1) Incidents

The inner bubbles are color coded to denote the priority levels. All the Incidents with P1 priority are highlighted with red circular border. The Incident Cluster chart includes an indicator for P1 incidents at the bottom, marked with a red highlight border. Refer the below screenshot:Figure: P1 Incidents indicator

Potential Major Incidents

Incident Cluster highlights the potential major incidents in a orange color bubble in case of the following conditions:

• If there are more than three Priority 2 Incidents of the same category, and;
• If there are Incidents related to system crash, slow performance, or data loss.

The Potential Major Incidents are also marked in the left side list, with details shown in brackets. This helps in prioritizing critical issues, understanding the scope and impact of incidents, and making informed decisions for effective incident management and resolution. Refer the below sample screenshot:

Figure: Potential Major Incidents

When you hover over the orange color bubble, then the tooltip displays the following information:

Figure: Potential Major Incidents - tooltip

Potential Problem Detected

Incident Cluster detects and highlights Potential Problems if there are three Incidents grouped under same category with a semi-transparent red color bubble. Refer the below sample screenshot:

Figure: Potential Problem Detected tag

When you hover over the red color bubble, the tooltip displays details about the Potential Problem Detected along with the Incident symptom. Refer the below screenshot:

Figure: Potential Problem Detected - red bubble

Major Incidents

Incident Cluster flags a ticket with Major Incident flag. Flagging a Major Incident in the Incident Cluster provides a clear and immediate overview of critical issues within the context of all incidents. This helps prioritize response efforts and streamline communication among support teams. Additionally, it offers a comprehensive understanding of the incident landscape, enabling faster resolution of high-impact problems and minimizing downtime.

RCA Summary

Copilot generates the RCA (Root Cause Analysis) summary for a Problem Record and provides a concise overview of the underlying causes of recurring issues. It leverages data analysis and machine learning to identify patterns and root causes from past issues, presenting the findings in an easily understandable format. This helps teams to quickly grasp the essential factors contributing to problems, facilitating more effective resolution and prevention strategies.

To view RCA Summary, perform the following steps:

1. Navigate to Problem > Manage Problem Records. List page of all the Problem Records is displayed.
2. Click on the Problem Record ID for which you want to view the RCA Summary. The details page is displayed.
3. Change the status of the Problem Record to Resolved or In Progress.
4. Navigate to Solution tab. Click Generate RCA Summary.
   

   Figure: Generate RCA Summary button
   
   A pop-up displays RCA summary. Refer the below screenshot:
   Figure: RCA Summary
   Click Append to the Solution if you want to change the RCA summary. Also, if you want to copy the RCA summary, click Copy the Solution (Replace).

Resolution Summary

The Resolution Summary feature of Copilot significantly enhances analyst productivity by transforming brief input notes into comprehensive and well-phrased resolution summaries. This functionality ensures that analysts can quickly generate detailed and accurate resolution notes without spending excessive time on writing and editing. By automating the creation of these summaries, Copilot helps maintain high-quality documentation that can be seamlessly integrated into knowledge articles.

This reduces the need for extensive review and editing, ensuring that information is consistently accurate and detailed. Consequently, enterprises benefit from improved documentation practices, streamlined workflows, and a more efficient transfer of knowledge within the organization. This not only saves time for analysts but also ensures that valuable insights and resolutions are readily accessible for future reference.

To view Resolution Summary, perform the following steps:

1. Navigate to Incident > Manage Incidents. The list of all Incidents is displayed.
2. Open an Incident by clicking on the Incident ID on the list page.
3. Change the status of Incident to Resolved.
4. A button to Generate Resolution Summary appears under the Solution tab.
   Figure: Generate Resolution Summary button
5. Click Generate Resolution Summary. A new pop-up displays the resolution notes.
   Figure: Resolution Summary pop-up
6. Click Override the Solution to reject or modify the generated Resolution Summary. Click Append the Solution to add the generated Resolution Summary to the record. You can also integrate the generated output with other information to create a more comprehensive or complete record. Click Cancel to close the Resolution Summary pop-up.

The generated Resolution Summary can be leveraged to create knowledge articles on Apex ITSM App and follow the rest of the knowledge article publishing process.

Major Incident Summary

Copilot generates a comprehensive Major Incident summary by thoroughly analyzing all data and interactions associated with the Incident. The summary provides a detailed overview of the major incident, outlining the key events and timeline. It captures essential information to give stakeholders a clear understanding of what was the issue and how it was addressed and mitigated.

Generative AI capability of Copilot also identifies and describes the root cause of the incident, pinpointing the underlying issues that led to the problem. This helps in understanding why the incident happened and aids in preventing future occurrences. Copilot evaluates the extent and severity of the incident's impact on business operations, customers, and other stakeholders. This analysis helps in assessing the damage and prioritizing future responses.

Copilot also suggests actionable steps to prevent similar incidents in the future. These recommendations are based on the analysis of the root cause and the resolution process, ensuring continuous improvement in incident management practices. This holistic approach ensures that all relevant information is captured, analyzed, and presented in an accessible manner, enhancing incident management and resolution strategies.

To view Major Incident Summary, perform the following steps:

1. Navigate to Incident > Manage Incidents. The list of all Incidents is displayed.
2. Open an Incident by clicking on the Incident ID on the list page.
3. Click General tab. Enable Major Incident switch.
4. Change the status of Incident to Resolved.
5. A button to Generate Major Incident Summary appears under the Solution tab.Figure: Generate Major Incident Summary button
6. Click Generate Major Incident Summary. A new pop-up displays the Major Incident summary details.
   Figure: Major Incident Summary pop-up
7. Click Print Summary to print the details of Major Incident. If you want to copy the Major Incident summary, click Copy to clipboard.

Generate Insights

The Generate Insights feature showcases the remarkable ability of Copilot to enhance the productivity of Analysts by providing concise and comprehensive summaries of Incidents. This feature encapsulates critical details, including an Incident Summary, which offers a clear overview of the situation at hand. It also includes a record of investigations performed so far. This saves Analysts' time by presenting a quick review of past actions.

Additionally, it evaluates Sentiment in terms of confidence score by analyzing user feedback and communication tones, giving Analysts a sense of user satisfaction levels. The AI Suggestion component leverages advanced algorithms to propose actionable recommendations, aiding Analysts in making informed decisions swiftly. Copilot also provides actionable advice - Sentiment booster tip - to improve sentiment, helping to maintain positive stakeholder engagement. It also assesses the potential risk of escalation, identifying issues that could lead to increased dissatisfaction or complaints. So, Analysts can prioritize and address these concerns proactively. Generate Insights feature is an indispensable tool for Analysts to streamline their workflow and enables more effective incident management.

To view the Incident Summary through Generate Insights, perform the following steps:

1. Navigate to Incident > Manage Incidents > Incident List. The list of all Incidents is displayed.
2. Open an Incident by clicking on the Incident ID on the list page. Details screen of the Incident is displayed.
3. Click icon, which displays the dropdown menu:
   Figure: Dropdown menu
4. Click Generate Insights. A pop-up is displayed shows insights about the Incident:
   Figure: Generate Insights pop-up
   
5. Click X to close the Insights Summary pop-up.

Enterprise Knowledge Search

Applicable Channels: Web, MS Teams

Enterprise IT's Copilot serves as a valuable resource for employees by providing instant access to enterprise-related knowledge, such as company travel policies and other organizational guidelines. By leveraging Natural Language Processing and Machine Learning, Copilot quickly retrieves and summarizes relevant information from company documents and manuals.

Such immediate access to accurate and up-to-date information streamlines decision-making and ensures employees adhere to company policies. This enhances efficiency and compliance across the organization.

One of Copilot's most powerful features is its ability to have advanced question-answer conversations. It uses Retrieval Augmented Generation (RAG) to answer questions based on specific information sources.

To find enterprise related knowledge and information through Copilot in web application, perform the following steps:

1. Clickat the bottom right corner to start conversation with Copilot.
2. Provide the input to Copilot for which you need information. For example: "what are the top three company travel policies?".
   Copilot displays the relevant information. Refer the below screenshot:Figure: Enterprise Knowledge - Web

You can also find enterprise related information through Copilot on MS teams:

1. On MS Teams, open the chat window with Copilot.
2. Provide the input to Copilot for which you need information. For example: "what are the top three company travel policies?".
   Copilot displays the relevant information. Refer the below screenshot:
   Figure: Figure: Enterprise Knowledge - MS Teams

Agent Chat Summary

Copilot has the capability to improve the productivity of the Live Agents (human agents) by generating a summary - Live Agent Insights and Actions - of the conversation they are handling.

The Live Agent Insights and Actions summary includes the following information:

• Chat Summary
• Service Summary
• Recommended Next Steps
• Actions
• Create ticket

Live Agent Insights and Actions summary automatically condenses and highlights key points of the ongoing chats being handled by Live Agents. This helps the Agents to quickly understand the context and essential information about the chat they are handling without having to manually review the entire conversation.

To view Agent Chat Summary, perform the following steps:

1. Login to the application.
2. Click on Live Agent iconfrom the top panel, which opens the Live Agent Hub in a new tab.
3. All the active Live Agent Transfer requests appear under New Requests section. Click on the name of the end user with whom you want to start the conversation.
4. Once you click on the name of the end user, the complete chat details are populated in the middle panel.
5. Clickto generate Generate Chat Insights as shown in the below screenshot.
   Figure: Generate Chat Insights button
6. A pop-up displays all the details related to the chat. Refer the below sample screenshot:Figure: Agent Chat Summary pop-up

Below table describes all the information present in the Agent Summary pop-up:

Section	Description
Chat Summary	Chat Summary provides a concise overview of the interaction between Digital Agent and the end user along with the end user’s satisfaction level. It captures key points, issues faced by the user, and relevant information exchanged during the conversation. This summary offers a snapshot of the chat’s essential aspects, enabling quick comprehension of the conversation without the need to review the entire conversation. The end user's satisfaction level is displayed in the Chat Summary, which is represented using emojis and color coding as explained below: Green - Indicates high satisfaction level. Refer the sample screenshot of the chat summary below: Figure: Green color coded sentiment Yellow - Suggests moderate satisfaction level. Refer the sample screenshot shown below:Figure: Yellow color coded sentiment Red - Denotes low satisfaction level. Refer the below sample screenshot:Figure: Red color coded sentiment
Service Summary	The Service Summary section provides a summary of all the tickets raised by the end user the Live Agent is currently talking to. It offers key details such as ticket numbers and a summary of the issues or requests associated with each ticket. This enables Live Agents to quickly assess the user's history, understand their issues, and prioritize responses accordingly.
Recommended Next Steps	Recommended Next Steps section in the Live Agent Insights and Actions pop-up provides actionable recommendations for addressing the tickets raised by end users. Leveraging AI, advanced algorithms and historical data, our system analyzes the nature of each ticket and proposes potential or next steps that Live Agents can take to resolve the issue efficiently.
Actions	Copilot suggests Actions when it detects that the user is experiencing an issue and needs to log an Incident. By analyzing the user's input, the Copilot identifies the problem and recommends steps to create and manage an incident report efficiently. Symptom: This section mentions the issue that the user is facing and wants a solution.  Description: Copilot suggests a description of the issue by analyzing the conversation between itself and the end user. As the user describes their problem, Copilot interprets and formulates a clear, concise issue description to ensure accurate and efficient resolution.
Create Incident	Click this button to create an incident for the issue that the end user is facing.

Personally Identifiable Information (PII) Data

Enterprises frequently deal with sensitive Personally Identifiable Information (PII) in their communications. Accidentally sharing PII can lead to serious compliance, privacy, and security issues. To prevent these risks, it's essential to identify PII, and our system can detect and mask PII data in real-time.

Copilot has a PII detection system by leveraging natural language processing (NLP) techniques and machine learning algorithms to identify and mask sensitive PII data in the Apex application. The system will integrate seamlessly with Apex application used within the organization, providing real time monitoring and intervention.

Key features of PII Data detection system

• Real-time PII Detection: Uses Natural Language Processing (NLP) techniques and machine learning models to analyze text input in real-time. The system will constantly monitor and identify occurrences of PII, such as names, addresses, phone numbers, email addresses, social security numbers, and more.
• Alerting and Actions: When PII is detected, a pop-up alert notifies the user to prevent accidental exposure. If the user attempts to save the record, the system shows a pop-up message identifying the PII data and provides two options. This alert prompts the user to review the sensitive information and choose whether to mask it or give consent for its use.
  • Mask and Submit: The masking option signifies the user's choice to have the detected PII replaced with placeholders or generic descriptors to protect privacy using the standard masking techniques.
  • Give Consent and Submit: The consent option indicates the user's consent to use the detected PII as it is without redaction and submits the data.
• Audit Trail and Reporting: The system keeps a record of detected PII instances and the actions taken. This allows organizations to prove compliance and investigate incidents if needed.

Refer the below steps on how PII Data identification functionality works in a Incident or Service Request creation flow:

1. Navigate to Incident > click My Incidents. The list page of all the logged in Incidents is displayed.
2. Click New to log a new Incident. Details page is displayed. Fill in all the details. If there are any personal information in any of the fields and you click Save, then a pop-up displays all the PII data. Refer the below screenshot:Figure: PII Data pop-up
3. There are two options: Give Consent and Submit, and Mask and Submit. Click Give Consent and Submit, if you want to provide your consent to use the details and create the ticket. If you want to create the ticket by masking the details and then create the ticket, click Mask and Submit.
4. Everywhere in the application the details are masked for this record. Refer the below screenshots:
   Figure: Masked details
   
   Figure: Masked details in list page