Matrix42 provides a predefined set of UI AI-powered actions that include:

Detect Category

Detect Category analyzes the selected ticket—primarily its subject and available ticket information—to determine the most suitable category.

After the action is executed, M42 Assist displays:

• The ticket for which the action was run, for example TCK00002 – VPN failed.
• A short explanation of the categorization result, based on the ticket content.
• A Suggested Category field with the recommended category and its confidence level, e.g. Service Desk > Incidents > Security (90%).

The suggested category is shown in a selection field, allowing the end-user to review the recommendation and select a different available category before confirming.

Clicking Apply assigns the currently selected category to the chosen ticket. Clicking Cancel closes the suggestion without changing the ticket category.

Detect Impact

Detect Impact analyzes the selected ticket’s summary and description to recommend an appropriate impact level.

After the action is executed, M42 Assist displays:

• The ticket for which the action was run, for example TCK00002 – VPN failed.
• A brief explanation of the recommendation, considering the scope and potential effect of the issue.
• A Suggested Impact field containing the recommended impact value and confidence level, e.g. Workgroup (80%). 

The suggested impact is shown in a selection field, allowing the end-user to review the AI recommendation and select another available impact level if necessary.

Clicking Apply assigns the currently selected impact to the chosen ticket. No change is made until the end-user confirms it. Clicking Cancel closes the suggestion without updating the ticket.

Detect Type

Detect Type analyzes the selected ticket’s summary, description, and related information to recommend the most suitable ticket type.

After the action is executed, M42 Assist displays:

• The ticket for which the action was run, for example TCK00003 – Multiple Laptops Encrypted by Ransomware!
• A brief explanation of the recommendation, based on the ticket’s content and potential business impact.
• Suggested ticket-type options, such as To Incident or To Service Request.

The end-user can select the appropriate ticket type and review or adjust the related values shown in the selection fields, such as Category, SLA, and Role, before confirming.

Clicking Apply changes the selected ticket to the chosen type and applies the selected related values. No change is made until the end-user confirms it. Clicking Cancel closes the suggestion without updating the ticket.

Detect Urgency

Detect Urgency analyzes the selected ticket’s summary and description to recommend an appropriate urgency level.

After the action is executed, M42 Assist displays:

• The ticket for which the action was run, for example TCK00003 – Multiple Laptops Encrypted by Ransomware!.
• A brief explanation of the recommendation, considering the issue’s severity, scope, and need for a timely response.
• A Suggested Urgency field containing the recommended urgency value and confidence level, e.g. Complete Failure (90%). As you can see from the screenshot, the current Urgency value for the ticket set to “Malfunction” that does not match based on its description, summary suggested by AI recommendation.

The suggested urgency is shown in a selection field, allowing the end-user to review the AI recommendation and select another available urgency level if necessary.

Clicking Apply assigns the currently selected urgency to the chosen ticket. No change is made until the end-user confirms it. Clicking Cancel closes the suggestion without updating the ticket.

Find Solution

Find Solution analyzes the selected ticket’s summary, description, and related context to identify a potential solution. It can be used for incidents, service requests, and other tickets.

After the action is executed, M42 Assist displays:

• The ticket for which the action was run, for example TCK00002 – VPN failed.
• A result header indicating whether a solution was found and the confidence level of the proposed result.
• A brief explanation of the identified issue.
• Recommended resolution steps that the end-user can review before taking further action.

The result header color indicates the confidence level:

• Red: 0% confidence
• Yellow: 1–70% confidence
• Green: 71–100% confidence

The M42 Assist provides actions based on the detected solution. Depending on the result, the end-user can:

• Select Send Solution to Requestor to send the proposed resolution to the ticket requester.
• Select Close Ticket to close the ticket after reviewing the suggested solution.
• Select Cancel to close the result without sending the solution or updating the ticket.

When the end-user selects Close Ticket, the Close ticket wizard opens.

The Solution field is prefilled with the solution proposed by M42 Assist. The end-user can review and modify the suggested text, add a knowledge-base article or attachments if needed, and then select Specify Details to continue with the ticket-closing process.

Selecting Close Ticket does not immediately close the ticket; the ticket is closed only after the end-user completes and confirms the closing wizard.

When the end-user selects Send Solution to Requestor, the Send E-mail for specified ticket wizard opens.

The message field is prefilled with the solution proposed by M42 Assist. The end-user can review and modify the email subject and message, add attachments, and select a signature if required.

The end-user can then select Pause Configuration to continue the email process. Selecting Send Solution to Requestor does not immediately send the email; the solution is sent only after the end-user completes and confirms the email configuration.

Summarize

Summarize analyzes the selected ticket’s title, description, comments, and related context to generate a concise, standardized overview.

After the action is executed, M42 Assist displays:

• The ticket for which the action was run, for example TCK00002 – VPN failed.
• An Overview summarizing the issue, affected service, key symptoms, and current status.
• A Communication section summarizing incoming and outgoing communications for the ticket.
• An Actions Taken section focuses on the actions taken on the ticket.

The summary helps the end-user quickly understand the ticket history and current situation, supporting faster triage and resolution.

Unlike actions that recommend changes to ticket fields, Summarize does not update the ticket or provide actions to apply the result. Selecting Close only closes the summary in the M42 Assist panel; it does not modify the ticket.

Generate KB Draft

Generate KB Draft creates a draft knowledge base (KB) article from information in one or more selected closed tickets. It analyzes ticket titles, descriptions, solutions, and related context to turn resolved issues into structured, reusable documentation.

This action is available only for closed tickets. The end-user can select one or multiple tickets to generate a single KB article draft.

After the action is executed, M42 Assist displays:

• The selected source tickets.
• A suggested title for the KB article.
• A draft introduction describing the issue or topic covered.
• Proposed solution steps based on the selected tickets’ information.

The end-user can review the generated draft before confirming it.

Clicking Apply creates a new KB article in Draft status. M42 Assist then displays a confirmation message with a link to the created KB article and links to the source tickets used to generate it.

Selecting the created KB article link opens a preview of the new article, where the end-user can review, edit, and complete the draft before publishing it. Clicking Cancel closes the generated draft without creating a KB article.

Find Related

Find Related analyzes the selected ticket’s subject and contextual information to identify relevant knowledge base articles and similar historical tickets. It helps the end-user find existing solutions, related incidents, service requests, and self-service resources.

After the action is executed, M42 Assist displays:

• Up to three Knowledge Base Articles that are most relevant to the selected ticket.
• Up to three Similar Tickets with matching symptoms, topics, or contextual patterns.
• Key information for each related item, such as its title, short description, status, and rating where available.

The end-user can select one or more related items. The Generate Solution button becomes available only after at least one item is selected.

Clicking Generate Solution creates a proposed solution based on the selected knowledge base articles and/or similar tickets. The generated result includes recommended resolution steps for the selected ticket.

After a solution is generated, the end-user can:

• Select Send Solution to Requestor to open the Send E-mail wizard with the generated solution prefilled in the message.
• Select Close Ticket to open the ticket-closing wizard with the generated solution prefilled in the Solution field.
• Select Related Items to return to the related-items list, select different items, and generate another solution.

Select Cancel to close the result without sending a solution or updating the ticket.

Personalization AI Preferences

AI Preferences can affect how M42 Assist generates and presents results. The AI Preferences personalization panel can be accessed either through the User menu,

from the M42 Assist panel,

Depending on the available settings, the end-user can configure:

• Response language — the language used for AI-generated content.
• Writing style — the preferred tone and style, such as casual or formal.
• Technical level — the depth and complexity of explanations for the intended audience.
• AI action preferences — settings that control which records are considered by applicable actions, such as the ticket statuses included when finding related tickets or the knowledge article statuses included in results.
• Optionally, users can also configure whether they want to view M42 Assist:
  • Side Panel (default): M42 Assist is integrated into the main workspace as a side panel.
  • Floating Window: M42 Assist appears as a detached, floating assistant window.

After changing preferences, the end-user can regenerate the result to apply the updated settings.

Preventing Unnecessary Executions

If an AI Function is run again using the same ticket information and the same AI Function settings as a previous run, a confirmation step appears before the function executes again. The confirmation step offers three choices: view the existing result, cancel the request, or regenerate the result anyway.

If the relevant ticket information or the AI Function's settings have changed since the previous run, the function executes again without this confirmation step. Additional safeguards also prevent accidental duplicate executions caused by repeated clicks or delayed responses.

This reduces unnecessary consumption of Action Credits from re-running a function on unchanged input, while still allowing an intentional re-run when a different result is wanted.

Migration from AI Assist

The AI Assist actions previously used for summarization, knowledge drafting, finding related records, and solution recommendations have been migrated to Native AI Functions, listed above. These capabilities are now part of the same unified AI experience described in this article.

Related Documentation 

For details on configuring AI service providers, managing AI Functions, supported deployment models, prerequisites, and troubleshooting, refer to: 

AI Functions & AI Service Providers: Configuration, Audit & Troubleshooting 

This article also provides an overview of predefined AI Functions, their configuration, and execution auditing.