Create a Custom Insight

KaiNexus includes a library of Stock Insights that are ready to use out of the box, but every organization has unique processes, terminology, and goals. Creating Custom Insights allows you to tailor the AI experience to fit your organization's needs.

You can control how an Insight behaves, what information it uses, and how it communicates with users. Use these settings to create Insights that support your organization's workflows and provide relevant, actionable responses.

Getting Started

To create a Custom Insight, you can start from scratch or copy an existing Insight.

Create a Custom Insight

• Navigate to the Admin section and open the KaiNexus Intelligence page.
• Select the Insights tab.
• Select Add and choose Custom Insight from the dropdown.
• Configure the options below and select Save. Note that the Knowledge Base file upload will not become available until after the Insight is saved for the first time.

Copy an Insight

• Navigate to the Admin section and open the KaiNexus Intelligence page.
• Select the Insights tab.
• Hover over the Insight you want to copy and select the Copy icon. In the confirmation window, select Copy to create a new Custom Insight with Access cleared to None.
• Configure Access and the other options below and select Save.

Configuration Options

The following settings control your Custom Insight's appearance, availability, information sources, and responses:

• Basics
• Access & Availability
• Context Sources
• Interaction
• Prompt & Model
• Template Overrides
• Knowledge Base

Click a section header in the Create Insight window to expand or collapse its settings.

Basics

Control how the Custom Insight is identified and described to users.

• Name (required): The name displayed for this Insight across KaiNexus. Maximum 100 characters.
• Description: Enter a brief description of the Insight. This text appears beneath the Insight name when users are choosing which Insight to run. Maximum 200 characters.
• Icon: Select an icon to represent this Insight. When you select one of the listed icons, the icon's code will automatically appear in the Icon Code field. This icon will appear next to the Insight's name across KaiNexus.
• Icon Code: To use a unique icon, select the Font Awesome link to browse available icons, then copy the icon's code and paste it into this field. Leave this field blank if you do not want an icon. You can also clear an existing code and save the Insight without an icon.

Access & Availability

Control who can access the Insight and where in KaiNexus it is available.

• Access: Select who can run this Insight based on Location, individual users, System Roles, Team Roles, or Personas. Users only need to meet one of the selected criteria to run the Insight. No criteria are selected by default.
  
  • Selecting a Location also grants access to users in nested Locations.
  • If you select a Team Role, make sure that Team Role is included on each Template where the Insight is available. Access by Team Role is not available for Item List or Notification Insights.

• Available On: Choose where in KaiNexus the Insight is available: Item, Item List, or Notifications.
• Templates (only for Item Insights): Choose which Templates the Insight is available on. None is selected by default.

Context Sources

By default, Item Insights use the Item's Title, Template, Status, non-private Fields and Attributes, Team members, Locations, Dates, Comments, Resolution information, current Milestone and Stage Gate details on the Item, and a parent reference when the user can access it. The Stock Comments Summary Insight is an exception: it uses the selected Comments, Notes or Updates and their discussion details rather than the broader Item record.

The settings below let you add context sources, including Template and workflow configuration. Enable the sources that support the Insight's purpose, such as workflow details for process guidance or Chart data for analyzing results.

• Include Workflow Context (not available for Notification Insights): When enabled, the AI uses additional information about the Item's workflow and Template to generate its response, including:
  • Template settings: Team Type, available planning options, required Fields, and whether open nested Items prevent completion. This also includes whether Honor Roll, Convert, and Recur are enabled and the available Conversion Types.
  • Resolution and approvals: Resolution and approval expectations, plus the Template's configured Impact Types, No Change Reasons, Third Option Reasons, and Investment Types.
  • Milestones and Stage Gates: The Template's Milestone and Stage Gate configuration, including Milestone dates and approvers.
  • Date settings: Whether Created and Completed Dates can be edited, date restrictions, whether Due Dates update parent or child Items, and minimum and maximum Due Date time frames. This also includes which Team Roles can edit Due Dates and whether the Responsible person can edit the Item's Due Date.
  • Item relationships and progress: Child Items, dependencies, and the Item's Status and lifecycle stage.
    • Direct children: Visible, incomplete direct children only. Includes ID, Title, Status, Due Date, and whether the child is overdue. Full child contents and further descendants are not included.
    • Downstream dependencies: Visible, incomplete Items that depend on the current Item. Upstream dependencies (Items the current Item depends on) are not included.

• Include User Details (not available for Notification Insights): When enabled, the AI uses the current user's name, Team Roles, System Roles, permissions, Persona, Locations, and User Attributes, including Titles, Positions, Certifications, and Employment Status, as additional context. This is enabled by default for new Custom Insights. Turning it off does not remove names already included in Team information or Comments.
• Include Item Images (not available for Item List or Notification Insights): When enabled, the AI analyzes image attachments on the current Item as part of generating the Insight. Other attachment types are not included by this setting.
• Include Chart Data (not available for Item List or Notification Insights): When enabled, the AI incorporates the current Item's Chart data into its analysis.
• Include Fishbone Data (not available for Item List or Notification Insights): When enabled, the AI incorporates the current Item's Fishbone data into its analysis. Links in the Fishbone do not retrieve the full contents of linked Items.
• Include Citations (only for Item List and Notification Insights): When enabled, initial and follow-up responses include Show Sources controls that reveal the Items behind that part of the response.
• Include Item Links (only for Item List and Notification Insights): When enabled, recognized Item IDs in responses are clickable and open the Item in a new browser tab. Item titles without an Item ID are not linked, and links do not grant additional Item access.

Interaction

Control how users continue the conversation and when they are encouraged to run the Insight.

• Enable Replies: When enabled, users can reply to the Insight after the initial AI response to continue the conversation. Enable this when users may need to ask questions or explore the response further.
• Enable Suggested Follow-up Replies: When enabled, the AI displays suggested follow-up replies after its initial response.
• Enable Nudge: Select this checkbox to add a Nudge to an Insight that appears on Items. A Nudge is a prompt that appears when an Item meets configured criteria and encourages users to run the Insight. Then configure the following settings:
  • Nudge Message: The message displayed to users when the Nudge appears. Keep it short, clear, and action-oriented. Describe the help the Insight offers so users understand why they should run it.
  • Nudge Criteria: The criteria an Item must meet for the Nudge to be displayed.
    • The check uses a shortened version of saved Item information and the Insight’s Include Workflow Context, Include User Details and Context Size settings.
    • The check does not include the Insight’s Prompt, Preamble, Knowledge Base, Item Images, Chart Data or Fishbone Data. Put conditions for showing the Nudge in Nudge Criteria.
    • When the user selects View Insight, the Insight runs using its normal configured sources.

Prompt & Model

Define what the Insight should do, how it should respond, and the AI resources it uses to complete the task.

• Model: Choose Efficient for routine tasks with the lowest AI Credit usage, Balanced for more complex analysis with moderate credit usage, or Advanced for demanding tasks with the highest reasoning capability and credit usage. Changing the Model applies to new conversations; an ongoing conversation continues using its original Model. Nudge Criteria are evaluated using Efficient; the Insight uses the configured Model.
• Thinking Level: Controls how deliberately the AI reasons before responding. The options are Automatic, Minimal, Low, Medium, High.
• Context Size: Choose Small (the default), Medium (3 times Small), or Large (6 times Small) to control how much textual source information the AI analyzes. Larger sizes can help with complex inputs but may use more AI Credits.
  
  This setting does not increase limits for the Prompt, Preamble, Knowledge Base files, selected files or images, Chart or Fishbone data, conversation history (the exchange between the user and the AI), or the three Comment per Item limitation for Item List Insights. 
• Prompt (required): Defines the instructions the AI uses to generate the Insight. For Item-level Insights, use #itemName or #itemId to reference the current Item.
• Preamble: Optional background context or tone to shape how the AI responds.

Compare response quality, response time, and AI Credit usage during testing to choose Model and Thinking Level settings that fit the task.

Template Overrides

For Item Insights, Template Overrides let you replace the default Prompt and optionally the Preamble for specific Templates. This is useful when an Insight needs to behave differently depending on the type of Item it's being used on. Leave the override Preamble blank to use the default Preamble.

To add a Template Override, expand Advanced, then:

• Select Add within the Template Overrides section.
• Choose one or more specific Templates to override. The Show All option is not available.
• Configure the Prompt and, if needed, the Preamble for those Templates.
• Repeat as needed for additional Template groupings, or select Remove to delete an override.

Knowledge Base

The Knowledge Base allows you to upload reference documents the AI can draw on when generating responses for this Insight.

Use it when the Insight needs organization-specific definitions, standards, or procedures. See Build an Effective Knowledge Base for guidance on preparing these documents.

To upload reference documents:

• Save the Insight for the first time to make the file upload field available.
• Edit the Insight and expand Advanced.
• Select Add File in the Knowledge Base section to upload a file. Accepted file formats are PDF, plain text, markdown, JSON, or XML.
• Select Save when you finish.

Test Your Insight

You may need to adjust your Insight before it consistently produces useful responses. After saving, run it where you've configured it to appear and try it with different information, including cases where details are missing or conflicting.

Check whether the response completes the intended task, uses the available information accurately, and provides the level of detail you need. If it falls short, review the context settings and references, then adjust one important setting or instruction at a time and test again.

After changing the configuration or test data, use Refresh to generate a new response using the current settings and information. Reopening a saved response or continuing an existing conversation may use earlier information.

If replies are enabled, test a follow-up question as well as the initial response.

For more guidance on evaluating and improving responses, see Write Effective Prompts and Preambles.