Topics Overview

Overview

Topics represent subject areas that users can ask questions about in Ask Insights, a generative BI feature. Topics are made up of fields from one or more datasets and can include customizations such as synonyms, suggested questions, and custom instructions.

Once you’ve set up a topic, users can ask questions about the topic and receive answers in the form of AI-generated summaries and visuals. You can then review the topic’s performance, read and address user feedback, mark answers as verified, and refine the topic to provide more accurate and relevant answers.

Managing Topics

To manage a topic, select Insights > Dashboards Home in the left navigation. Then select Topics from the left menu.

On the topics list, the following information is displayed:

• Name—The name of the topic. This name also appears in the Ask Insights side panel.
• Owner
• Status—The date the topic was last refreshed. This column may also display the progress of creating or refreshing a topic.
• Last Modified—The date the topic was last edited.
• Usage—How many questions have been asked of the topics and how many thumbs up, thumbs down, and comments the topic has received.
• Actions—The following actions are available:
  • Edit topic
  • Update
  • Delete
  • Duplicate
  • Add to folder

Select a topic from the topics list to manage its settings. The following tabs and subtabs are available.

Summary Tab

Select the Start Review button in the Improve your topic section to begin a guided walkthrough of refining a topic. If you’ve already started a review, this section displays the review steps you’ve completed and need to complete.

The Topic details section displays the name, description, last modified date, last refresh date, and generative BI options. You can edit the topic name and description from here.

The Statistics section displays analytics for the topic. You can view a Feedback over time or Total distribution of feedback chart. Select the Options button to change the information shown in the graph.

Data Tab

Datasets Subtab

In this subtab, you can add, remove, and manage the datasets for the topic. The datasets added to a topic determine the data fields that are available to the topic.

Select the More button for a dataset to refresh the dataset or remove it from the topic.

Select the Expand button for a dataset to set the dataset’s default date dimension—which is used when answering user’s questions in the Ask Insights feature—or to add a refresh schedule.

Data Fields Subtab

In this subtab, you can include or exclude data fields, add synonyms that map Kantata field names to your organization’s specific terminology, configure fields, add calculated fields, and add filters.

Named Entity Subtab

In this subtab, you can add named entities, which support the informal names that users may use to refer to multiple data fields. For example, your users may say “task estimates” when referring to both task estimated hours and the task budget. Adding named entities makes generative BI features more accurate.

User Activity Tab

In this tab, you can see analytics for the topic, view the questions users have asked, mark responses as verified, and review user feedback and comments. The information in this tab helps you refine and improve the topic. For example, you can review unanswerable questions and evaluate if more synonyms, named entities, or data fields need to be added to the topic.

To verify a response, select the View button for a question in the table. A side panel opens. Select the Mark as verified button.

Suggested Questions Tab

Verified Subtab

This subtab shows a list of verified responses, which are also shown as suggested questions in Ask Insights.

Select the Expand button to see the analytics for a verified response.

Select the View button for a question to preview the response. In the side panel that opens, you can change the visual type of the visuals in the response or unmark the response as verified.

Select the Add custom message button to display a custom message in Ask Insights when certain keywords are included in a question. Custom messages provide guidance to users on what data and questions are supported for specific topics. For example, you can let users know that questions about certain subjects are not supported by a specific topic or redirect users to a different topic for certain keywords.

AI Generated Subtab

This subtab shows the list of AI-generated questions that are shown as suggested questions for the topic in the Ask Insights side panel.

Select the Options button to show or hide all AI-generated suggested questions or regenerate the questions.

Select the Expand button to see the analytics for an AI-generated suggested question or remove the question.

Custom Instructions Tab

In this tab, you can add more context to the topic to improve the accuracy of the Ask Insights feature. We recommend first adding context through synonyms, field descriptions, the topic description, and dataset descriptions before adding custom instructions.

For best practices for writing custom instructions, see this external article.

Standard Topics

The following topics are available by default:

• Fees—Supports questions about fees (actual, allocated, scheduled, and projected), margin, over/under budget. You can ask about data by project, client, or time period.
• Revenue Recognition—Supports questions about recognized revenue, accrued and deferred revenue, projected revenue recognition. You can ask about data by project or month.
• Invoicing—Supports questions about invoice amounts (paid, unpaid, and overdue), invoiceable amounts, current AR, opening AR, and last month's collections.
• Utilization—Supports questions about utilization rates (allocated and actual), billable vs. non-billable hours. You can ask about data by user, role, or manager.

How to Create a New Topic

1. In the left navigation, hover over Insights, then select Dashboards Home.
2. Select Topics.
3. Select Create topic, then select Create legacy topic.
4. Enter a Topic name and Description.
5. Select Continue.
6. Select a dataset. This will determine the data fields available to the topic.
7. Select Create.
8. Wait for the topic to finish creating. The topic opens.
9. Make the following updates in the Data tab:

   1. Add more datasets in the Datasets subtab if desired.

      Note: If you don’t see the Datasets subtab, try clicking on a different tab, then click the Data tab again.

   2. Include or exclude Data Fields in the Data Fields subtab.

      Tip: To ensure the quality of AI answers, limit the data fields you include.
   3. Add Synonyms for included data fields to help the AI understand users’ natural language questions.
   4. In the Named Entity subtab, group related data fields with a term or phrase that users may refer to. For example, the project lead, project name, project budget, and actual fees fields can be grouped together with the phrase “project information”.
10. Select Open Q&A to test the Ask Insights feature for the topic.

As people ask questions of the topic and submit feedback, you can review performance, read user feedback, and refine the topic.

How to Add Synonyms for Fields

Synonyms map Kantata field names to your organization’s specific terminology, ensuring the AI understands your organization’s specific language in natural language prompts. For example, the “Invoice Amount Field” field can have a synonym of “cash collected”.

1. In the left navigation, hover over Insights, then select Dashboards Home.
2. Select Topics.
3. Select a topic.
4. Select the Data tab.
5. In the Data Fields subtab, locate the data field you would like to add a synonym to. The Include toggle should be activated for the data field.
6. In the Synonyms column, select the Edit icon. If the data field already has synonyms, this will be an Add icon instead.
7. Enter a keyword.
8. Press Enter on your keyboard or click outside of the field to save the synonym.

How to Verify Responses and Add Suggested Questions to Ask Insights

When you verify responses, the questions are displayed as suggested questions in the Ask Insights feature, helping users get started with a topic.

1. In the left navigation, hover over Insights, then select Dashboards Home.
2. Select Topics.
3. Select a topic.
4. Select the User Activity tab.
5. Select View button for a question in the table. A side panel opens.
6. Select the Mark as verified button. The button changes to Verified.
7. To unverify a response and remove it as a suggested question, select the Verified button, then select Unmark as verified.

Tips to Improve the Performance of a Topic

Turn off any unused data fields

Limiting the number of activated data fields helps ensure the quality of AI answers. 

This can be done in the Data tab, in the Data Fields subtab. Set the Include toggle for data fields to Off.

Add synonyms for data fields

Synonym ensure the AI understands your organization’s specific language in natural language prompts. 

This can be done in the Data tab, in the Data Fields subtab. Select the Edit or Add button in the Synonyms column for a data field.

Add friendly names and descriptions to datasets

Give a dataset a friendly name and description to provide information about the data in it. 

This can be done in the Data tab, in the Datasets subtab. Select the Edit icon in the Friendly Name column to update the friendly name. Select the Expand icon to the right to update the Description.

Add friendly names and descriptions to data fields

Give data fields friendly names and descriptions to help the AI understand the terms and phrases users may use to refer to them in Ask Insights. 

This can be done in the Data tab, in the Data Fields subtab. Select the Edit icon in the Friendly Name column to update the friendly name. Select the Expand icon to the right to update the Description.

Add a description to the topic

The topic description can provide information about the topic's overall purpose. 

This can be done in the Summary tab, in the Topic details section. Select the Edit icon next to the Description.

Add custom instructions to the topic

Custom instructions add more context to the topic. 

This can be done in the Custom Instructions tab. Enter custom instructions, then select Save changes.

For best practices for writing custom instructions, see this external article.