⚠️ Check Your Version.
The steps in this article are for Dashboards & Analytics (Amazon Q) users only. If you use Reports (Microstrategy), proceed to these articles instead.
This article covers the following:
You must have the Author Pro license to perform the actions described in this article.
- Topics in Dashboards & Analytics
- Creating and Configuring a Topic
- Reviewing and Verifying Suggested Questions
- Tips for Reviewing Inaccurate Suggested Questions
- Enabling Q&A for Dashboards
- Building Visuals Using Natural Language
- Generating Executive Summaries
- Creating Data Stories
Topics in Dashboards & Analytics
Topics teach the AI how your organization refers to its data. They connect business language, friendly names, and synonyms to the actual dataset fields used to answer questions.
When you configure a Topic to add friendly names and synonyms, you tell the AI, “When users say this, they mean that.” For example, if your users say “firms”, you can link that term to the LawFirmName field. And if they may ask about “total spend,” you can link that to the ApprovedMatterSpend field.
See this video for more on Topics, including creating them.
Why topics matter
- They improve natural language interpretation.
- They let you expose only relevant fields for Q&A.
- They support curated, verified questions for end users.
Without a configured topic, the system may understand the dataset structure but not the terms your organization uses every day.
Best Practices
- Include all relevant fields in the topic.
- Replace technical field names with clear friendly names.
- Add synonyms for common business language.
- Verify suggested questions before exposing them to users.
- Refresh or update the topic if the dataset changes.
Creating and Configuring a Topic
Creating a Topic
- Go to Topics.
-
Click New Topic. The New Topic window displays.
- Enter a unique, descriptive topic name.
- Enter a description if needed.
- Click Continue.
- Select the dataset.
- Click Create.
Configuring Topic Fields
-
Click the topic name.
- Optional: Click Start Review for a guided workflow on completing the topic configuration. Use the diagonal arrows to minimize the panel as you work through each step. Stop the guided review at any step and return to complete the remaining steps later.
-
Select the Data tab.
- Toggle Include on for fields that should be available for Q&A.
-
Review the Friendly name for each field. Click the pencil icon to change the name to one that users will recognize when using Q&A.
- Add Synonyms for common business terms. Add synonyms for terms your users commonly use, even if they do not match the technical field name.
Reviewing and Verifying Suggested Questions
Suggested questions help users discover useful ways to query the data. Verifying them ensures the answers are relevant and accurate.
- Open the topic.
-
Select the Suggested Questions tab.
- Select the AI Generated tab.
-
Review the AI-generated suggestions.
- Click View to test a question.
- If the result is accurate, click Mark as Verified.
- If the result is inaccurate, use the edit controls to adjust the interpretation and retest. See this section for more information.
- Remove irrelevant questions or regenerate the list when needed.
Adding Additional Questions
-
Click Open Q&A.
- Enter a question in the search box.
- Click Ask.
- Review the answer. Click the pencil icon to make edits if needed.
- Click Mark as Verified to include it in the list.
Tips for Reviewing Inaccurate Suggested Questions
If your Suggested Questions are inaccurate, consider what changes could improve them.
- Does the topic include all relevant fields? If not, relevant dataset fields within the topic can lead to incorrect interpretations and incomplete answers.
- Are synonyms added for commonly used business terms? Add variations for how users naturally refer to data.
- Are friendly field names clear and descriptive? Avoid abbreviations or technical names.
- Does the question use vague or ambiguous language? Rephrase the question using specific fields or metrics.
- Does the topic need to be refreshed to ensure any dataset updates have been captured? From the topics list, click the Actions menu icon and select Update. Then configure the new fields to include them in your topic.
- Are the suggested questions still relevant? You can regenerate suggestions if fields were added or renamed in the dataset. Open the topic and select the Suggested Questions tab. On the AI Generated sub-tab, select Option and click Regenerate. Review and verify the new questions.
Enabling Q&A for Dashboards
To make dashboard Q&A available to end users, you must link a topic to the analysis and enable Q&A during publishing.
- Open the analysis.
- Click the Amazon Q icon.
-
Select Set up Q&A.
The Manage Q&A window displays.
- Turn on the option to link a topic for Q&A.
- Select the topic that matches the dataset.
- Apply the changes.
- Click Publish.
- Confirm Allow data Q&A is selected.
- Move the dashboard from your personal folder to the Shared Folder.
- Open the dashboard.
- Click Share.
- Grant dashboard permission to the appropriate users. See this topic for more information.
Building Visuals Using Natural Language
Author Pro users can generate visuals by describing the desired output in natural language.
See this video on more for building visuals with AI.
- Open an analysis.
-
Click Build visual with Q.
- In the slide out panel, enter a prompt, such as
Show law firm total approved spend for 2025 as a bar chart. Or, choose a prompt from the list of suggested questions. - Click Build.
- Review the visual.
- If needed, revise the prompt or choose an alternative interpretation.
- Click Add to Analysis when the visual is ready.
Refining an Existing Visual
- Hover over the visual.
-
Select Edit with Q.
The Edit visual window displays.
- Enter a description of the change.
- Click Apply.
Generating Executive Summaries
Executive summaries provide a high-level written overview of the trends and findings shown across a dashboard’s visuals.
- Navigate to an analysis you've created.
-
Click Publish and confirm the dashboard was published with the Allow executive summaries option enabled.
- If not, enable that option and click Publish Dashboard.
- If previously enabled, close the Publish dashboard window.
- Open the dashboard.
-
Click Summarize.
-
Review the summary panel.
- Optional: Click Copy to Clipboard to copy the information.
Creating Data Stories
Stories turn selected visuals into an AI-generated narrative report. You can then edit the text, layout, and content blocks to produce a presentation-ready output.
See this video for more on Data Stories.
Building a Story
- Click Stories from the left navigation panel.
- Click New Story.
-
In the Build story panel, cescribe the story you want to generate.
- Click +Add to add appropriate visuals to the story.
- In the Add visuals window, search or navigate to select a dashboard that contains the relevant visuals.
-
Place a check mark in the top-right corner of visuals on that dashboard that you want to add to the story.
You can select up to 20 visuals from various dashboards.
- Click Build.
Editing a Story
You can choose different visuals and build the story again if needed. Or, you can take the draft story and refine it. Hover over a block to make changes. Editing options include:
-
Text editing – Select any text to change the narrative. Highlight a word to display a formatting toolbar.
- Visual editing – Select any visual to display the toolbar. You can edit the properties, refresh the data, remove the visual from the block, or expand it to fit the full width.
- Add content block – Hover between two sections of the story to display and click the + icon. Choose a layout and a block is added to that spot. Follow the block instructions to add text or type / to insert a visual or image.
- Reorder blocks – Click the vertical ellipse and choose Move block up or Move block down to reposition it in the story.
- Duplicate block - Click the vertical ellipse and choose Duplicate to create a copy of the selected block to edit.
- Delete block - Click the vertical ellipse and choose Delete to remove a block from the story.