Amazon Q in QuickSight has enhanced how organizations extract meaningful insights through natural language querying. However, its true potential lies in strategic topic curation, much like a well-organized library. Proper topic curation is crucial for enhancing user experience and customer satisfaction, which are paramount for business success.
In this post, we discuss Amazon Q in QuickSight topic curation strategies that can help you enhance user experience, improve answer accuracy, and drive better business decisions.
Amazon Q in QuickSight is a machine learning-powered natural language querying capability that enables users to ask business questions in plain English and receive instant visual answers. It eliminates the need for complex SQL knowledge by understanding everyday business terminology and automatically generating appropriate visualizations, tables, or narratives.
Amazon Q in QuickSight topics are the foundation that bridges business questions with accurate data insights. These carefully curated data collections define how Amazon Q interprets and responds to natural language queries. When properly maintained and enhanced, topics drive superior answer quality, boost user confidence, and enable faster insights.
Prerequisites
- To create Amazon Q topics in QuickSight, you need an active AWS account with QuickSight Enterprise Edition subscription.
- The user must have QuickSight Author or Admin role along with necessary IAM permissions for Amazon Q integration. Refer to Using Amazon QuickSight with IAM for more details.
- Additionally, the account should have at least one configured data source and published dataset, and you need to be an Author Pro to edit topics. For pricing, refer to Amazon QuickSight pricing.
Create an Amazon Q in QuickSight topic
Complete the following steps to create an Amazon Q in QuickSight topic:
- On the QuickSight dashboard, choose Topics in the navigation pane.
- Choose New topic.
- For Topic name, enter a name of the topic. (For our example, we name our topic Retail Q Topic).
- For Description, enter a clear, concise topic description that guides users to the right data. An effective description highlights key metrics, business context, and available insights, helping users quickly determine if this topic answers their business questions.
- Choose Continue.
- Select the dataset you want to be part of the topic. (For our example, we use the dataset
Retail_Daily.) - Choose Create.
The dataset will be analyzed as Amazon Q indexes the data and sets up field configurations as part of topic creation.
After creation, you can find your new topic listed in the Amazon Q Topics section.
Q&A topics can be created directly from an analysis, automatically incorporating all relevant datasets.
Exploring Amazon Q topics
In this section, we explore the different tabs on the Amazon Q topic details page.
After you create a topic, or when you choose an existing topic from the list on the Topics page, the topic opens to that topic’s workspace. Four tabs appear here that you can use as described in the following sections.
Summary tab
The Summary tab contains the following sections:
- Improve your topic – This section allows you to review your topic settings to optimize performance. Chose Start review to begin evaluating your topic settings.
- Topic details – This includes the following information:
- Name of the topic
- Description
- Last modified date and time
- Last refresh date and time
- Statistics – This section contains statistics on the topic usage and feedback. Detailed usage metrics are not available for newly created topics. After a topic is published and users begin accessing it, this section will display comprehensive insights into usage patterns and user sentiment.
Data tab
The Data tab contains the following sections:
- Datasets – Shows you the list of datasets associated with this topic.
- Data fields – Provides comprehensive details about all the data fields along with synonyms. Amazon Q automatically includes the most relevant columns and adds field name synonyms in this initial version of the topic.
- Named entity – Named entities are groupings of data fields that collectively represent a business concept and are used to enhance the Q&A experience. This section lists the created named entities.
User Activity tab
On the User Activity tab, you can track user engagement through comprehensive activity metrics, including total question volume, answer success rates, and user feedback analysis. This dashboard provides valuable insights into both answerable and unanswerable queries, along with detailed positive and negative feedback statistics.
This information can help you analyze negative feedback patterns to identify areas for improvement and take targeted action. These insights enable you to enhance user experience through strategic refinements and optimizations.
Suggested Questions
The Suggested Questions tab categorizes questions into two key types:
- Verified questions – Provides a list of verified questions that are thoroughly reviewed. Giving users this level of transparency and validation can instill greater confidence.
- AI-generated questions – Provides a list of AI-generated questions based on your enabled data fields in the topic.
Enhancing Amazon Q in QuickSight topics
Enhancing Amazon Q in QuickSight topics is crucial for delivering an optimized data analytics experience, because it directly impacts how users interact with and extract insights from their data. By fine-tuning topics through the addition of relevant synonyms, named entities, and verified questions, organizations can significantly improve natural language query accuracy and reduce user frustration. This optimization not only leads to higher user adoption rates and satisfaction, but also empowers business users to make data-driven decisions more efficiently, ultimately resulting in better business outcomes and a stronger return on investment in analytics infrastructure.
In this section, we discuss various ways in which you can enhance your topics.
Include or exclude fields
Amazon Q in QuickSight curates a selection of high-impact fields to optimize performance and relevance. Amazon Q initially selects a focused subset of fields to keep the scope narrow. Fields should only be included if the user truly needs them for their questions.
As the domain expert of your data and business requirements, you maintain full control to customize field selection, enabling or disabling specific fields to align with your analytical needs.
You can manage field visibility by selecting your desired fields, allowing you to precisely control which fields are included or excluded from your analysis.
Toggle the button on to include the field, or off to exclude it.
Add field synonyms
Field synonyms in Amazon Q in QuickSight significantly enhance natural language query accuracy by recognizing alternative terms users might use when asking questions about their data. Field synonyms are meant to capture business synonyms that are specific to your domain that Amazon Q wouldn’t know.
For example, if your data field is labeled Ratings but some business users might use the term “product rating” or “product score,” adding Product Rating and Product Score as synonyms makes sure Amazon Q understands and correctly interprets these variations.
You can enhance field recognition by adding synonyms by choosing either Add alternative terms for field or the plus sign.
Add value synonyms
Adding value synonyms in Amazon Q in QuickSight directly enhances user and customer experience by making data exploration more intuitive and user-friendly. When users can access data using familiar terms rather than exact database values, it creates a more natural and efficient analytical experience.
Consider a scenario where your dataset contains Australia in the country column. Without value synonyms, when a user queries “Total Amount for Country Aus,” the system fails to recognize “Aus” as equivalent to “Australia,” resulting in the following message.
Complete the following steps to add value synonyms for the Country column:
- Choose the field (Country) to expand its properties.
- Choose Configure field value synonyms.
- Choose Add.
- Enter the value (for example, Australia).
- Add synonyms (for example, Aus and AUS).
- Choose Done to confirm the changes.
This configuration makes sure that queries using common abbreviations will successfully retrieve the intended data.
After you implement value synonyms, queries such as “Total Amount for Country Aus” will successfully retrieve data for Australia, delivering accurate insights and providing a seamless user experience.
Add default aggregation
Default aggregation in Amazon Q in QuickSight topics significantly enhances the user experience by automatically applying the most appropriate mathematical operations to numeric fields based on business context. This feature provides consistent and meaningful data representation across queries.
This intelligent pre-configuration empowers users to focus on extracting valuable insights rather than managing calculation methods, delivering more efficient and reliable data analysis outcomes.
If the business needs Count distinct as the default aggregation, you can set it using the Default aggregation option. When users need to analyze unique records through queries like “Show Products by Month,” the system will return unique number of products per month.
Complete the following steps to configure Count distinct as your default aggregation method:
- Find the field requiring unique count aggregation. (In our example, we use the field
products.) - Choose the field to expand field properties.
- For Default aggregation, choose Count distinct.
- Choose Save.
With Count distinct now set as the default aggregation, queries such as “Show Products by Month” will automatically display unique product counts, delivering precise insights for accurate decision-making.
Exclude aggregations
Configuring Not allowed aggregation in Amazon Q in QuickSight topics enhances analytical accuracy by preventing unintended calculations, making sure data insights align with business requirements and analytical best practices.
Consider a business scenario where unique product counts need to be restricted for the product field. Configuring count distinct as a disallowed aggregation protects your analytics integrity by preventing unexpected unique counts.
Complete the following steps to configure Count distinct as Not allowed aggregation:
- Find the field requiring unique count aggregation. (In our example, we use the field
products.) - Choose the field to expand field properties.
- For Not allowed aggregation, choose Count distinct.
- Choose Save.
Add semantic info
Adding semantic information in Amazon Q in QuickSight topics enhances natural language understanding by providing contextual meaning to your data fields and relationships. By defining semantic relationships between fields, units of measurement, and business contexts, the system better interprets user queries and delivers more accurate, relevant responses. This helps Amazon Q answer questions like “who, when, and where.”
To better handle people-related queries, we demonstrate setting the semantic type of the Name field to Person. Complete the following steps:
- Find the field requiring unique count aggregation. (In our example, we use the field
Name.) - Choose the field to expand field properties.
- For Semantic type, choose Person.
- Choose Save.
Add a calculated field
Adding calculated fields in Amazon Q in QuickSight topics enhances analytical capabilities by creating custom metrics that align with specific business requirements. This feature allows you to define complex calculations that automatically execute when users ask related questions.
Consider a business scenario where users need to analyze true product costs. By creating a calculated field that determines true cost using the Sales and Profit columns, users can query this metric without performing manual calculations. This enables accurate cost analysis through simple natural language questions.
Complete the following steps to create a calculated field:
- On the Data tab, choose Data fields.
- Choose Add calculated field.
- Provide a descriptive name for the field. (In our example, we name it
TrueCost.) - Enter the calculation formula. (In our example, we enter (Sales)- (Profit).)
- Choose Save.
After you create the calculated field, you can immediately incorporate it into your natural language queries. For example, when asking “Show me TrueCost by Product,” the system automatically applies the calculation and displays precise cost metrics for each product.
Add a filter
Adding filters in Amazon Q in QuickSight topics enhances the user experience by enabling more precise and relevant data exploration. When properly configured, filters help users narrow down information based on specific criteria, making sure the user receives exactly the insights they need. This streamlined approach not only improves answer accuracy and reduces query refinement time, but also enables users to make better-informed business decisions through more focused and contextual data analysis.
Business users often ask questions using terms that correspond to multiple data values. For example, if the business users frequently inquire about combined sales data for East Coast states in the US, they might refer to this group as EastCoast. To accurately answer such queries, the data needs to be filtered specifically for the states under EastCoast.
Complete the following steps to configure the appropriate filters:
- On the Data tab, choose Data fields.
- Choose Add filter.
- For Name, enter a name for the filter. (In our example, we name it
EastCoast). - For Field, choose the field to which you want to apply the filter. (In our example, we apply it to State).
- For Filter type, choose the filter type. (In our example, we choose Custom filter list).
- For Rule, choose Include as the rule to specify the values. (In our example, we use North Carolina, New Jersey, and New York).
- Choose Save.
After the EastCoast filter is configured, users can query Amazon Q for “Total Sales for EastCoast” and receive accurate data for the states you added.
Refer to Adding filters to an Amazon QuickSight Q topic dataset for more details.
Add a named entity
Named entities in Amazon Q in QuickSight topics are strategic groupings of related data fields that represent meaningful business concepts, enhancing the natural language query experience and data comprehension.
Named entities excel in diverse business scenarios by transforming complex data relationships into intuitive business concepts. By logically grouping related fields, they enable users to ask natural questions and receive comprehensive insights.
Consider a business scenario where multiple product-related fields exist, such as Product Type, Product Brand, and Product Category. To streamline product analysis, we create a named entity called Product_Details by grouping these related fields. When users include Product_Details in their questions, they automatically receive comprehensive product information across all relevant fields. This intelligent field grouping transforms complex data relationships into intuitive insights, enhancing the analytical experience.
Complete the following steps to create a named entity:
- On the Data tab, choose Named entity.
- Choose Add named entity.
- For Entity name, enter a name for your entity. (For our example, we enter
Product_Details.) - For Description, enter a description.
- Add the following fields: Product Type, Product Brand, and Product Category.
- Optionally, add relevant synonyms.
- Choose Save.
With named entities now configured, users can efficiently retrieve consolidated product information. When querying “Show me Product_details for EastCoast,” the system intelligently presents data from all associated fields, demonstrating how named entities transform complex data relationships into accessible insights.
Add multiple datasets
Adding multiple datasets to an Amazon Q in QuickSight topic enhances the analytical experience by creating a unified data environment that enables comprehensive business insights.
For cross-dataset joins, you can perform these operations upstream during data preparation to create a single dataset.
Complete the following steps to add multiple datasets to a topic:
- On the Data tab, choose Datasets.
- Choose Add datasets.
- Select the dataset you want to add to the topic. (For our example, we add the dataset
Retail_Monthly.) - Choose Add Datasets.
You can now see your dataset on the Datasets tab.
Add a custom message
Adding custom messages in Amazon Q in QuickSight topics enhances the user experience by providing tailored guidance and contextual information when specific scenarios occur.
Complete the following steps to add a custom message:
- On the Suggested Questions tab, choose Verified.
- Choose Add custom message.
- Choose the plus sign to add the keywords. (For our example, we add
switzerlandas the keyword.) - For Description, enter the custom message that users will see when they ask questions that include the keywords you’ve defined for this topic.
- Choose Save.
When users ask questions containing the defined keywords, the system will display the message you specified.
Hide AI-suggested questions
In addition to verified questions, Amazon Q enhances the user experience by displaying AI-generated questions. However, you might want to hide the suggested questions.
To hide the AI-generated questions, complete the following steps:
- On the Suggested Questions tab, choose AI-generated.
- On the Options dropdown menu, choose Hide from suggestions.
Amazon Q will no longer display AI-generated questions.
Link Amazon Q topics to analyses and dashboards for discoverability suggestion
By linking Amazon Q topics to analyses and dashboards, QuickSight creates a powerful discovery engine that anticipates users’ analytical needs. The system uses the context of current visualizations and user interaction patterns to suggest relevant questions, making data exploration more intuitive and guided.
Complete the following steps to link an Amazon Q topic to an analysis:
- On the QuickSight console, navigate to the analyses where you want to link an Amazon Q topic.
- Choose the options menu (three dots) next to Build visual and choose Topic linking.
- Enable Link topic for Build Visual and Q&A.
- Choose the topic to be linked.
- Choose Apply changes.
Complete the following to link an Amazon Q topic to a dashboard:
- Choose Publish to publish your dashboard.
- For Publish new dashboard as, enter the dashboard name.
- Choose Link topic.
- Enable Link topic for Build Visual and Q&A.
- Choose the topic to be linked.
- Choose Apply changes.
- Choose Publish dashboard.
Conclusion
This post explored strategic approaches to enhance Amazon Q in QuickSight topics, demonstrating how effective topic curation can transform your organization’s data analytics experience. Through careful topic organization, strategic keyword definition, and customized responses, businesses can empower users to access precise insights using natural language queries. The built-in user activity tracking capabilities provide administrators with valuable metrics to continuously optimize content based on actual usage patterns and feedback. This systematic approach leads to improved answer accuracy and more efficient data-driven decision-making across your organization.
To learn more about Amazon Q in QuickSight, see Amazon Q in QuickSight. Additionally, check out the following resources:
- Getting started with Amazon QuickSight Q
- Answering business questions with Amazon QuickSight Q
- Using Generative BI with Amazon Q in QuickSight
About the author
Satish Bhonsle is a Senior Technical Account Manager at AWS. He is passionate about customer success and technology. He loves working backwards by quickly understanding strategic customer objectives, aligning them to software capabilities and effectively driving customer success.
This is a companion discussion topic for the original entry at https://aws.amazon.com/blogs/business-intelligence/curate-topics-in-amazon-q-in-quicksight-to-derive-maximum-value-and-improve-user-experience/