AWS Business Intelligence Blog
Enhance your Amazon Redshift business intelligence workloads with gen AI capabilities using Amazon Q in QuickSight
Amazon QuickSight is a unified business intelligence (BI) service that helps users meet their various analytics needs through building modern interactive dashboards, performing one-time analysis and getting business insights from their data quickly and efficiently with less friction. Amazon QuickSight has been designed to solve complex and costly infrastructure problems by bringing the scale and flexibility of the Amazon Web Services (AWS) Cloud to business intelligence.
Amazon Q in QuickSight is an innovative feature within Amazon QuickSight that lets users interact with their data through natural language queries. By typing a question in plain language, such as “What were our top five products by sales last year?” Amazon Q in QuickSight instantly interprets the query and provides relevant visualizations and insights. Powered by natural language processing (NLP) and machine learning (ML), Amazon Q in QuickSight removes the complexity of data exploration so nontechnical users can engage with their data without needing to write SQL queries or understand the technical details. This democratizes data access, making it easier for everyone in the organization to uncover insights.
Amazon Redshift is a fast, scalable, and fully managed cloud data warehouse you can use to process and run your complex SQL analytics workloads on structured and semi structured data. It also helps you securely access your data in operational databases, data lakes, or third-party datasets with minimal movement or copying of data. Tens of thousands of customers use Amazon Redshift to process large amounts of data, modernize their data analytics workloads, and provide insights for their business users.
Combining the powerful data processing capabilities of Amazon Redshift with the intuitive AI-driven interface of Amazon Q in QuickSight offers a game-changing solution for modern businesses. Together, they enable:
- Faster, smarter decisions – By using the real-time analytics of Amazon Redshift and the NLP of Amazon Q in QuickSight, companies can quickly derive actionable insights, even from massive datasets. This shortens decision cycles and allows teams to react more swiftly to emerging trends.
- Accessibility for all – With Amazon Q in QuickSight, everyone in the organization, from data analysts to business users, can quickly access and interpret complex data without requiring technical expertise. This fosters a culture of data-driven decision-making across all levels.
- Enhanced business insights with generative AI – By embedding generative AI capabilities into BI workflows, companies can automate the discovery of trends, detect anomalies, and forecast outcomes based on historical data, making predictive analytics more accessible and actionable. This also reduces the learning curve for authors on how to build dashboards, reports and metrics.
In this post, we explore how you can use Amazon Redshift and Amazon Q in QuickSight to enhance your BI workloads with generative AI, unlocking new capabilities in data-driven decision-making.
Solution overview
The solution architecture is made up of four components, as illustrated in the following diagram:
- Amazon Simple Storage Service (Amazon S3) – Contains raw data files
- Amazon Redshift – Ingests raw data and applies business logic through data transformation, data marts, or views
- Amazon QuickSight – Generates dashboards from Amazon Redshift datasets
- Amazon Q – Generate dashboards and reports using generative AI
Prerequisites
To complete this solution, you need to have the following prerequisites in place:
- An Amazon Redshift provisioned cluster or serverless endpoint. To create one, refer to Creating a cluster or Creating a data warehouse with Amazon Redshift Serverless. Alternatively, you can launch the following stack to launch AWS CloudFormation in your AWS account.
- An Amazon QuickSight (enterprise account) administrator or author with a connection to Amazon Redshift.
- Make sure that your Amazon QuickSight user has a Pro role (Admin Pro or Author Pro for this post).
Data ingestion into Amazon Redshift
In this post, we use the TPC Benchmark H Standard Specification (TPC-H) data model. The following schema explains the relationships between the different tables.
There are four available datasets of different sizes (10 GB, 100 GB, 3 TB, and 30 TB), and you can load any of them into Amazon Redshift. In our test, we ingest into Amazon Redshift the 10 GB dataset by executing the CREATE TABLE and COPY commands that are defined in CloudDataWarehouseBenchmark-TPCH-10GB. These commands can be executed using Amazon Redshift Query Editor V2 or other SQL clients with a Java Database Connectivity (JDBC) or Open Database Connectivity (ODBC) connection (using Amazon Redshift drivers).
Make sure that your Amazon Redshift cluster or endpoint has an AWS Identity and Access Management (IAM) role that has read access (GetObject and ListBucket) on the S3 bucket.
After the data is ingested into Amazon Redshift, create the following view to use as a dataset in Amazon QuickSight.
Amazon QuickSight configuration
If your Amazon Redshift data warehouse isn’t already configured as a data source in Amazon QuickSight, then, before creating a dataset, you need to configure the connection between both services and enable access to your data warehouse.
Establishing the connection between both services depends on whether your Amazon Redshift is accessible through a public network:
- If Amazon Redshift data warehouse is in a public subnet and configured as publicly accessible, you can add the Amazon QuickSight IP address range to the inbound rules of your Amazon Redshift security group.
- If it isn’t publicly accessible, make sure that the virtual private cloud (VPC) and security group(s) are configured according to the documentation for Amazon QuickSight to have access to Amazon Redshift.
When Amazon Redshift data warehouse is accessible from Amazon QuickSight, you can proceed with creating a new dataset. Follow these steps:
- If this is the first time you connect to Amazon Redshift, you can use Auto-discovered or Manual connect to define it as a data source, as shown in the following screenshot. Select your existing data source if you’ve already connected to Amazon Redshift.
- Choose the Amazon Redshift schema where you created the view and then select the view orders_agg. You also have the option to edit the data before creating the dataset using the Edit/Preview data button, as shown in the following screenshot.
- Choose Import to SPICE or Directly query your data. For this post, we choose Directly query your data.
- Visualize the dataset takes you to the analysis page where you can create and define charts.
Using Amazon Q
Amazon Q in QuickSight accelerates your BI workloads by introducing the following generative BI capabilities and functionalities:
- Building dashboards by using natural language prompts to create calculated fields and build and refine visuals
- Creating topics that will provide a Q&A experience where answers are enhanced by multiple visuals based on related data
- Creating data stories by explaining your data through visuals and providing ideas to help improve your business and getting executive summaries of key insights from a dashboard
Scenarios with Amazon Q in QuickSight is a new capability and feature that was recently released in public preview. It allows Amazon QuickSight users to analyze complex business problems with simple natural language and help them perform tasks that simplify in-depth analysis and save hours of manual data manipulation.
In this post, we cover visuals, topics, and data stories.
Build a visual
To build a visual, follow these steps:
- On the Analyses page, open the analysis that uses the dataset that was created from the Amazon Redshift view orders_agg.
- Under Visuals, choose +ADD to manually add and define your chart. Choose BUILD to automatically build a chart using natural language and add it to your analysis. For this demo, choose BUILD, as shown in the following screenshot.
- The Build a visual popup window appears at the right side of your page. Start describing what you want to visualize and build. In the following example, we entered a query to visualize the top five countries with the highest revenue.
- (Optional) To change the chart type, view insights, or export to .csv before adding the chart to your analysis, use the icons at the top right of the chart.
- To add the generated chart to your analysis, choose ADD TO ANALYSIS.
- (Optional) Change the y axis, x axis, or other options such as the aggregation function or sorting order.
- To create and add calculated fields using natural language, choose + CALCULATED FIELD.
Amazon QuickSight authors can also use generative BI to refine and edit their visuals using natural language prompts.
Create a topic
To create a topic, follow these steps:
- On the QuickSight start page, choose Topics.
- Choose New Topic.
- Enter a topic name and a brief description and turn on Use new generative Q&A experience.
- Select a dataset (in this post, we use the dataset created from the Amazon Redshift view orders_agg).
After creating a new topic, a topic’s workspace page with the following four tabs opens:
- Summary – Contains statistics and suggestions to help improve your topic.
- Data – Shows the fields of the chosen dataset and the ability to configure the topic metadata.
- User Activity – Displays key performance indicators (KPIs) on the questions asked.
- Suggested Questions – Lists the verified and AI-generated questions.
In the Data tab, you can add new datasets and edit metadata for a dataset using the following subtabs:
- DATASETS – To add new dataset(s)
- DATA FIELDS – To edit metadata and add calculated fields and filters
- NAMED ENTITY – To enhance the Q&A experience
Named entities are considered one of the main components of topic curation and authoring Q&A. The information and field ranking defined in named entities will play an essential role in providing contextual and multi-visual answers in response to vague questions (for example, when users refer to multiple columns of data without stating each column explicitly).
Choose OPEN Q&A and start asking questions. In the following example, we asked for a pie chart showing the total number of orders per region in 1998 and marked it as verified. The generated output usually consists of a summary of visuals that highlights key insights, a center visual that directly answers the question and other visuals on the right that provides relevant KPIs and context.
You can save the chart by adding it to the pinboard so it can be used for further analysis or data stories. Here are other examples:
- Total revenue by market segment in EUROPE in 1998
- Top five countries with highest revenue in EUROPE in 1998 for the segment AUTOMOBILE
Create a data story
To create a data story, follow these steps:
- On the QuickSight start page, choose Data Stories
- Choose New Data Story
- Describe the story that you want to generate
- To add visuals from your pinboard, choose + ADD VISUALS
- Choose BUILD
You can edit and share a generated data story on the Amazon QuickSight console. The following screenshots show examples of the output. A data story explains your data with visuals, insights, and ideas that help enhance your business.
A generated data story draft isn’t meant to replace your ideas, analysis, or interpretation. It’s usually considered as a starting point that you can customize and edit based on your requirements and circumstances.
Conclusion
In this post, we’ve shown that getting started with Amazon Redshift and Amazon Q in QuickSight is straightforward. We’ve shown how you can transform raw data into business-critical insights with minimal setup.
Using Amazon Redshift for powerful data processing and combining it with Amazon Q in QuickSight to integrate generative AI capabilities can transform how businesses handle their analytics workloads. This approach not only simplifies access to critical insights, but also democratizes data analysis across the organization. From building dashboards with natural language prompts to creating topics that enable enhanced Q&A experiences and generating automated data stories, this solution allows companies to unlock actionable insights quickly and efficiently.
By introducing generative AI–driven BI workflows, businesses can automate complex analyses, detect trends and anomalies, and generate forecasts. Ultimately, this combination empowers both technical and nontechnical users to make smarter, data-driven decisions, driving innovation and improving competitiveness in a rapidly evolving market.
Now it’s time to take your BI workloads to the next level by embracing the future of analytics with generative AI and Amazon Q.
About the Authors
Ziad Wali is an Acceleration Lab Solutions Architect at Amazon Web Services. He has over 10 years of experience in databases and data warehousing, where he enjoys building reliable, scalable, and efficient solutions. Outside of work, he enjoys sports and spending time in nature.
Semir Naffati is a Sr. Specialist Solutions Architect in Analytics at AWS, helping customers across the EMEA region design, implement, and refine cloud infrastructure, databases, DevOps, and business intelligence solutions. With over 20 years of experience in IT, particularly in the banking and finance sectors, he is dedicated to supporting organizations in adopting scalable, cloud-native solutions. Semir advises C-level executives on cloud strategy and digital transformation to make sure solutions are aligned with long-term business objectives and innovation.