Now more than ever, having the capacity to quickly access and evaluate data is essential for maintaining a competitive advantage. With reliable data gathering and robust infrastructure being the standard across industries, business leaders must take the next step: maximizing the valuable insights gleaned from this abundance of data. Business intelligence (BI) tools are typically the access point for stakeholders to get their reports and dashboards; however building out and maintaining a robust reporting ecosystem can require significant time and effort from analysts, engineers and BI teams.

Decision-makers have constantly evolving needs but the data required to make informed choices isn’t always readily accessible. Compounding the issue is that building ad-hoc reports demands considerable time and involves numerous back-and-forth with the BI team. This presents significant bottlenecks in time-to-insight. But what if business users could directly access and analyze their data without relying on technical expertise?

What Amazon Q in QuickSight Is

Amazon Q in QuickSight is a generative AI assistant embedded within QuickSight, the Amazon Web Services browser-based cloud Business Intelligence tool. The tool allows users to ask natural language inquiries about data and receive prompt, accurate answers accompanied by instant visualizations. Developers can also leverage natural language questions to create visuals, executive summaries, forecasts and anomaly reports. 

Amazon Q in QuickSight has the potential to significantly reduce time to insight and lighten the load on dashboarding teams. While BI analysts focus on building data sources and critical reports, business users can generate ad-hoc reports in seconds rather than hours. QuickSight operates securely within your Amazon QuickSight account, requiring only clean data sources, well-structured metadata, and well-formulated questions.

What it Is Not

Amazon Q in QuickSight is not a replacement for QuickSight or any other BI tool, nor is it a general-purpose AI virtual assistant. Unlike general-purpose AI assistants, such as Amazon Q Business, which use data from across the organization to answer broad business questions, Amazon Q in QuickSight is designed for narrow, data-specific inquiries. It works best with business-specific questions related to datasets that have been prepared for natural language queries, providing focused, data-driven answers tailored to the user’s needs

Leveraging Amazon Q in QuickSight

For Chief Data Officers

Empower your teams to explore data independently with Amazon Q in QuickSight. This tool allows everyone in the organization to ask questions and gain insights without relying on data analysts for every query. By democratizing data access, you can foster a culture where business decisions are quicker and more informed. This increased autonomy also frees up your data teams to focus on complex analyses and strategic initiatives, driving greater value across the organization.

For Business Intelligence Managers

Streamline your team’s workload by routing frequently asked data questions through Amazon Q in QuickSight. This approach significantly reduces the time spent on repetitive data requests and ad-hoc reports, allowing your team to focus on higher-value tasks. For example, you could create a Q Topic dedicated to quality assurance on dimensional data or for validating database metrics without writing SQL. By implementing Amazon Q in QuickSight, you not only boost your team’s productivity but enhance the efficiency of teams you work with as well.

For Analysts

Use Amazon Q in QuickSight to validate hypotheses and gather quick insights during exploratory analysis. Save time on running queries by asking data questions right in the Amazon Q in QuickSight interface. The tool also facilitates quickly switching between datasets, as all topics are available and can be accessed in any part of QuickSight. Amazon Q in QuickSight also supports building dashboard visuals and writing calculations directly in the Q interface, further accelerating report development.

Creating a Topic

Using Amazon Q in QuickSight is simple by design. Once enabled, the tool appears as an icon within QuickSight, which the user can click, ask a question about data and receive an answer.

To successfully implement Amazon Q in QuickSight, data topics must first be set up, which requires thoughtful planning.

Start Asking the Right Questions

Before even choosing the right dataset to answer questions, it’s important to establish a clear business case and determine what questions stakeholders need answered. This process is iterative and adaptable.

Start by engaging with your key stakeholders to understand their pain points and goals. What insights are driving business decisions? Which reports are they using frequently? Which are they rarely using? How much time is being spent on ad-hoc reporting? The insights gained from this process is invaluable, and ensures the stage is set for Amazon Q in QuickSight implementation.

Building Topics

Topics for Amazon Q in QuickSight are organized sets of data designed to give context for the AI, allowing users to ask questions and receive answers in seconds. When an end user interacts with Q, they will do so by selecting a topic within the QuickSight interface and asking questions related to that topic.

To create a topic, you must first select the datasets the topic will use, and manipulate the metadata to fit the chosen business case to make it user-friendly.  

Identify Jargon

During this process, note down any business-specific language used by stakeholders. When creating a topic, you can give data fields synonyms. This helps Amazon Q in QuickSight understand questions and correctly identify the relevant field to use in its analysis. If a stakeholder asks a question using business jargon known to them, they should still get an answer they expect. Organizations that maintain a glossary/dictionary of terms for their business can use this document to further speed up adoption of Amazon Q in QuickSight by their stakeholders.

Topics: Broad or Narrow?

Should your topic be about sales across the organization, or only one product? Will it contain sales data, inventory data, or both? These are critical decisions to make when setting up your topics in Amazon Q in QuickSight.

Topics encompassing a broad range of target data can answer diverse questions and provide a comprehensive view of the business. They are useful for high-level reporting and for users who require a general overview.

Topics that are narrower in focus might include sales data, customer accounts, key financials or other limited scope data. These focused topics offer more targeted insights, making it easier for end users to get relevant answers from Amazon Q in QuickSight quickly.

Your exact topic setup can vary widely pending the business need defined at the outset, data available and audience for the topic. One helpful feature offered by QuickSight that helps with this is Row Level Security available natively within topics. When there is Row Level Security enforced on a QuickSight dataset, and a topic is built using that dataset, the user will only receive insights about data they have access to, with no additional setup required.

Choosing the Right Data

Once you have a clear understanding of common stakeholder data questions, pain points, and how they are talking about their data, you will then naturally find the data sources containing the information best equipped to answer them. The data you choose will be the bedrock of your Amazon Q in QuickSight solution.

Setting Up Topics

QuickSight will perform an initial ‘cleaning’ of your data on its own, assigning common synonyms to field names (e.g., a field called ‘Cost of Goods Sold’ will be given the synonym ‘COGS’ automatically). You will want to double check this for accuracy, as it can make mistakes.

Once data is selected, QuickSight will prompt you to ‘Start Review’ of the topic data. This is almost always your immediate first step. Proper data cleaning and preparation are crucial steps to ensure that Q in QuickSight provides accurate, actionable insights that align with your business goals.

Field Selection

Review each field to ensure that only relevant and high-quality data is included. Consider which fields are adding value to the analysis, and which are adding noise or aren’t useful for the analysis. Include and exclude fields as needed.

Synonyms

Add synonyms to fields to accommodate different terminologies used by your business users. This will make the topic (and results) more user-friendly and intuitive.

Data Roles and Formatting

Define roles (such as dates, measures, dimensions, etc.) and apply consistent formatting. You can manually set fields to be formatted as Currency, which will maintain the clarity and overall quality of the insights.

Data Aggregations

Set up the data aggregations of measures to set up meaningful summarizations. You can set default aggregations (e.g., a ‘Sales’ field can be set to SUM or AVERAGE automatically) and disallow other aggregations that either don’t make sense or are potentially sensitive.

Testing, Deployment and Beyond

Amazon Q in QuickSight encourages continued evaluation and performance monitoring after a topic is deployed to end users. For each topic, you can see how many questions have been asked and what they are, and view user feedback of their responses. Each time a user asks a question, they have the option to give the response a thumbs up or thumbs down, and give long text feedback.

End users should be highly encouraged to interact with the results they get from Amazon Q. Their input is extremely valuable and can be used to determine what is working well with the topic and what could use improvement. Additionally, you can see how many questions have been asked in total, what portion of those questions were answerable, and what portion weren’t. Use these metrics to assess engagement with the solution and calibrate or update the topics accordingly.

Listening to Feedback

From the outset, deploying Amazon Q in QuickSight works best as a collaborative process. Integrating it into your BI ecosystem is best done by working closely with the target end users at all stages of development, from ideation to deployment. Their expectations and usage of the tool will be the biggest variable impacting the success of the project, and their continued engagement after deployment will ensure issues are identified, fixed and maintained.

Negative feedback from users should always be addressed and corrected quickly. If end users find many of their questions unanswerable or are consistently getting unexpected results from the topic, it can lead to frustration and abandonment of the tool. Remember, the value of Q in QuickSight is being able to get the information you need quickly and easily. If users are using synonyms that the development team had not considered, if their queries are too complex, or irrelevant to the topic, they will have trouble getting the results they are looking for.

A great way to start users out using Amazon Q in QuickSight is with Suggested Questions. These are questions shown to users by default when they open a topic in the Q in QuickSight interface within the QuickSight console. These can be validated by the developers of the solution. Questions can be manually validated by the developer of the topic and shown to end users by default. These can also be re-validated periodically as the data refreshes. End users are shown the last validation time.

Setting Expectations

Given the rise in popularity of other AI-powered tools, it’s important to set user expectations about what Amazon Q in QuickSight is and isn’t useful for. It is designed for quick, high-level insights while also giving a user the ability to dig deep to answer specific questions about their data. Questions posed to a Topic should always be pertinent to the data in the topic.

Amazon Q in QuickSight is not a replacement for mission-critical dashboards, nor is it a general purpose, world-wise AI virtual assistant. Situations involving intricate calculations or carefully formatted reports are better carried out in a QuickSight Analysis. Additionally, while Amazon Q in QuickSight streamlines the querying process, it still requires well-prepared and clean data to function effectively.

Final Thoughts

Amazon Q in QuickSight is an AI solution completely integrated into a full-fledged BI tool. It is secure, and enables end users to perform their own ad-hoc reporting done at their convenience. By empowering users to query data directly and intuitively, Amazon Q in QuickSight reduces the load placed on technical teams, accelerates decision-making processes, and fosters a more data-driven culture within the organization. These make Amazon Q in QuickSight a valuable asset for any organization looking to enhance its reporting capabilities and drive positive business outcomes..