AWS on Tuesday launched the public preview of new generative AI capabilities in QuickSight, the tech giant's analytics platform.
AWS CEO Adam Selipsky unveiled Amazon Q during his keynote at AWS re:Invent 2023, the company's user conference in Las Vegas.
Amazon Q is AWS' version of a generative AI-driven chatbot, similar to Microsoft's Copilot and Duet AI from Google. And just as Microsoft and Google are doing with Copilot and Duet AI, respectively, AWS is planning to integrate Amazon Q with some existing platforms, including QuickSight.
Doug Henschen, an analyst at Constellation Research, called Amazon Q the most significant of the generative AI capabilities unveiled by Selipsky, noting that it will not only enable conversational interactions but also apply machine learning to understand a user's needs.
"Amazon Q was the most broadly compelling and exciting GenAI announcement," he said. "Amazon Q … will engage in conversations based on understanding of company-specific information code and technology systems. The promise is personalized interactions based on user-specific roles and permissions."
QuickSight already included a natural language processing feature called QuickSight Q, first introduced in September 2021, that enabled some natural language interaction with data. In July, AWS unveiled two generative AI capabilities planned for QuickSight Q.
Now, those two generative AI capanilities are included in Amazon Q for QuickSight, along with two others.
Now in preview
Analytics has long been the domain of trained data experts.
Attempts have been made with varying degrees of success to make business intelligence platforms accessible to business users. But despite those attempts, including low-code/no-code tools and limited natural language processing (NLP) capabilities, analytics use within organizations has largely been stagnant over the past decade.
Generative AI chatbots have the potential to change that by enabling true conversational interactions with data. Their large language models (LLMs) include extensive vocabularies that can eliminate the need to write code or learn the business-specific language required by the NLP tools developed by analytics vendors themselves.
As a result, many analytics vendors beyond just the tech giants have made generative AI a focal point of their product development over the year since OpenAI launched ChatGPT.
For example, Tableau is incorporating Einstein GPT from Salesforce, its parent company, and Qlik unveiled Staige, a portfolio of tools aimed at helping customers integrate generative AI with their data.
The new Data Q&A capability in QuickSight, fueled by Amazon Q, will enable the conversational interactions between users and their data that will expand analytics use to potentially any employee within an organization whose role can benefit from data.
Therefore, even though Amazon Q in QuickSight is similar to Copilot in Power BI from Microsoft and Duet AI in Looker, it is a significant development for QuickSight customers, according to David Menninger, an analyst at Ventana Research.
David MenningerAnalyst, Ventana Research
"It's significant from the perspective that natural language makes analytics much more accessible," he said. "Prior to ChatGPT [from OpenAI, in November 2022] hitting the market, natural language was not widely used in analytics. Amazon Q in QuickSight delivers the natural language capabilities."
Using Data Q&A, QuickSight customers will be able to explore data beyond existing reports and dashboards.
Using existing dashboards and reports as a starting point, customers -- using conversational language rather than code -- will be able to ask questions of the underlying data whose answers aren't seen within the existing data products. For example, if a dashboard displays sales figures for a region, users can dig deeper by asking about sales figures for a specific city within that region.
Once they get a response, they can keep exploring with follow-up questions that help them better understand what is happening in their business.
Beyond Data Q&A, Amazon Q in QuickSight includes Stories -- which was previously named Ask Q -- Build for Me and Executive Summaries.
Stories is a data storytelling tool that automatically develops narratives that explain data.
Decision-making is often a collaborative process. Rather than require the time-consuming task of developing PowerPoint presentations and other ways of sharing information, Stories enables users to create narratives and visuals that describe data by simply going to the "Build" menu in QuickSight.
Build for Me is a tool for dashboard developers. Rather than write the reams of code previously required to cultivate data and build a dashboard, developers can type commands in natural language.
And Executive Summaries automatically highlights interesting facts and statistics within reports and dashboards and deploys LLMs to write digestible summaries. Its intent is to save business users time by providing snapshots in natural language.
Beyond being able to ask questions of data, Stories is perhaps the most significant of the capabilities enabled by Amazon Q in QuickSight, according to Menninger.
"Stories is important and valuable because [presentations] are the way things are done in businesses," he said. "We need ways to combine narratives and visuals, and most of us do that in PowerPoint. We need the visuals and we need the explanations. A tool that helps put those things together makes a lot of sense."
From a comparative standpoint, Amazon Q in QuickSight represents one of the most advanced set of analytics-related generative AI capabilities, Menninger continued.
However, it is among the vanguard, rather than alone.
"It's at the forefront but not necessarily distinguishable from the others," Menninger said. "This is the state of the market, which is a natural-language Q&A interface, story building, the ability to ask follow-up questions -- those are things AWS talked about that seems to be the state of the art."
Henschen, meanwhile, noted that the totality of Amazon Q in QuickSight goes well beyond what AWS introduced with QuickSight Q in July. Those initial iterations of generative AI tools were being developed using Amazon Bedrock, an AI service that provides access to LLMs.
"The idea with Amazon Q is to deliver an AI assistant that will understand the context of data, text and code," he said. "It's pretty clear that Amazon Q will be the one, uber AI assistant that will provide more nuanced, contextual understanding of what users are seeking when they ask questions in natural language."
With Amazon Q in QuickSight's initial capabilities now in public preview, one opportunity for further growth of the tool is to expand Q's understanding of code, according to Henschen.
While many analytics vendors have focused on natural language query as a means of incorporating generative AI, many data management vendors have used LLMs to develop text-to-code translation capabilities.
"Having a more capable Amazon Q service with broader and deeper understanding of not just data but also text and code will benefit users wherever that service shows up," Henschen said. "QuickSight will be one of the starting points."
Menninger, meanwhile, said he's curious to see how uniformly AWS is able to incorporate Amazon Q into its various tools.
Beyond QuickSight, Amazon Q is being introduced in customer service and supply chain management tools, among others. If AWS is able to effectively infuse Amazon Q throughout its vast array of platforms, that could help differentiate Q from other chatbots.
"If you think about having a common platform for all these different natural language interfaces for all these different parts of your business operation, that's pretty exciting," Menninger said. "That would be a way for AWS to distinguish themselves certainly from the analytics-only players."
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.