I tried Quicksight Q for a month and here is what I found…
I recently learned about Generative BI at a tech conference, and the presentation on Quicksight Q was impressive. I decided to try it for a month and write a detailed review from a data analyst’s point of view.
What’s Generative BI?
Generative Business Intelligence (BI) refers to the integration of generative artificial intelligence (AI) techniques with traditional BI tools to enhance data analysis, visualization, and reporting capabilities. This integration leverages AI models, such as large language models (LLMs), to enable users to interact with data more intuitively, often using natural language.
In essence, Generative BI refers to the application of AI in the realm of data analytics.
Amazon Q in QuickSight
It costs an additional $300 per month on top of QuickSight subscriptions but luckily it is available as part of the 30-day Amazon QuickSight Enterprise Edition free trial. To try Amazon Q in QuickSight, a Pro user role needs to be invited and assigned.
In the beginning, I thought only by providing data to AI and it would understand everything and do the analysis. It turns out that some prep work are required to make it happen.
Prepare Q
1. Create a topic, from a dataset or an existing analysis
2. Create named entities and curate other topic metadata
3. Start asking questions, and share with business users
According to guidelines, topic preparation is needed to provide context to the data provided. It’s not all magic after all, it requires human input set-ups to produce reports.
Adding Friendly name and Synonyms to let AI understand different words
Named entity
Named Entities are groupings of data fields that collectively represent a business concept and are used to enhance the Q&A experience.
In my case, the real estate dataset that includes property listing details and lister information, I have put alternative names and chosen data columns that I want to be included.
As you can see, most suggested questions could have been more insightful…
My two cents
- Lack of contextual knowledge is no joke, dump questions are given even after the Data Topic is prepared. Not all business models are like the marketing dataset used in the demo. I guess that’s why human input is important here, training Q to understand the contextual knowledge is a bit time-consuming than I thought.
- Some plot type is not integrated, and impossible to be presented by Q no matter how many times it’s trained. For instance, unlike the world map, it seems hard for Amazon Q to understand Australian suburbs only by providing postcodes. I attempted to update the tree map to a regional map using alternative plot types manually but failed.
- Some might wonder if Quicksight Q is worth the investment. In my opinion, it is not. I found it more troublesome than useful, and it is also quite expensive, given the monthly subscription only costs $24 per author, spending $300 for what feels like a cumbersome Gen BI tool doesn’t seem justified. I found that ChatGPT offers better data analysis and insights.
I do like the executive summary at the top, it can discover trends like no sales were found after 14 May, which can be useful for anomaly detection.
Another “more advanced” Gen BI feature by Quicksight is Paginated Reports, a fully managed cloud-based business intelligence (BI) solution for creating, scheduling, and sharing reports and data exports. Unfortunately, there is no Free Tier option. It costs $500 for 500 reports, or $1 per report unit, which isn’t a terrible deal, but it’s beyond my budget.
Amazon Quicksight isn’t the only tool with Gen BI integration; nearly all popular visualization tools have it. Here’s an overview of their pros and cons.
Can generative BI replace data analysts?
Here is an answer AI gives — Generative BI is unlikely to completely replace business/data analysts, but it will significantly augment their capabilities. The future will likely see a collaborative relationship where AI handles data-intensive tasks and analysts provide the strategic oversight, contextual understanding, and creative problem-solving needed to drive business success.
Agreed, Gen BI is a tool that enhances analyst work efficiency but contextual understanding and strategic guidance are still required from human. At least, for now.
Maybe in 10 years, analysts will be replaced by new roles — “BI Model Validator,” and “BI Strategy Consultant”, just as new technologies have historically led to the creation of new jobs, we expect that generative BI will contribute to the emergence of novel job opportunities designed to harness this technology.
The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.
We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.
Submit your own writing here if you’d like to become a contributor.
Happy learning!
-Team Maven
