Answers generated from Webex transcript and chat interactions
Q: Does Q business support incremental indexing for unstructured files?
A: Yes, it does. Once it is uploaded, it will start reindexing based on any new information that comes in.
Q: Can you join data that’s indexed data? For example, products that are in one file and sales that are in another file - can you join product and sales in a dataset from indexed files?
A: Yes, as long as the information is indexed and based on the prompt, if Q thinks there’s relevant information that needs to be combined from two different sources within the same application, it does join and responds back to the prompt.
Q: Why does Product L through M show different numbers from the first answer versus unstructured data in the insights (178K vs 191K)?
A: There might be some additional information flowing from the insights. In real world use cases, you will see valid numbers depending on the source that is getting into the Q&A. If there is no additional information flowing from unstructured data through the file, then it will show the same numbers.
Q: Does the data stories feature automatically include visuals to support stories?
A: No, visuals need to be added manually. In data stories, as a user you might have access to multiple dashboards and you can combine visuals/information from different dashboards. Data stories gets its information from these visuals as a structured source, so they have to be manually selected from the dashboards you have access to.
Q: Can we create automations for dynamic Q business indexing of incoming unstructured files?
A: Yes, there are APIs available on the Q business side (not part of Quick Sight) which you can use to trigger the indexing and add sources from S3 to automate it. There is a published workshop that has steps showing how to automate that part.
Q: If you type in the same question again and considering the backend data didn’t change, would you get the same response and in the same document format?
A: Mostly it is the same. Sometimes it tries to reiterate and give a better response if it wasn’t able to get some information on the first try, but it will be about 80% same if there isn’t much change in the data points or unstructured information.
Q: Does Q support doing sentiment analysis on unstructured data (like marketing lead data)?
A: Yes, Q can do sentiment analysis with unstructured information, though this wasn’t specifically demonstrated in this session. The Scenarios feature may be particularly useful for this kind of analysis.
Q: Can we create multiple Q business apps and grant different access?
A: No, this is not supported today.
Q: Can we connect Athena as a data source and use it in Amazon Q?
A: Yes, whatever data sources are supported in Quick Sight (like Athena, Redshift, etc.) are the same data sources supported for Amazon Q in Quick Sight.
Q: Can Quick Sight Q business read P&L documents and give insights?
A: Yes, if the document contains P&L information, it should be able to provide insights from it.