Authoring Q&A
Converting to the Generative Q&A experience
If you have existing Amazon Q topics, you can easily convert these to leverage our new generative capabilities. Navigate to a topic, and then choose Convert next to the topic name. You will then be prompted to Duplicate & Convert Topic in a dialog box. We duplicate your topic for you so that the conversion to our beta experience does not impact your end users. Once you are satisfied with topic performance in the new experience, you can unshare the original topic and share the new one.
Named entities
Named entities are one of the most important components of topic curation. The information contained in named entities — specifically, the ordering of fields and their ranking — is what makes it possible to present contextual, multi-visual answers in response to even vague questions. Authors can find named entities by navigating to a topic, choosing the Data tab, and then choosing the Named Entities. From here, authors can preview or edit existing named entities, and create new ones.
Authors can configure the following facets of named entities:
Fields: Choose a dataset, and then choose which fields from that dataset to include. This defines the scope of data that will be considered when using this named entity to answer enduser questions.
Field Rank and Presentation: The relative rank of the dimensions and measures in a named entity determines how those fields are used when generating contextual, multi-visual answers. Note in the following demo that adjusting the relative rank of Profit so that it is higher than Sales leads to different data being displayed. By default, the order of fields in the table visual is the same as the field rank. However, you can control these two individually by turning off Sync table view with field ranking.
Show / Hide in Presentation: Fields that are included in named entities can simultaneously be hidden from the tabular presentation of the named entity, while still providing additional context in other components of the answer.
Measure aggregations
Authors have fine-grained control over how Amazon Q aggregates measures in topics. Across QuickSight,
measures are defaulted to SUM,unless
they have custom aggregations defined in a calculated
expression.
To change this in Q, navigate to the measure in the list of data fields, and specify
a different default aggregation. You can also disallow aggregations, which will
prevent them from being applied even if a user specifically asks for them. Lastly,
you can specify that a measure is non-additive. This is useful for pre-computed
metrics, such as percentages, which should not be re-combined in any way. Doing so will force Amazon Q
MEDIAN or AVG depending on your use case.
