Power BI Blog: Q&A Improvements for Inferred Results
1 July 2021
Welcome back to this week’s edition of the Power BI blog series. This week, we will look at how Q&A is being improved in the June update of Power BI Desktop.
When your end-users’ questions are ambiguous or incomplete, Q&A will need to make some assumptions to produce a visual for these kinds of questions. For these scenarios, Q&A would return a visual and natural language restatement (shown under the (i) icon on the right), based upon Power BI’s interpretation of the end-user’s question. However, it was not evident to users which parts of the result were inferred, if any. In the June update of Power BI, this experience has been modified, by adding a ‘Showing results for’ restatement for these cases, with explicit emboldening for the inferences made by Q&A:
For example, for ambiguous questions, Q&A assumptions could be based upon the model or use of certain syntax, such as implicit aggregations or other derived entities. In the example question below, we see that the question “top products” is too vague, but Q&A made multiple assumptions based off the model and returned the result for Top 10 products by product unit price.
In the ‘Showing results for’ restatement, “10” and “by product unit price” are shown in bold because Q&A made these assumptions to return a useful result. The user can then click on the restatement for the visual to automatically update the question and tweak the question further.
That’s it for this week!