Natural language enhancements to the Assistant

The Assistant now understands comparatives and additional filters. You can also associate a term to a field name so that the Associate recognizes the term to be a synonymous word for the field name.

Important: Text input is supported in English only.

Comparatives

You can ask the Assistant to compare multiple columns or two data values with or without a timeline. For example, you can ask the Assistant:
  • How do marketing costs for the last two years compare?
  • What is the difference in marketing costs between the last two years?
  • Compare the average ticket of Store A and Store B

When there is one data value and a timeline, the Assistant compares the same data value at different times. If you want to compare multiple measure columns with a timeline, the timeline only works as filter value.

For more information, refer to Comparing data in the Assistant.


Filters and modifiers

The Assistant now understands additional filters and modifiers such as above, highest, most, best and worst. For example, you can ask the Assistant the following:

  • What city has average ticket above 90,000?
  • Which product has the highest revenue today?
  • Which product has the worst NPS Score?

Previously, the Assistant understood only certain filters such as top and bottom. When you asked the Assistant to show the top product by sales, it returned the top five results. In Cognos Analytics 12.0.0, the Assistant returns the top result only.

Another update to filters in Cognos Analytics 12.0.0 is that the Assistant now applies filters after any aggregation is applied to the data. Previously, the Assistant applied filters to the data before aggregations.

For more information, refer to Filtering in the Assistant.


Synonymous terms

You can associate terms to field names in your data so that the Assistant recognizes these terms as synonymous terms. For example, if you tell the Assistant that revenue means sales, then when you ask the Assistant to show revenue, it returns results related to the sales in your data.