Use AI to provide a summary of text fields

databases provide limited capability for aggregating text - Min / Max / Count is about it.

But AI is demonstrating how it can provide a summary from a lot of text input. We could use AI to provide a text summary.

For example, we could have a product returns table that contains a text field detailing why the product was returned. We should be able to use AI to summarise all the text fields and provide an overview of all the returns descriptions.

If a filter of the table returned three records, and the three records looked like:

"product was not green coloured, more blue"

"product was a funny blue green colour"

"colour not like photo"

This would work in the exact same way as every other aggregation type. Lets call the aggregation type 'Precis'. On a text field within Data Modules you could set the aggregation to Precis.

The precis aggregation of the above 3 records would be something like

"blue products are not blue as in photo"

When the field is used in a report or dashboard, if the product encounters a single record it just gives the text as entered. If the produce encounter more than 1 record it provides a precis of all the text.

There could even be a version that works on a single field. For example, if the text entered in a field is very long, the aggreegation could summarise it down. The user could specify a max text length and when the field is greater than that length a precis is performed.

Our actual use case for this is with Planning Analytics sourced data - users provide a reason why the Forecast values vary from Actuals. This provides thousands of records. We want an easy way of summarising all those variance descriptions.