Skip to main content

Detect chatter information

Overview

Smart Detection uses machine learning to create a more intelligent chat experience. By detecting important chatter information such as names, locations and dates, and presenting the chatter with a list of previously mentioned entities that look like matches, chatters are saved from having to repeat themselves.

In the below example, the chatter opens the conversation by introducing herself, and asks to book a flight to Vancouver. By providing a list with the captured text, along with one or more prediction options that can be clicked, the bot saves the chatter’s details and requests.

smart_detect.jpg

Configure Smart Detection

To configure Smart Detection:

smartdetect.jpg
  1. Drag a Capture block into an Answer.

  2. In the Content field, enter a question that will help your bot confirm chatters’ information.

  3. Under Validation select Text.

  4. Under Save Response As, select or create a new variable that categorizes the information you are confirming, and assign the variable to an entity type. Current entities are name, location, date or time.

  5. Under Smart Detection select the corresponding entity type, under either Name, Location, or Date or Time.

smart.jpg

For non-English chatters, we will attempt to detect entities in the translated-to-English version.

This feature is currently only supported on the Chat platform. In other channels, the chatter experience will not include Smart Detection.

If you have additional entities you’d like to have detected, please share them with your ACX consultant.

Have any questions? Contact your Ada team—or email us at help@ada.support.