Smart Intents concepts: Overview

End-users contact you when they need help. Nonetheless, communication barriers may disrupt their best attempts to describe what’s wrong, as well as how to fix what’s wrong.
  • Your end-users may send confusing or ambiguous requests for help.
  • They might use different words than you would use or might use familiar words in a different order.
  • Ambiguity can delay and disrupt your delivery of product support.
Smart Intents addresses these communication worries from two angles at once.
  • Rather than requiring your end-users to write about their difficulties, you can show options to them in an intent menu. By clicking well-crafted menu options, any end-user can tell you all about their intent.
  • You can even train Smart Intents’ natural language processing (NLP) engine to drive which intent menus your end-users see.
End-user selections in an intent menu can then trigger automations and bots to run, eliminating most of the product support delays that ambiguity might otherwise cause.


  • intent — [EXAMPLE: “lost password”] A label, whose topic underlies a problem or a question that end-users raise while requesting product support. Also, a container for training samples.
  • intent group — [EXAMPLE: “billing” or “gameplay”] Given a set of similar and/or closely related intents, an intent group is:

    ‣  the workspace where you create intents of that particular type.
    ‣  the broad but distinct product support category where you collocate intents of that particular type.
    ‣  a navigational heading that your end-users see in their Smart Intents menu.
    ‣  an organizational convenience that helps you to plan and populate intent menus.

  • intent menu — [EXAMPLE: “{your app name}”]

    ‣  You populate an intent menu with intents, by way of their intent groups.
    ‣  You configure one or more apps to show the published intent menu.
    ‣  End-users confirm their intent by choosing it from the intent menu that you show to them.

  • training sample — One uniquely phrased way that an end-user could describe the problem (or phrase the question) that gives them their intent.

    ‣  Intent-matching accuracy improves with each new training sample that you add.
    ‣  When you enter and save new samples, improvements to intent-matching accuracy might not take effect for 20 minutes.

Administrator process flow

As an administrator, your process to configure and use Smart Intents includes six distinct parts.
Illustrated summary of steps described elsewhere in text.

End-user process flow

For an end-user, the process to trigger Smart Intents and benefit from it is nearly as simple as sending a chat message.
Illustrated summary of steps described elsewhere.

Articles about Smart Intents include the following.

Helpshift (Main Site)