About actions for AI agents (Ultimate)

Actions let your AI agents perform certain tasks during conversations with end users. Actions can be used to:

Enrich the conversation experience for the end user by leveraging parameters to create more dynamic and responsive conversation flows.
Retrieve or send conversation details to your CRM platform so you can use important customer information during conversations or store for later.
Improve your understanding of your customer conversations overall by applying situation-specific labels to conversations for later analysis.

Actions can be applied at different points in the conversation journey, giving you flexibility to manage conversations, personalize responses, and interact efficiently with your CRM system.

This article contains the following topics:

Types of actions
Where actions can be applied

Types of actions

There are two types of actions:

Conversation actions
CRM actions

Conversation actions

Conversation actions allow AI agents to perform tasks during conversations with end users. These actions are versatile and can be used for various purposes, including:

Creating parameters: Save parameters to use in conditional logic within the dialogue builder, enabling more dynamic and responsive conversation flows.
Modifying parameters: Update existing parameters, such as changing the language of the conversation based on user preferences or other criteria.
Labeling conversations: Tag conversations, making them easier to track and analyze within conversation logs and analytics.

For a list of available conversation actions, see Available conversation actions for AI agents (Ultimate).

CRM actions

CRM actions enable communication between AI agents and your CRM platform. These actions are platform-specific, meaning different actions are available for each CRM platform.

For a list of available conversation actions for each CRM platform, see the articles below:

Where actions can be applied

Actions, whether they’re conversation actions or CRM actions, can be applied at three different levels:

Bot-level actions
Intent-level actions
Block-level actions

Bot-level actions

Bot-level actions are executed in every conversation your AI agent engages in. They require an event and an associated action.

For example, you could create an action associated with the “Chat started” event to automatically pull in visitor information from your CRM when a conversation begins.

To create a bot-level action, see Adding an action at the bot level.

Intent-level actions

Intent-level actions are executed when a specific intent is triggered during a conversation. You select the action you want to execute when the intent is triggered. Intent-level actions are particularly useful if you want to apply actions to all replies within an intent.

For example, you can apply conversation tags or internal notes based on the intent.

To create an intent-level action, see Adding an action at the intent level.

Block-level actions

Block-level actions are executed when a particular block within a conversation flow is reached.
These actions are useful for more granular control within your dialogues.

For example, you could add labels to conversations, check for specific parameters, or store detailed information relevant to the ongoing conversation.

To add a block-level action, see Adding an action at the block level.