Skip to main content

​
AI Employee

AI employee action
Hands the conversation over to an AI employee you choose. The AI replies in the chat until one of your exit conditions is met or the reply timeout expires. Use this to automate first-line support, qualify leads, or collect information before an agent takes over.

​
Fields

  • Channel (required): Select the channel where the AI employee will reply.
  • AI Agent (required): Choose which AI employee handles the conversation.
  • Context (optional): Extra instructions or context for the AI (e.g. product info, policies).
  • Exit conditions (optional): Define when the AI should stop and hand off (e.g. when the contact asks for an agent or says “goodbye”).
  • Reply timeout (optional): Toggle and set a duration after which the AI stops waiting for a reply and the automation can continue.

​
Use cases

  • First-line support: Use the action to hand the chat to an AI employee that answers FAQs and collects info; add exit conditions (e.g. “ask for agent”) so the automation continues to Assign chat or Move chat when the contact wants a human.
  • Lead qualification: Let the AI ask qualifying questions in the chat; use exit conditions and reply timeout so the flow continues to Set attribute, Add label, or Assign chat based on the AI’s conversation.
  • Collect info before handoff: Use Context to give the AI product or policy details; when the AI finishes or times out, use the next steps to assign the chat and optionally send a summary or Set attribute with what was collected.

​
AI Intention Recognition

AI intention action
Uses AI to classify the contact’s intent into one of the options you define, and stores the result in a workflow variable or contact attribute. Use this to route conversations, tag contacts, or branch your automation based on what the contact wants (e.g. “support”, “sales”, “complaint”).

​
Fields

  • Context (optional): Description or instructions so the AI understands the conversation (e.g. “Customer support chat”).
  • Field (required): The workflow variable or contact attribute where the matched option will be saved.
  • Options (required): List of possible intents (one per line). The AI picks the best-matching option; you can add up to 20.

​
Use cases

  • Route by intent: Use the action to classify the contact’s message (e.g. “support”, “sales”, “complaint”) and save to a variable; add a Condition right after to Assign chat, Move chat, or send different messages based on the stored value.
  • Tag contacts for campaigns: Save the matched option to a contact attribute (e.g. “intention”) so you can segment or target contacts in campaigns and views based on what they wanted.
  • Branch without full AI chat: When you only need to detect intent (not hold a long conversation), use AI Intention Recognition after “Save text reply to” and branch the automation on the result (e.g. support vs. sales flow).

​
AI Transform

AI transform action
Uses AI to transform text from a workflow variable or contact attribute according to your instructions, and writes the result to another variable or attribute. Use this to summarize messages, translate, extract key info, or reformat text for downstream steps.

​
Fields

  • Input (required): The workflow variable or contact attribute that contains the source text.
  • Instructions (required): What you want the AI to do (e.g. “Summarize in one sentence”, “Extract the email address”).
  • Output (required): The workflow variable or contact attribute where the transformed text will be saved.

​
Use cases

  • Summarize or extract from the reply: Use the action with Input = the variable where you saved the contact’s reply (e.g. from “Save text reply to”); set Instructions to “Summarize in one sentence” or “Extract the order number”, then use the Output in the next message or Condition.
  • Translate for internal use: Set Input to the contact’s message (variable or attribute) and Instructions to “Translate to [language]”; save Output to a variable or attribute so agents or later steps see the translation.
  • Reformat for downstream steps: Use AI Transform to clean or structure text (e.g. “Extract only the email address”, “Convert to bullet points”) and save to a variable or attribute for use in an API call, template, or contact record.