Skip to main content
The Intellixent tool is the core connector between your AI calling system and your automation flows. It gives you dedicated triggers that fire on call events and actions that let you manage leads, inject context, and drive post-call workflows — all without code.
In the flow builder’s search sidebar, the Intellixent tool appears at the top of the list whenever you search for a trigger or action.

​
Triggers

​
Call Ended

The Call Ended trigger fires immediately after an AI call completes. Use it to kick off post-call workflows like CRM updates, follow-up scheduling, or team notifications. Data available in the trigger payload:
  • Complete conversation transcript with timestamps
  • AI assistant responses and decisions made during the call
  • Call duration and technical metadata
  • Customer phone number and caller information
  • Call outcome and sentiment analysis
  • Variables set during the conversation
To configure:
1

Add the trigger

In the flow builder, click the trigger block and search for Intellixent. Select Call Ended.
2

Choose an assistant

Select the specific AI assistant you want to monitor. The trigger fires for calls handled by that assistant.
3

Map variables

Configure how call data maps to the variables your downstream actions need.

​
Inbound call variable injection

The Inbound Call trigger fires before your AI assistant picks up an inbound call. Use it to fetch real-time data about the caller and return variables that get injected directly into the AI’s prompt. This lets your AI greet callers by name, reference their account history, and adapt its behavior before saying a single word. Data you can inject:
  • Customer profile and preferences from your CRM
  • Conversation history summary
  • Business rules and routing constraints
  • Custom AI behavior flags
  • Dynamic response templates
To configure:
1

Add the trigger

In the flow builder, select Intellixent → Inbound Call.
2

Define required variables

Specify the variables your AI assistant’s prompt expects. These are the keys your flow must return.

​
Actions

​
Add lead to campaign

The Add Lead to Campaign action enrolls a contact into one of your Intellixent outbound calling campaigns. Use it to automatically follow up with leads after they interact with your AI. This action supports creating leads from external sources including Google Sheets, HubSpot, and Facebook Lead Ads. To configure:
1

Select the target campaign

Choose which AI campaign the lead should be added to.
2

Map contact details

Map the caller’s phone number, name, and other fields from the trigger data to the lead record.
3

Add custom attributes

Include any additional context — like call outcome or expressed interest — that the AI should reference on the follow-up call.

​
Return variables

The Return Variables action is used exclusively with the Inbound Call trigger. It sends a structured response back to the AI assistant so the injected variables are available in the conversation prompt. Variables you can return:
  • Customer profile and preferences
  • Conversation history summary
  • Business rules and constraints
  • Custom AI behavior flags
  • Dynamic response templates
To configure:
1

Structure the response

Build the JSON payload your AI assistant expects. Each key should match a variable defined in your assistant’s prompt.
2

Map customer data

Use the Data to Insert panel to pull values from your CRM lookup or other upstream steps into the response fields.
3

Set AI parameters

Include any conversation parameters — like tone, language, or routing logic — that should influence the assistant’s behavior.
4

Configure timeout handling

Provide fallback values for every variable so the AI can proceed gracefully if a data fetch fails or takes too long.

​
Example flows

Trigger: Call Ended
↓
Extract conversation insights
↓
Update HubSpot record
↓
Add to follow-up campaign if qualified

​
Best practices

Keep the processing in your Inbound Call flow under 2 seconds. The AI assistant waits for the injected variables before answering the call, so a slow flow creates a noticeable delay for the caller. Set sensible defaults for every variable so the flow can return quickly even if a data source is slow.
Match the variable names and data types your AI assistant’s prompt expects exactly. Validate all values before returning them, structure nested objects clearly, and document any custom fields so future edits are straightforward.
Always provide fallback values for every variable. Log injection failures so you can spot patterns, monitor response times in the Runs view, and handle missing data gracefully rather than returning an empty or malformed payload.