Salesforce: How Assembled Voice Agents Create Cases – Assembled

This article explains how Assembled’s Voice AI integrates with Salesforce to create or enrich support cases during customer phone calls.


Requirements

Before using Voice AI with Salesforce, ensure:


Overview

Assembled Voice AI processes every customer call and ensures a Salesforce Case is created or enriched with the full context of the conversation—whether the call is handled by the AI or handed off to a human agent.


1. Live Call Flow

During the call:

Regardless of the path, a Salesforce Case must capture:


2. Contained Call Path

For calls handled entirely by the AI agent:

  1. Assembled creates the Salesforce Case directly using the Salesforce REST API.

  2. The created case includes:

This ensures full documentation of automated conversations.


3. Handed-Off Call Path

When a call transitions to a human agent:

3.1 Case Creation & Screen Pop

  1. A Salesforce Case is created automatically by the customer’s telephony/CTI system.

  2. When the agent answers the call, Salesforce performs a screen pop that opens the created case.

Typical telephony + Salesforce CTI setups include (non-exhaustive):

These systems are responsible for:


3.2 How Assembled Detects the Case

Assembled integrates tightly with Salesforce to match the newly created case with the Voice AI conversation:

  1. Incoming-call-created Cases are detected by a Salesforce Apex Trigger.

  2. The Apex Trigger notifies Assembled that a new case was created from an incoming phone call.

  3. Using this signal, Assembled matches the Salesforce Case to the corresponding Voice AI conversation.


3.3 Enrichment of the Case

Once a match is found:

  1. Assembled updates the Salesforce Case using the Salesforce REST API.

  2. The case is enriched with:

This ensures agents have full visibility into the Voice AI portion of the conversation.


Need Assistance?

Contact Assembled Support if you need help configuring your Salesforce integration or connecting your telephony provider to Voice AI.