How OutCallerAI works
Understand how use cases, agents, leads, calls, knowledge, and integrations work together.
OutCallerAI turns workspace configuration into voice and messaging workflows. Operators configure the behavior in the dashboard; OutCallerAI executes conversations and records the resulting activity for review.
It is designed around a simple separation of concerns: use cases decide what the workflow should accomplish, agents decide how the conversation should behave, and integrations decide how information enters and leaves the workspace. Keeping those responsibilities separate makes an agent reusable and keeps operational changes easier to audit.
Core workflow
- Create a use case to organize an outbound workflow.
- Configure an agent with a voice, objective, instructions, and conversation boundaries.
- Add knowledge when the agent needs approved product, policy, or organization information.
- Import leads manually or through an integration.
- Run a test batch and review conversations, recordings, transcripts, and outcomes.
- Connect downstream systems with calendars, messaging, CRM integrations, webhooks, or the public API.
- Monitor performance from the dashboard and refine the workflow.
How the pieces relate
- A workspace owns configuration, access, usage, knowledge, and integrations.
- A use case groups an outbound workflow and its leads.
- An agent defines the conversational behavior used during a call.
- A lead contains the customer information and campaign-specific context.
- A conversation records what happened, including transcript, recording, status, and outcome.
- An integration moves information into or out of the workflow.
A practical example
For an appointment-qualification workflow, the use case stores the objective, schedule, retry behavior, and lead list. The assigned agent asks the qualifying questions and follows the conversation instructions. A knowledge base supplies approved service details. Google Calendar or Calendly checks availability during the call, while HubSpot or a webhook receives the final result.
The conversation record connects those pieces after execution. It gives operators the transcript, recording when available, status, summary, and captured context needed to understand the outcome.
Choose the right surface
| If you need to… | Start here |
|---|---|
| Understand call volume or a sudden performance change | Dashboard |
| Build a repeatable outbound workflow and manage its contacts | Use Cases & Leads |
| Control voice and conversational behavior | Agents |
| Review what happened in an individual interaction | Conversations |
| Supply approved business information to an agent | Knowledge Base |
| Connect a CRM, form, calendar, mailbox, or webhook | Integrations |
| Automate workspace records from your own software | API setup |
Start with a controlled rollout
Create the smallest version of the workflow that can complete the business objective. Test it with representative leads, inspect the evidence in Conversations, then increase volume after the agent and integrations behave consistently.
Before increasing volume, confirm consent and calling-hour requirements for the regions you serve, limit workspace access to the people who need customer data, and verify that downstream systems handle duplicate or delayed events safely.
Continue with Use Cases & Leads or Agents.
How is this guide?