AI VOICE AUTOMATION / LIVE SAAS
An outbound AI voice agent that acts the moment a towing job arrives.
A production call pipeline connecting live job events to contextual Retell AI conversations, outcome tracking and controlled retries.
01 / The real problem
Automation starts where the process breaks.
A new towing job could require an immediate outbound call, but a dispatcher might be busy, the destination might change the right offer, and an unanswered call could disappear from the workflow. The product needed automation without taking operational control away from each towing company.
Context was fragmented
Vehicle, destination and service details arrived as job data, while the call script needed all three to sound relevant.
A call was not a result
Completed, failed and unanswered calls needed distinct states so the team knew what happened and what to do next.
One rule did not fit every tenant
Each operator needed auto, manual or off modes plus service-level controls instead of a platform-wide switch.
A clear path from trigger to outcome.
Detect
A Towbook job event enters the pipeline with its vehicle, service and destination context.
Classify
The flip engine evaluates the destination and selects the appropriate offer and conversation path.
Call
Retell AI places a contextual outbound call using the tenant's enabled service rules.
Resolve
The system records the outcome, exposes it in the dashboard and makes failed or unanswered calls retryable.
03 / What I built
The implementation.
04 / Result
The finished pipeline turns an incoming job into a traceable operational action. Calls are faster, scripts use the right context, outcomes remain visible, and operators can choose exactly how much of the process is automated.
✓05 / In the product
Real screens. Real workflow.
01 — Every automated call remains visible with its outcome and retry state.
02 — A live command view gives operators oversight without interrupting the automation.