Atenea Call is an AI-powered last-mile delivery assistant designed to resolve real delivery exceptions through phone calls.

The project was inspired by recurring operational problems in last-mile delivery: customers not answering, unclear delivery instructions, identity verification, authorization limits, secondary contacts, delivery restrictions, and situations that require escalation.

Atenea Call connects an operational interface with CALL-E so the system can:

identify the order and contact context; verify identity before authorizing sensitive actions; interpret the customer’s request; apply order-specific permissions and restrictions; communicate a clear decision to the customer or courier; escalate when the requested action is not permitted; keep an auditable trace of the interaction.

The MVP includes multiple fictional delivery scenarios, contact orchestration, primary and secondary contacts, authorization rules, retry logic, escalation, timeline monitoring, transcripts, summaries, and JSON/text exports.

A real CALL-E integration was successfully demonstrated during development and is shown in the submission video. We later updated the interface so judges could select an order, choose a scenario, and launch the demo from a single button.

The final validation from this updated interface was blocked by CALL-E issue #229, which remained unresolved at submission time. The issue was documented with the provider, including Plan IDs, timestamps, logs, and controlled test results.

Despite that external blocker, the product workflow, orchestration logic, UI, monitoring layer, and real-call integration were fully developed and demonstrated during the hackathon.

Built with: Node.js, JavaScript, CALL-E CLI/MCP, local orchestration services, real-call monitoring, and a custom last-mile simulation interface.

Built With

Share this project:

Updates

Submission history