Dev.to
7/10/2026
The original title is: "Open-source AI voice agent automates first-round candidate phone screens with structured scorecards"
Original: I Built an AI That Calls Candidates and Scores Them So Recruiters Don't Have To
Short summary
A developer built an AI voice agent using Telnyx AI Assistant and GPT-4o that conducts first-round phone screens, asks five questions, and produces a structured scorecard with scores and recommendations within seconds. The architecture replaces webhook-based call control with a simpler three-API-call model: trigger the call, let Telnyx host the conversation, then poll for the transcript and run a second inference pass to extract the scorecard. The code is open-source on GitHub.
- •AI voice agent conducts 5-question phone screens and generates scorecards automatically
- •Built on Telnyx AI Assistant with GPT-4o; three API calls replace a webhook state machine
- •Open-source code includes dashboard, call triggering, and transcript-based scoring
Generated with AI, which can make mistakes.
Is this a good recommendation for you?
