Build a Voice AI Agent to Schedule Interviews

Most small staffing agencies start by sending candidates Calendly links. This is passive and puts the burden on the candidate. For high-demand roles, less than 50% of candidates will book a time, and the ones who do not are often the most valuable. This approach cannot proactively follow up or escalate when a top candidate goes dark.

A few agencies try off-the-shelf call center AI tools from providers like Dialpad. These systems are designed for sales or support, not recruiting. They are expensive, with per-seat fees often exceeding $75/month for AI features. Worse, they cannot read real-time availability from multiple recruiter calendars or connect deeply with an ATS like Bullhorn. The agent can take a message, but it cannot complete the primary job: booking a confirmed time on the calendar.

This forces a broken workflow. A recruiter has their ATS open, their calendar in another tab, and their phone dialer in a third. They spend their day cross-referencing information and leaving voicemails instead of talking to qualified candidates. The core bottleneck, the manual act of scheduling, remains untouched.

Syntora's approach to automating interview scheduling begins with a thorough discovery phase. We would audit your existing Applicant Tracking System (ATS) API capabilities, current recruiter scheduling workflows, and identify all specific business rules for booking interviews.

Following discovery, we would design and implement a direct, read-write API connection to your ATS and each recruiter's Google or Microsoft calendar. Utilizing Python with the httpx library, the system would be architected for efficient, asynchronous calls, enabling rapid retrieval of candidate lists and recruiter availability.

The voice agent's conversational logic would be powered by the Claude API, managed by a robust state machine that tracks the conversation from introduction to confirmation. A specialized text-to-speech API would be integrated to provide low-latency, natural-sounding audio, with the system designed to achieve rapid round-trip response times. The agent would be engineered to intelligently handle ambiguous candidate responses, such as "sometime next Thursday afternoon," by offering specific time slots like "Does 2:00 PM or 4:00 PM work?"

The application's backend would be built with FastAPI and deployed as a serverless function on AWS Lambda. This architecture is designed to be highly efficient and scalable, triggered by a webhook from your ATS when a candidate enters a designated stage like 'Phone Screen.' This serverless approach minimizes operational overhead and keeps infrastructure costs low, even for high volumes of outbound calls.

Every call attempt, transcript, and outcome (e.g., booked, voicemail, no answer, follow-up required) would be logged to a Supabase database. A custom dashboard would be developed to provide real-time insights into booking success rates and call durations, enabling data-driven tuning of the agent's script and performance post-deployment.

Syntora has extensive experience building similar conversational AI and document processing pipelines using Claude API for sensitive financial documents, applying the same robust engineering patterns to other complex data interaction challenges. A typical engagement for a custom voice AI system of this complexity, from discovery to initial deployment and tuning, can range from 6 to 12 weeks. Key client contributions would include providing access to ATS API documentation, dedicating stakeholder time for workflow definition, and clearly outlining all desired scheduling rules. Deliverables would include the deployed, custom voice AI system, comprehensive documentation, and knowledge transfer to your team.