The production infrastructure behind a high-volume voice-AI operation
Published customer result: $430M+ additional annual revenue across a voice operation handling 400K+ outbound calls per month.
A production voice-AI operation for American Way Health, built around Bland AI. The engineering contribution here is the production layer that keeps the agent calling, compliant and measurable at 400K+ calls per month.
US healthcare enrollment. The published Bland AI case study reports the customer outcome; the engineering described here is the production layer our founder built around the agent at BlandLabs.
Role: Production orchestration, safety and observability engineering around the voice agent
Founder contribution delivered at BlandLabs, built around Bland AI's voice platform.
- Additional annual revenue
- $430M+Additional annual revenuePublished by Bland AI
- Outbound calls per month
- 400K+Outbound calls per monthPublished by Bland AI
- Transfer rate
- 25% → 45%Transfer ratePublished by Bland AI
- Numbers in the outbound pool
- 128Numbers in the outbound poolFounder-built production layer
The system
What was actually built
- Convoso lead ingestion and outbound call orchestration
- Bland AI API integration with transcript and callback processing
- A 128-number outbound pool with intelligent caller-ID rotation and decay scheduling
- Scheduling rules, daily call limits and redial controls
- Do-not-call handling and failed-call blocking
- Warm-transfer state tracking to human agents
- Carrier-health monitoring and number-reputation protection
- Operational statistics, reconciliation and daily reporting
- Pickup-rate heatmap and automated call analysis
- Multi-account alerts portal and production observability
Why it was hard
Engineering complexity
- Sustaining 400K+ outbound calls per month without burning number reputation
- Improving transfer rate from 25% to 45% while staying inside compliance controls
- Keeping a live revenue operation observable enough to debug in production
- Multi-tenant alerting so operators see carrier problems before customers do
Technology
Verified stack
Only technologies confirmed in source code, project documentation or the live deployment are listed.
- Node.js
- TypeScript
- REST APIs
- Webhooks
- Bland AI APIs
- Convoso APIs
- AWS EC2
- Docker
- PM2
- NGINX
- TLS/SSL
- Scheduled background processing
Delivery
Scope of responsibility
- Control plane and routing
- Safety rails and compliance controls
- Observability and dashboards
- Number-pool orchestration
- Warm-transfer detection
- Carrier and reputation monitoring
Evidence
Published evidence
Related
Nearby systems
MyPlanAdvocate$1.2M/yr
An AI-assisted Medicare and ZIP-code qualification workflow with real-time decisioning, structured hand-off and operational reporting, credited with a reported saving of about $1.2M per year.
Mixed Reality Desktop App42 days
A mixed-reality "Selfie Point" desktop application for the Lucknow Vidhan Sabha: live camera feed, real-time object detection and positional tracking compositing visitors into mixed-reality content, delivered in 42 days.
Build a system with similar complexity
Book a free 30-minute consultation, or write to us — you will always talk directly to the person who builds your product.
Techon Pixel · info@techonpixel.com