Event-Driven Customer Engagement Chat Application
Built a resilient chatbot that improved customer engagement while handling unpredictable demand.
Impact: Reduced time-to-value by ~35% (first completed story) and improved completion rate by ~28% (draft -> publish) on customer story creation.
View architecture decisions
Architecture: Event-driven workflow using serverless functions, queue-backed task processing, and a messaging integration with Twilio.
Challenge: The system had to support unpredictable usage spikes while preserving response quality and low latency.
Design Decisions & Tradeoffs: Integrated with 3rd party APIs for direct Twilio communication and Azure Functions for scalable compute, while implementing robust retry and fallback logic to handle external service variability.
Scalability & Performance: Designed for bursty traffic with horizontal scaling and lightweight caching for high-frequency prompts.