AI Competitive Intelligence Agent
Built an autonomous outbound AI agent automating market research through phone-based competitor analysis at scale. Winner at the YC Hackathon.
View sourceOverview
Built an autonomous outbound AI agent that automates market research by conducting phone-based competitor analysis at scale. The agent calls competitor businesses, gathers structured intelligence, and populates an analytics dashboard — replacing hours of manual research with an automated pipeline.
This project won the YC Hackathon, competing against teams building on cutting-edge AI infrastructure.
Voice pipeline
Engineered a Pipecat voice pipeline combining Gemini 2.5 Flash for reasoning, Deepgram for speech-to-text, and ElevenLabs for text-to-speech. The pipeline achieves an 85% call completion rate through careful optimization of call flow, interruption handling, and natural conversation pacing.
Analytics dashboard
Shipped an analytics dashboard built with Next.js visualizing data from 50+ competitors. Includes ML-based detection for distinguishing human respondents from AI phone systems, allowing the agent to adapt its conversational strategy in real time. Built on Python and FastAPI with Twilio handling telephony.