Product Management
Conversational AI Voice Agent, HEARD 




HEARD uses AI to moderate short interviews with your customers and discover what they need, fast. After launching this customer research tool, teams were able to interview hundreds of customers in a few short hours instead of spending weeks scheduling and moderating a handful of in-person customer interviews.


The Challenge
How can we make the conversation between research participants and HEARD more natural and lifelike while also managing many different scenarios within an interview?
Work
Led development of a real-time voice agent that conducts interviews with customers. 


reativecommons.org/lic
Hanna Barakat  & Archival Images of AI + AIxDESIGN / betterimageofai.org   ses/by/40
Process
I supported our engineers in designing infrastructure pipelines to ensure we balanced turn detection, speech-to-text accuracy, flexibility and model latency. 

Throughout this process I lead the evaluation of multiple set-ups and six models from various companies to arrive at the best solution, powered by a proprietary text-based AI model. Among the conversational AI platforms tested were OpenAI’s Whisper, Google’s voice synthesis, Elevenlabs’ Conversational AI, LiveKit, Gemini AI and browser-based TTS. 


Today, HEARD’s voice agent increases user insights three-fold, and offers companies a way to collect richer data on-par with in-person conversations.