Fieldwork 🧭
Fieldwork is an AI-native UX research platform that solves the scaling problem of modern research by using Gemini as a collaborator and a 24/7 researcher.
Pitch Video: https://youtu.be/erRJxkb0BQw
The Scaling Problem in UX Research
UX research workflows have a significant scaling problem. Traditionally, conducting high-quality studies requires extensive manual effort to create guides, recruit participants, and moderate sessions. This manual bottleneck limits the frequency and depth of insights teams can gather.
The Solution: Gemini as your Research Collaborator
Fieldwork transforms the research lifecycle:
- Guided Creation: Anyone can create a professional research guide simply by answering questions about their goals, audience, and preferred style. Gemini collaborates with you to refine the depth and focus of the study.
- Gemini Live Interfacing: Deploy a bi-directional research agent available 24/7. Participants can engage in natural, voice-first conversations that capture nuanced feedback that static surveys miss.
- Multilingual Support: Capture feedback from diverse, global audiences with native multi-language support.
- Instant Insights: All feedback is captured as high-fidelity transcripts, with actionable insights generated on demand.
Meeting Hackathon Requirements
Leverage a Gemini Model Fieldwork is powered by the Gemini 2.5 family of models: Real-time Moderation: Uses gemini-live-2.5-flash-native-audio for low-latency, bi-directional voice interactions. Synthesis & Analysis: Uses gemini-2.5-flash for generating tailored research guides and synthesizing session insights.
Built using Google GenAI SDK The agent infrastructure is built using the Google Cloud Vertex AI Node.js SDK (@google-cloud/vertexai). We utilize: Multimodal Live API (WebSockets): For real-time, bi-directional agent interactions. Tools / Function Calling: For interactive multimodal UI elements (e.g., multiple-choice questions). Stream Generation: For real-time text synthesis and guide generation.
Use Google Cloud Services Fieldwork is a cloud-native application deeply integrated with the Google Cloud Platform (GCP): Vertex AI: Hosting all Gemini 2.5 models and managing the real-time agent infrastructure. Cloud Run: Containerized deployment of the Next.js application, utilizing service-to-service authentication (ADC). Cloud Storage (GCS): Secure persistence of interview recordings and transcribed data via signed URLs. IAM / Service Accounts: Secure, ambient authentication across all services without hardcoded keys.
Built With
- gcp
- gemini
- vertex
Log in or sign up for Devpost to join the conversation.