Sales Sensei

Most sales dashboards are graveyards of data—static, complex, and silent. I built Sales Sensei to turn that data into a living conversation.

By leveraging the Gemini 1.5 Flash model via the Google GenAI SDK, I’ve created a multimodal agent that doesn't just 'read' spreadsheets; it 'sees' the physical charts you hold up to your webcam. Using Google Cloud Firebase, the agent scales instantly to handle real-time sales queries.

The 'magic' happens in the interaction: with Native Barge-in logic, the Sensei listens as well as it talks, allowing for a natural, interruptible flow that mimics a real CFO sitting across the desk. It’s not a chatbot—it’s an Action-Oriented Sales Partner that brings 2022-2025 data to life through sight and sound.

🏁 Technical Summary To summarize the tech stack: Sales Sensei is a cloud-native Angular application hosted on Google Cloud. It utilizes the Google GenAI SDK to facilitate an interleaved stream of text and vision.

By grounding the model in a structured 2022-2025 Knowledge Base, we eliminated hallucinations and ensured every prediction is historically accurate. We’ve solved the 'latency' problem of traditional AI by using Gemini’s low-latency Flash architecture, delivering a sub-second response time that makes live, multimodal business intelligence a reality. Thank you for your time.

🏆 Quick-Reference Card

Component Technology Contest Requirement Met Brain Gemini 1.5 Flash Leverage Gemini Model SDK @google/generative-ai Google GenAI SDK Hosting Google Firebase Google Cloud Service authentication Google Firebase Google Cloud Service Input Audio + Video Frames Multimodal Interleaved

Built With

• angular.js
• antigravity
• api
• firebase
• gemini
• live