🚀 GroMoPulse – Your AI Growth Companion

🌟 About the Project

GroMoPulse is a hyper-personalized AI-powered sales companion built specifically for GroMo Partners (GPs). It transforms the traditional sales enablement experience by providing real-time, multilingual, voice-based coaching, smart lead generation, and adaptive growth tracking — all tailored to the unique journey of each individual GP.

💡 What Inspired Us

While studying the diverse challenges faced by field agents and micro-entrepreneurs in Tier 2/3 cities, we discovered that most digital tools are one-size-fits-all dashboards that ignore regional language needs, emotional support, or personalized growth paths.

We realized that GPs don’t just need a platform — they need a companion who can:

Understand them,
Talk to them in their language,
Predict their struggles,
And guide them like a coach.

That’s when GroMoPulse was born — not as a dashboard, but as an AI Growth Companion.

🛠️ How We Built It

We structured the platform into modular AI-driven components:

1. AI GP Persona Engine (AGPE)

Built dynamic profiles using past activity, behavior, and sales data.
Continuously evolved recommendations using real-time event triggers.

2. Conversational Co-Pilot with Multilingual Voice AI

Integrated Whisper for speech-to-text and HuggingFace models for text generation.
Added support for Hindi, Tamil, and Bengali to begin with.
Simulated voice roleplay and live call feedback sessions.

3. AutoLeadAI – Contextual Lead Engine

Used spatial-temporal models with GPS and local event APIs to suggest high-potential leads.
Auto-generated WhatsApp and SMS templates in regional languages.

4. MicroCoach

Replaced long boring training with 30-second capsules triggered by contextual cues.
Created gamified nudges to improve user engagement (XP, badges, rewards).

5. PulseScore – Predictive Growth Tracker

Built a simple predictive model using XGBoost that estimated income potential and stagnation risk.
Displayed real-time growth tips based on local trends and peer performance.

Tech Stack:

Voice AI: OpenAI Whisper + HuggingFace + Google TTS
AI/NLP: OpenAI API + Pinecone vector DB
Frontend: React Native
Backend: FastAPI + AWS Lambda + Firebase for push notifications
Data: Integrated local weather & festival APIs for lead prediction context

🤯 Challenges We Faced

Multilingual Voice Understanding:
- Handling accents and dialects in real-time was tricky.
- We had to fine-tune models for regional language accuracy.
Dynamic Personalization Without Overwhelm:
- It was a challenge to deliver useful AI insights without bombarding the user.
- We used a progressive disclosure approach — only showing relevant data when it was most needed.
Balancing AI Suggestions vs. GP Autonomy:
- Striking a balance between guiding and overwhelming required user testing.
Offline and Low-Connectivity Optimization:
- We compressed models and preloaded critical modules to make the app usable in poor network zones.

📚 What We Learned

Voice-first interfaces are not just a convenience — they are a necessity in Bharat.
Personalization is powerful, but it must be context-aware and emotionally intelligent.
GPs value growth — not just earnings, but confidence, skill, and community status.
Modular AI design lets you scale from MVP to enterprise without rearchitecting the system.