Cally-IO: AI-Powered Conversational Sales Platform

Inspiration

We recognized that modern sales teams struggle with inefficient workflows, scattered customer data, and the inability to leverage AI effectively. Traditional CRMs are often complex and don't integrate well with AI capabilities. We built Cally-IO to bridge this gap - combining the power of conversational AI with practical sales automation in an intuitive interface.

What it does

Cally-IO is an intelligent sales co-pilot that:

• 🤖 Answers questions using your company knowledge base and real-time web search
• 📞 Automates outbound calls with AI-generated personalized scripts
• 📊 Visualizes sales pipelines with a Kanban-style lead management system
• 📈 Provides real-time analytics on team performance and content effectiveness
• 🔍 Tracks unanswered questions to identify knowledge gaps

How we built it

Tech Stack

• Framework: Next.js (App Router)
• UI: ShadCN UI + Tailwind CSS
• AI: Google Gemini via Genkit
• Backend: Appwrite (Auth, DB, Storage)
• Calling: Twilio Voice API
• Search: Tavily AI

Key Features

• Role-based access control (Admin/Agent)
• Persistent chat history with multi-modal support
• Dynamic call script generation
• Automated calling with speech recognition
• Real-time analytics dashboards
• CRM webhook integrations

Challenges we ran into

1. Real-time voice processing: Implementing smooth speech-to-text during calls required significant Twilio configuration
2. RAG accuracy: Tuning the retrieval system to provide relevant document excerpts took multiple iterations
3. State management: Coordinating chat history, lead data, and analytics across components was complex
4. Permission systems: Implementing granular role-based access without compromising performance

Accomplishments that we're proud of

✅ Creating a fully functional AI calling system that handles natural conversations
✅ Building an intuitive UI that makes complex AI features accessible
✅ Achieving <500ms response times for RAG queries
✅ Implementing comprehensive analytics that provide actionable insights
✅ Partnering with leading AI infrastructure providers

What we learned

• The importance of streaming responses for conversational UX
• How to effectively monitor and improve LLM output quality
• Best practices for secure AI agent deployment
• The value of structured data for analytics vs unstructured chat data
• How to balance automation with human oversight in sales workflows

What's next for Cally.io

🚀 Multi-channel expansion: Adding email and SMS automation
🔌 More integrations: HubSpot, Salesforce, and Zapier connectors
🌎 Multilingual support: For global sales teams
🛠️ Custom AI workflows: Visual builder for sales processes
📱 Mobile app: Dedicated iOS/Android applications

Built With

• appwrite-(auth/db)
• google-gemini-ai
• next.js-14
• shadcn-ui
• tailwind-css
• tavily-search
• twilio