Projects using PostgreSQL in PunjabProjects using PostgreSQL in PunjabKorely Voiceperi – AI Voice Assistant Platform
Developed Korely Voiceperi, an AI-powered voice assistant platform that enables businesses to create intelligent voice agents trained on their own business information. Built with React.js, NestJS, and PostgreSQL, the platform allows users to configure their business profile, knowledge base, and call handling preferences, allowing the AI assistant to answer customer calls with accurate, context aware responses. Implemented business onboarding, AI agent configuration, knowledge management, and real-time call workflows with a strong focus on scalability, reliability, and security. Integrated Retell AI for AI-powered voice call orchestration and ElevenLabs for high quality, natural sounding speech synthesis, delivering human-like conversational experiences and 24/7 automated customer support while reducing operational overhead. The client was losing high-value business leads due to delayed follow-ups and manual data entry across disconnected platforms (CRM, email, and advertising channels). The internal team was spending upwards of three hours a day manually copying lead data, resulting in data fragmentation and an average response time of over 12 hours.
The Tech Stack
1. n8n (Advanced Workflow Orchestration).
2. Webhooks & REST APIs (Data Ingestion).
3. PostgreSQL / Airtable (Centralized Data Warehouse).
4. Slack API (Instant Internal Notifications).
The Solution.
I engineered an autonomous, multi-stage lead processing engine using n8n. The system captures inbound leads via instant webhooks from front-end marketing platforms, normalizes and enriches the data using secondary API lookups, and instantly populates the centralized data warehouse. Simultaneously, the workflow routes the lead based on custom logic (such as budget or region) and fires an immediate, structured notification to the sales team's Slack channel, allowing for a sub-5-minute response time.
The Business Impact.
The automated pipeline eliminated manual data entry, recovering roughly 15 hours of administrative time per week for the internal team. More importantly, reducing the lead response time from 12 hours down to under 5 minutes directly improved lead-to-opportunity conversion rates by over 30%. 
