Mirza Umer's Work | ContraWork by Mirza Umer
Mirza Umer

Mirza Umer

AI Chatbot and Automation Expert

New to Contra

Mirza is ready for their next project!

Cover image for WhatsApp AI Booking Agent -
WhatsApp AI Booking Agent - Voiceflow Developed a conversational AI receptionist for a busy private dental clinic, deployed directly onto WhatsApp using Voiceflow. The goal was to fully automate patient intake, lead qualification, and appointment scheduling to relieve an overworked front desk. What I Built: Intake & Lead Qualification: Designed a dynamic Voiceflow conversational agent that asks required screening questions (procedure type, insurance status) before allowing users to book. Real-Time Calendar Sync: Integrated the Calendly API to check the clinic's availability instantly and offer users open time slots directly within the chat. WhatsApp Deployment: Connected the agent to the WhatsApp Business API for seamless mobile access. Smart Confirmations: Set up automated booking confirmations and reminders to reduce patient no-shows. The Results: The AI agent successfully handles an average of 50–80 patient conversations per day with zero human intervention. By fully automating the standard booking flows, the clinic reduced their receptionist's administrative workload by roughly 4 hours every single day.
0
7
Cover image for Created a highly accurate Retrieval-Augmented
Created a highly accurate Retrieval-Augmented Generation (RAG) chatbot for a legal services firm. The client needed a secure way for employees to instantly query a massive internal knowledge base of 200+ complex contracts and compliance documents. What I Built: Data Processing: Used LangChain and Python to ingest, chunk, and embed hundreds of legal documents. Vector Storage: Integrated a Pinecone vector database for high-speed, semantic search and retrieval of the document chunks. Contextual AI: Layered OpenAI GPT-4 on top to synthesize the retrieved data and generate coherent, conversational responses. Anti-Hallucination & Citations: Engineered the retrieval pipeline so the chatbot explicitly cites its sources (e.g., "According to NDA_Template.pdf, Page 4") for every answer, ensuring zero hallucinations on strict legal terms. Custom UI: Deployed the entire system via a clean, secure web interface built with Streamlit. The Results: The chatbot reduced the time employees spent searching for specific clauses from an average of 15 minutes per query down to mere seconds, drastically improving team productivity and accuracy.
0
11
Cover image for Built an enterprise-grade AI assistant
Built an enterprise-grade AI assistant handling customer inquiries across Web, Facebook Messenger, WhatsApp, SMS, and Voice. Using n8n for workflow orchestration , the bot queries an internal knowledge base to generate quick quotes , books appointments directly into Google Calendar , and securely retrieves client invoices using Twilio SMS OTP verification. Voice calls are routed through an ElevenLabs AI agent for intake. Every interaction is automatically summarized and categorized into a Chatwoot ticketing system. Critical issues, such as business downtime or data loss, instantly bypass the bot to escalate to human staff. Tools used: n8n, Chatwoot, Twilio, ElevenLabs, Google Calendar, Invoice Ninja
0
27
Cover image for AI Voice Agent for Dental
AI Voice Agent for Dental Appointment Booking Built a fully automated AI voice agent named "Sarah" for a dental clinic to handle inbound appointment booking calls without any human involvement. The Challenge: Premier Dental's front desk was overwhelmed with repetitive appointment calls during peak hours, causing missed calls, long hold times, and lost bookings. The Solution: I deployed an AI voice agent on VAPI that handles inbound calls 24/7. The agent was configured to: Answer calls instantly and greet patients naturally. Collect patient name, preferred date, and appointment type. Check availability and confirm bookings in real time. Reschedule or cancel existing appointments on request. Escalate complex cases to a human receptionist. Key results visible in the dashboard: 1,240 total calls handled. 87% success rate. Average call duration of 2 minutes 34 seconds. Consistent positive sentiment across completed calls. The Result: The clinic eliminated front desk call overflow. Staff now only handle complex cases while the AI manages all routine booking calls around the clock, including evenings and weekends. Tools Used: VAPI, AI Voice Agent, Appointment Booking Automation, Custom Voice Configuration, REST API Integration.
0
38