Enterprise Conversational AI Platform
Chat AI, voice AI, documents and payments — one automated conversation flow, wired into live enterprise systems.
Context
As a software developer on an enterprise team, Farhaan worked on a conversational platform handling customer interactions at scale on WhatsApp — in a production environment built on .NET Core, Angular and containerized microservices, with Keycloak identity and Graylog observability.
Problem
Conversations, documents and payments lived in separate manual workflows, and none of it talked to the core business systems. The hard part was not the chatbot — it was making AI work inside a legacy enterprise landscape, over REST and SOAP, without breaking what already ran.
Solution
Farhaan contributed across the AI surface of the platform: WhatsApp chatbots built on Google Vertex AI and Python (PyTorch, Hugging Face), OCR and image analysis, voice AI with AWS Call Analytics and Transcribe, and payment integrations directly inside conversation flows. The signature piece: a real-time WhatsApp system — live chat over WebSocket — integrated into a decades-old legacy application, embedded browser and all, alongside REST/SOAP integration layers into the existing enterprise systems.
Outcome
The platform brought chat, voice analysis, documents and payments into one automated flow on the channel customers actually use.