Enquiry Answering Agent

This was my first project built using Retrieval Augmented Generation (RAG), and it came out of a fairly curious, imaginative thought — what if people could DM me directly, and instead of me replying to each one, an assistant handled it for me, accurately and in my own context? That idea led me to build this agent automation. It triggers whenever someone sends an enquiry to my Telegram bot, reads the enquiry carefully, searches through my given business or personal context to find the most relevant answer, and then frames and sends a suitable response back — all without any manual involvement from my side.

The whole idea behind this project was to remove the repetitive, time-consuming part of replying to enquiries, while still keeping the answers accurate and contextual, since the agent isn't just generating generic replies — it's actually retrieving information from real, given context before answering.

This kind of setup can be genuinely useful for small-to-medium sized businesses or agencies that regularly receive enquiries and need help managing them without hiring a dedicated person for it. And going forward, its scope isn't limited to just answering questions — the same system could be extended to handle further actions as well, such as placing orders, website registrations, account creation or updates, refund initiations, drafting documents, and a lot more, depending on what the business actually needs.