BuzzWomen Bot

BuzzWomen Bot

An AI enabled chatbot that simulates human-like conversations with users to do business counselling for low-income women entrepreneurs from rural areas.

BuzzWomen Bot

BuzzWomen Bot

An AI enabled chatbot that simulates human-like conversations with users to do business counselling for low-income women entrepreneurs from rural areas.

The problem BuzzWomen Bot solves

BuzzWomen bot is an AI enabled chatbot that simulates human-like conversations with users via text messages and voice recognition on chat. Its key task is to help them identify the right business ideas after assessing their passion, skills, aptitude, resource availability, need of the market and risk mapping. Buzzwomen Bot provides a user-friendly conversational counselling to local entrepreneurs. Our bot is trained using Natural Language Processing to enable quality responses to users's queries. Our Bot uses speech to converse with the user as well. For those who find it difficult to write, our bot understands the queries in user's language using speech recognition and responds in the same. It reads the texts out loud in users language for a better user experience. It enables users who are visually impaired or who find it difficult to read to access the bot with utmost feasibility. Our chatbot also supports over 100 languages including native and international languages. It overcomes linguistic barriers locally as well as globally in chat as well as voice assistance. It is integrated in whatsapp using twillio to allow remote access to the bot. This way, user can communicate whenever they want with their mobiles.

Challenges we ran into

We faced several challenges while working on this project. Firstly, we wanted our trained chatbot to work both on the web as well as on whatsapp concurrently. This was a tideous task indeed. Next, training the chatbot without any available data is tough. We had to build our own intents. Finding the most accurate translation apis and Building connection with twillio was difficult. We faced several response timeouts and bugs while doing it. In the end, we were able to resolve these issues.

Discussion