Grow scale

Inspiration

I recently purchased an item from an e-commerce website. But when I looked at the packet which was delivered to my home. it doesn't have the one which I've ordered. It got replaced with something else. So, I went to their Facebook page and commented on my issue there under a post. Then after I looked at some more comments over there. I could see many complaints. Then it just struck my mind that how the business owners can have a look at all these comments. If they can't then how are they gonna improve their business without reaching out to customer problems.

What it does

Our idea innovates the way how business owners can reach their customers. We made use of some recent advancements in Natural language processing which is a chatbot. Now a user who comments their problem on the page will be redirected to the messager with the help of a feature(private replies). Then the chatbot which we had trained and deployed in the messenger interacts with the user and tries to know his issue by providing the top issues faced by many customers in the form of quick replies. Each of these replies is the intent we took from the already available customer data on an Indian e-commerce website(Myntra). Now, these are some of the intents chosen from the problems faced by the customers majorly. If he doesn't find his issue in the list of quick replies then he needs to describe his problem with a short paragraph. This short paragraph is sent to a page that has a trained Wit.ai model running in the background which can find out the intent in the paragraph. Now the intent is stored into the database in its appropriate intent section. To analyze all the data stored in the database and know what's the most common problem faced by the customers, we've built a website(which looks like a Facebook page) with a histogram plot. This histogram is dynamically updated by fetching the data stored in the database. By looking at the plot business owners can solve the issues which customers face the most

How we built it

For the first task to redirect the customer from the business page comments section to his inbox we made use of a website called PostJelly which can send automated replies for negative comments posted by the customers. It also sends a private reply to the customer's inbox redirecting them from the business page to the messenger. Now from there chatbot takes over the control. we've designed the chatbot with the help of Gupshup.io just for the quick replies part. If the user chooses to describe his problem with a paragraph then in background the paragraph is sent to a page which has a wit.ai model trained in background finds out the intent and will be sent to the database.

Challenges we ran into

Accomplishments that we're proud of

We are proud of the idea we've developed which can impact the business on a large scale. This application helps the business owner to find out the areas where his/her customers' facing problems in his business faster, easier and also to reach his customer's problems within no time and no pain

What we learned

In developing up an application for this hackathon, we got good exposure with some tools like Wit.ai, Gupshup. We've also learned how to manage work among the team members and make things work out well.

What's next for Grow scale

We would like to add some more improvements in the future to make this application much more interactive to the business owners. If we can help find not only the most common problem faced by the customers but also from which section that particular problem is mostly pointed out by the customers then it could be more helpful. To exemplify, the histogram only shows the most general problems like whether it's a refund issue or product quality issue or it's from the customer service end. But what if we take one step ahead to know the exact brand which people are facing the issue when we talk in terms of product quality. With this kind of smart analysis, business owners can have a clear idea about everything and improve their customer's trustworthiness which in return produce better revenues