We meet atleast 40 new people every week who are product managers and business analysts whose job is to talk to users, empathize with them, understand their pains and come up with product features to solve those problems. Everybody we meet talk about the challenge of going to users and talking to them face to face being time consuming and difficult. This inspired us to build a bot that they can use for conducting user research.

What it does

It interviews users to understand the jobs that they do and pain they have in doing their job in the context of the problem that a startup or product team is trying to solve. It analyzes the response from users and create an opportunity matrix to help startups identify the opportunities to work on.

How we built it

A Ruby on Rails app is used to provide web UI and Amazon Alexa has been used to provide voice based interface for humanly touch using these a dataset of features is created from the interaction with user using AWS Lex on Alexa. Then AWS lambda is used to store the response into AWS DynamoDB and AWS Comprehend is used to produce sentiment score and entitities. The features are analyzed and clustered to generate opportunity matrix (opportunity, differentiator, dispute or discard). Entities are analyzed[Machine Learning] to decide which feature has highest opportunity score.

Challenges we ran into

Getting the semantics and emotions out of the simple sentences from users was a challenge. Building a human like responses on Alexa was a bit difficult given that there is less knowledge base available.

Accomplishments that we're proud of

We are proud of the fact that we were able to analyze the sentiments and make the bot respond more like a human to get users talking.

What we learned

We learnt a great deal about human sentiments and machine learning.

What's next for CXBot

CXBot should become part of people's life to provide feedback to service providers and product developers.

Share this project: