Our inspiration for building this chatbot was being able to work with interesting, practical technologies and a chatbot that is front-facing and will be likely to interact with customers with the advent of the internet and automation piqued our interests.
What it does
Our chatbot is able to query input parameters from customer responses and will select the relatively appropriate response to give the customer more information about their auto-needs or direct them to more detailed sources of information such as AutoNation's company websites.
How we built it
Our Graphical User Interface was built using Java (front-end) and Our Bot's logic and the way it handles the associated training data was built using Python, PyTorch, and Natural-Language-Processing techniques.
Challenges we ran into
Some of our challenges that we ran into were: -Learning the role NLP plays with regards to programming the logic of our chat bot. -Going through pytorch documentation and applying what's needed. -Choosing which language to build our front-end that is convenient for our use-case.
Accomplishments that we're proud of
-Proud that we we're able to dynamically learn and apply our knowledge that we learned during the Hackathon such as chatbot methodologies (Amazon Lex) -Proud that we we're able to communicate relatively effectively and setup workflows despite being first-hackthon-project-submitters.
What we learned
-We learned the importance and value of setting up a team-workflow whether it be through 'github repositories' or 'google cloud repositories' and having constant communication to account for potential changes or potential challenges. -We learned the importance of picking the appropriate tech-stack in increasing our coding output.
What's next for Autobot
-Finish GUI + Bot Logic integration... -Autobot integration with Google Maps API to assess customer's nearest location and direct bot's response logic according to AutoNation nearest provider. -Autobot integration with React or other technologies to be able to create a Hybrid-Mobile App that pairs well with web-dev or mobile-app methodologies.