For this 6th edition of hack week in iZettle, we counted with several high quality contributions from different teams throughout tech, risk and product.
This time the hackathon theme followed the motto “ship it”, where the main goal was to develop a new application or functionality that should be running in production, ready to be used, by the end of the week.
For this edition, the chosen winning team proposed and showed a proof of concept chatbot under the name Emma running on iZettle’s internal slack channel.
The name Emma, given by their team members; stands for Elegant Machine-Manufactured Answers.
The idea to make Emma was to assemble an iZettle bot assistant inspired by similar efforts made by companies such as Google (for example their future Google assistant chatbot). The assistant would interact with the user, in order to answer their questions and give back useful information based on their needs in real time.
iZettle’s chatbot Emma, was built from the ground up using open source tools and frameworks based on Java. The AI engine is based on an AIML 2.0 interpreter which is publicly available, which allowed the team to assemble the reasoning and induction mechanisms for Emma. Domain knowledge base was built based on iZettle’s internal employee references and FAQ for their product base, in addition part of its humanity was generated by using publicly available AIML files.
In the case of iZettle, our use cases focused on 2 scenarios were the bot will provide information to the person speaking with her.
In one scenario, we would show how a normal user will interact with Emma in order to get support questions answered regarding our products and applications. Emma would reply interactively with the user; offering options and answers to their inquiries.
The other scenario focused on showing how Emma could be used as a knowledge and information base user agent, were employees could ask her questions regarding other employees, departments and other information of interest for internal use. Some examples, included how Emma was able to distinguish and reason what person you were looking for in cases where people shared the same name.
Does this sound like fun? Then join us, we are hiring.