Macy's hires over 90,000 seasonal employees every holiday season to deal with demands. They sometimes have to interview up to 10 candidates for each position they need to fill. This takes up huge amounts of time that the hiring managers have to spend interviewing potentially unqualified candidates.
What it does
Our chatbot, Stella, talks to each candidate and extracts important information from their conversation in order to assign them a score which tells the hiring manager how likely the candidate is to do well in the role and how strongly Stella recommends proceeding with further interviews with the candidate. The score is calculated based on how well the candidate performed on a series of logic and math puzzles in addition to their communication skills and overall tone throughout the conversation.
How we built it
Our chatbot is built on Microsoft Azure's BotBuilder platform running in Node.js. It also uses Microsoft Cognitive for tone analysis. Our website is HTML/CSS/JS hosted on Azure's webapp platform.
Challenges we ran into
Setting up the database to get the bot responses into the hiring dashboard was challenging as the Azure Node.js framework did not support the way we wanted to get information from the bot. Setting up the website was a challenge since the Node.js web app that we originally set up did not support the bot.