Portland Community College was founded in 1961. The community college, which is located in Portland, Oregon, has events for students throughout the year. One of the biggest events of the year is the annual job fair which gets 700+ attendees. A big motivator for companies to be involved in the job fair is the access to key insights about the job fair attendees.
Organizing an event for 700+ attendees is filled with challenges. As the doors open, the day of the job fair, attendees rush all at once and back up the lines. A paper form to collect information about the attendee is required before the attendee is allowed in. This data is then aggregated multiple weeks afterwards through manual data entry to create the insights that lure big companies like Nike to participate. Another challenge is having knowledgeable staff members working the event who can answer common questions like “what companies are hiring”, “where are the bathrooms”, or “how do I best network”. Typically, PCC will need to train staff members working the event, weeks in advance, to make sure things run smoothly.
PCC needed a solution that would help streamline bottlenecks in attendee registration, automate manual data entry, answer frequently asked questions, and help create a unique experience for the job fair attendees.
PCC became interested in the idea of creating a chatbot after learning about the Natural Language Processing (NLP) capabilities that chatbots have and their ability to curate information to users through conversation.
PCC reached out to Serverless Guru to help evaluate what needed to be built and design an easy to use chatbot for the job fair. After the initial onsite consultations, it became clear that the chatbot was a perfect fit for an event driven design using serverless architecture on Amazon Web Services (AWS).
The architecture consisted of a series of API endpoints which acted as webhooks to send and receive messages from the chatbot in real time. The NLP (Natural Language Processing) was handled using an AWS service called, Amazon Lex.
As messages would be sent from the user, a lambda function would put these messages into a NoSQL database. The NoSQL database would automatically trigger another Lambda function that would analyze the message using NLP via Amazon Lex and construct a response.
The response would then be sent back to the Facebook Chatbot and the user would be notified.
PCC was able to save a lot of money and time by augmenting the serverless architecture design and chatbot development versus trying to build everything internally.
By working with Serverless Guru, PCC was able to rapidly accelerate the planning, development process, and the delivery of the chatbot in time for the job fair.
Working closely with the Serverless Guru team allowed PCC to better understand how to build serverless applications and how to design serverless architecture.
Due to the fast feedback loop between the PCC stakeholders and the Serverless Guru team, the entire project stayed agile resulting in an aligned finished product.
“We needed to build a chatbot within a short deadline to provide attendees with a unique event experience. Serverless Guru took this requirement and proposed/implemented an entirely serverless architecture on AWS including a machine learning component! The result was an easy to understand chatbot experience that even captured common phrases the attendees would ask. We would highly recommend them!”
%20(1).svg)