How Slalom and NFTE Used AI and Machine Learning to Mentor High School Entrepreneurs
As organizations increasingly focus on big data, machine learning and artificial intelligence, it can be difficult to identify practical applications of these new technologies. But once you cut through the noise of ongoing tech trends, you’ll find that there are countless ways in which these technologies can be leveraged for social good, in both for-profit and non-profit spaces.
In this post, I will explore how Slalom partnered with the Network for Teaching Entrepreneurship (NFTE) and used Amazon Web Services (AWS) to provide high school students with virtual coaching on their business plans. Slalom is an AWS Partner Network (APN) Premier Consulting Partner with AWS Competencies in Machine Learning, Financial Services, Education, Data & Analytics, Mobile, DevOps, Migration, Microsoft Workloads and Public Sector.
NFTE is an international nonprofit organization providing entrepreneurship training and education programs to young people from under-resourced communities to ensure their success and to create a more vibrant society. For the past five years, NFTE and Slalom Chicago have worked together to connect NFTE students with Slalom volunteers for coaching and mentoring opportunities.
In March 2019, Slalom hosted an inaugural, one-day client summit in Chicago called REALIZE. The summit featured keynotes, breakout sessions and digital experiences for guests, addressing topics ranging from customer engagement to experience design. Five of the experiences featured at the event used AWS services to bring to life Slalom’s core values, including “Build and Shape a Better Future.” This is precisely the motivation we bring to our deep partnership with NFTE.
One of NFTE’s programs focuses on equipping high school students with an entrepreneurial mindset. To do this, NFTE has developed a yearlong curriculum centered around the development of an original business plan. As part of this curriculum, many schools are able to partner with local companies like Slalom, whose employees volunteer to provide in-person coaching and mentorship for students as they work through their ideas. However, due to distance, scheduling or other factors, some schools don’t have the opportunity to work with in-person volunteers.
It was this problem that we set out to solve at REALIZE: could we create a meaningful, virtual coaching experience for both our guests and the NFTE students that would remedy this disparity in access? And in doing so, would we be able to lower the systemic barriers to entry for people to provide and receive ad hoc coaching and mentorship?
Imagine this: You walk into a phone booth. There’s a table in front of you with a tablet welcoming you and explaining that you’re about to watch a video and record feedback to help a high school student develop their business idea.
When you hit “start,” you’re shown a 30-second video of a student giving their business pitch, complete with product description, financials and competitive analysis. You are then prompted to verbally record your feedback for the student, without any typing or fumbling with an awkward tablet keyboard. With just a few taps of your finger, you’re able to go through the entire experience and create meaningful feedback that the student will receive later in the week.
How We Built It
Our solution to this problem employs serverless technologies, and revolves around how well AWS technologies work together to create a comprehensive product. We hosted our React app on Amazon Simple Storage Service (S3) and Amazon CloudFront. The 41 business pitch videos we included in the experience were uploaded to S3 and selected methodically in the app to ensure equal representation of videos. After each guest recorded their audio feedback for a pitch, we utilized Amazon Elastic Transcoder to convert web audio format to standard MP4. We then used Amazon Transcribe to convert their audio responses into text and then stored these responses in an Amazon DynamoDB table. We subsequently used AWS Lambdas to coordinate transcoding and transcription jobs and respond to requests; the Lambda triggers were Amazon API Gateway and S3 bucket notifications.
Without Elastic Transcoder and Transcribe, we would have needed our guests to type paragraphs of feedback, which would have taken longer and led to less overall feedback as well as a suboptimal user experience. Instead, our guests were able to record audio feedback, which Elastic Transcoder then converted into the appropriate format for Transcribe to consume the file.
Transcribe also allowed us to use the latest in automatic speech recognition (ASR) technology to process each audio file and give us the corresponding text file, along with the level of confidence for each word. Once Transcribe analyzed each audio file, we used custom logic based on the average level of confidence to identify which transcriptions required a quick manual check. From this output, we ended up with a CSV of all the feedback. Finally, after the event, Slalom created feedback cards to distribute to the students.
A Meaningful Coaching Experience
A custom set of Amazon CloudWatch event logs, paired with a dashboard, were used to track engagement at each experience, including custom metrics such as error rates (if any), individuals participating in coaching and more.
In the span of three hours, REALIZE guests completed 55 distinct coaching sessions, yielding over 41 minutes of feedback for the NFTE students. This feedback directly helped them further develop their business ideas by asking questions or stating concerns that could be posed by their teachers or judges, as well as by providing direct encouragement. As mentioned above, we set out to create a meaningful coaching experience that would give students greater access to entrepreneurial feedback, and that is exactly what we did.
We’ve only scratched the surface of what’s truly possible with AWS language services. Imagine the possibilities with bi-directional language translation: A Spanish-speaking student could pitch their business idea, the coach could hear it in English and give advice, and then the student would receive the feedback in their native language! AWS could also be used to generate subtitles for a video so that a coach with hearing loss could still view and provide feedback for a student’s business pitch. At Slalom, we are excited to continue working with NFTE to build on the experience we created for REALIZE and find more opportunities to use technology to serve our communities.