AWS Public Sector Blog
Building belonging in digital student communities with AI on AWS
In today’s rapidly evolving educational landscape, fostering a sense of belonging and engagement is crucial to student success and retention. An August 2024 report by the Institute for Higher Education Policy found that attending to students’ experience and degree of belonging improves academic performance and retention, especially for first-year students from historically underrepresented groups in higher education.
Improving student belonging is what Yellowdig was made for. Yellowdig is an education technology (EdTech) company that provides instructors and universities with a digital community-building platform that prioritizes high-quality student participation and relationship-building alongside learning. Powered by Amazon Web Services (AWS), Yellowdig fosters active, social, and experiential learning communities for online, hybrid, and in-person programs—both in and outside the classroom.
Using generative artificial intelligence (AI) built on AWS, Yellowdig recently launched a feature that enhances community engagement by providing instructors with core insights into digital student interactions—no matter how big their classroom size. This saves educators time, and also makes sure students get more of what they want: meaningful interactions with their instructors and one another.
Building belonging in higher education with Yellowdig
Backed by psychology and education research, Yellowdig stands apart from stale discussion boards, Q&A boards, and other peer-review tools, by focusing on building genuine human connections. While most interactive education platforms aim to facilitate dialogue among students and educators, they often miss the mark by treating engagement like a numbers game—rewarding post counts instead of real, meaningful interactions. Instructors attempt to spark discussion or debate by posting a question and requiring students to post a formal response, but this doesn’t produce natural interaction or the feeling of belonging that students need.
Rather, Yellowdig emphasizes agency and autonomy to boost real connection. Instead of prompting students to answer isolated questions, Yellowdig rewards students more for engaging with their peers than posting. “Real community requires actual interactions among people,” said Brian Verdine, senior vice president of academic engagement at Yellowdig. “Companies often refer to their social channels as ‘communities,’ but unless their customers are really interacting with each other, they’re really just an audience. The same applies to a classroom.”
Through a gamified point system, Yellowdig encourages organic conversations among students, leading to more genuine interactions. As students interact with one another, they earn more points. “Our system facilitates shifting from a mode of ‘I’m creating content to post here’ to ‘I’m coming to this community to learn with other people that I’m interacting and building relationships with,’” noted Verdine. He added that the number of comments per post in a ‘healthy community’ skyrockets as the semester progresses—a clear sign of deeper student connection and engagement.
Helping educators keep up with the conversation
The success of Yellowdig’s model led to a new challenge: how to keep faculty up-to-date on all of the conversations happening among students. In a large lecture class with several hundred students—or an online course with several thousand—it’s impossible for an instructor to read every post or comment by every single student. Some instructors feared losing track of the discussion or misbehaving commentors, and that they wouldn’t be able to rein it in.
To solve this, Yellowdig built a solution using AWS that helps educators stay informed without having to sift through thousands of interactions. With generative AI built on Amazon Bedrock, Yellowdig created the Community Recap feature. The solution generates summaries of all student conversations in a set period—in under four seconds—giving faculty a quick, real-time understanding of what has happened in the community since the last time they checked in. This lets educators jump right into the conversation—not only reassuring them that discussions are on track, but highlighting areas they can contribute to the community when it matters most.
Using AI tools to strengthen online student communities
Yellowdig has long used AWS for cloud services because the two share the same value of security being priority one. “We wanted to work with AWS because data integrity and privacy are essential in the higher education space,” said Anand Vamsi, the software engineer at Yellowdig who worked on the Community Recap solution. “Building the Community Recap feature on AWS was a no-brainer, we wanted both the data and the large language models to be in AWS infrastructure because of the security it provides,” Vamsi added.
The Community Recap feature is divided into two major parts. For the backend indexing, Yellowdig uses Amazon Kinesis, a fully managed serverless service that collects, processes, and analyzes real-time data streams, to ingest new posts and comments. The solution stores them in a vector database within Amazon Elastic Kubernetes Service (Amazon EKS). Then, this processed data is delivered to a large language model (LLM) built with Amazon Bedrock, which creates a comprehensive summary of the data and delivers one executive summary of posts back to the instructor. Not only does the summary help the instructor stay up-to-date on all of the conversations their students are having, but the model also produces an overall sentiment analysis—so educators can see the tone of their community right away and respond to any notable discussions.
Building the Community Recap feature took only four months to develop and deploy, with AWS providing consultation support and removing roadblocks as they came up. At first, Yellowdig rolled out the feature to a small selection of instructors and teaching assistants (TAs). Now, all community owners can include the summary feature in their digital classroom spaces—so instructors can reduce the tedium of monitoring hundreds of individual comments, release their worry about student misbehavior, and dive right into engaging conversations with their students.
Improving student outcomes with meaningful online communities
The Community Recap feature is relatively new, but Yellowdig is already exploring adding more capabilities to support students and instructors, like expanding the time range for which posts can be summarized, integrating more links within the summaries to specific student content, and more. Yellowdig already uses AI for other accessibility features, like auto-captioning videos, and plans to do more for things like generating image alt-text. To belong in a community, you have to be able to participate; Yellowdig is excited to use emerging AI to make the platform even more inclusive for students and faculty with different needs.
In light of the current mental health crisis among youth—and the epidemic of loneliness and isolation cited by the U.S. Surgeon General—creating environments where students feel seen, heard, and supported is more important than ever. Fostering a sense of belonging through learning communities like Yellowdig’s is a win-win for students and academic institutions. Yellowdig’s data shows that clients have seen improvements in student pass rates of 9 percent and increases in student persistence rates of 7 percent when incorporating Yellowdig.
“Belonging is the key word to what we’re seeking to do,” said Verdine. “You can only get that by interacting with people and feeling like you’re a welcome part of a community.” Using AWS, Yellowdig provides the platform for students to create these supportive, online communities that keep them on track in their learning journey.
EdTechs are using AWS to support students and educators around the world. Learn how EdTechs can optimize their business to deliver on their missions at the AWS for EdTechs hub.