Glean App’s New AI-Powered Community Feature

How do you turn early success into long-term growth? Glean App’s approach is to listen to users and deliver what they need. Their latest product release, an AI-powered community feature, is a great example of how they’re incorporating user feedback into their product roadmap.

With Glean’s users previously relying on external platforms like Facebook for community interaction, there was an opportunity to create a more tailored, moderated, and user-focused space within the app itself. This led Devika to design an integrated community feature that enhances functionality without sacrificing user experience.

Building a Community That Works

Through close collaboration with Glean App, we identified three key areas to focus on:

• Threaded Comments: Keeping conversations clear and organised.
• Streams: Personalized, goal-oriented sub-communities.
• AI-Powered Moderation: Ensuring safety without compromising authenticity.

Each of these features enhances usability, scales with Glean’s growing audience, and aligns with the app’s mission to provide a seamless fitness experience.

AI Moderation – Built on AWS & Amazon Bedrock

To moderate user-generated content at scale, Devika engineered a custom AI-powered moderation engine using Amazon Bedrock and other AWS services. The system automatically reviews user posts across public feeds, scores them for appropriateness using prompt-engineered LLMs, and flags content for review only if it crosses a custom threshold – ensuring balance between safety and authenticity.

Key Highlights:

• Moderation via Bedrock Models Leveraging foundation models from Amazon Bedrock, Glean’s moderation engine rates content on a 0–100 scale for potential harm or violation, allowing nuanced and context-aware moderation.
• DynamoDB Stream Handlers + AWS SDK Devika built a real-time content review system using DynamoDB streams, triggering automated reviews for every post submitted. AWS SDK integrations forward the data for LLM analysis via Bedrock.
• Human-in-the-loop Review Coaches can review and override AI decisions via a dedicated admin interface. This ensures moderation is flexible and not overly restrictive – important for a platform focused on adult users where swearing or gym selfies are often appropriate.
• Cost-Efficient Staffing Strategy Initially, Glean considered hiring full-time moderators. Instead, the AI system allowed existing Coach staff to manage flagged content, saving on new hires and keeping operational costs predictable while scaling.
• Cloud Monitoring via AWS CloudWatch Performance and moderation volumes are tracked in real time, with CloudWatch providing logging and basic alerts to ensure the system remains efficient and reliable.

1. Threaded Comments: Keeping Conversations Organised

Users frequently engage in discussions about workouts, nutrition, and progress. To improve readability and engagement, threaded comments allow for structured conversations, ensuring responses stay connected to the right topics.

The Result: Users can now reply to specific messages, tag coaches or peers, and easily follow discussions. This reduces confusion and fosters more meaningful interactions within the community.

2. Streams: Tailored Communities for Every Fitness Goal

Fitness journeys are personal, and Glean’s new Streams feature allows users to join groups based on specific interests – whether it’s weight loss, bodybuilding, or dietary needs. This customisation fosters deeper connections and more relevant discussions.

For trainers, Streams enhance efficiency by consolidating commonly asked questions and enabling broader yet personalised engagement. This feature also streamlines navigation, making the user experience smoother and more intuitive.

3. AI-Powered Moderation: Balancing Safety and Authenticity

Traditional moderation methods can be slow and inconsistent. By integrating AI, Glean ensures a safer, more inclusive community without excessive manual oversight.

How It Works:

• AI performs real-time sentiment analysis on text posts to detect and remove harmful content.
• Image and video moderation ensures inappropriate material is flagged while allowing progress photos.
• Customisable rules align with Glean’s guidelines, filtering out toxicity while maintaining a natural, engaging conversation style.

With AI handling moderation, the Glean team can focus on community engagement and growth rather than constant content policing.

A Case Study in Strategic AI Implementation

This latest update demonstrates how listening to users and strategically leveraging AI can drive engagement, improve retention, and future-proof a product. Glean’s approach – taking insights from real users and continually iterating – ensures they remain at the forefront of digital fitness innovation.

For startups looking to create scalable, AI-driven solutions, Devika’s approach to product evolution – planning, design, development, launch, and ongoing iteration, ensures sustained success.

**Your Vision, Our Expertise –**We work with a select few high-growth startups each year. Think your project has what it takes? Let’s talk. https://devika.com/contact

Results

By integrating AI moderation, Glean avoided hiring an estimated 1–2 full-time moderators, saving on two full time salaries. Moderation response time has improved from several hours to near-instantaneous flagging. Coach staff can now handle flagged content in under 60 seconds per case, improving operational efficiency while preserving community safety.

Read the blog about Glean App’s 2024 Launch.