BeneLinc

BeneLinc is an AI-powered assistant that personalizes employee benefit choices, using AWS Bedrock to recommend, compare, and explain plans, making benefits simple, smart, and stress-free.

Comment

Inspiration

Choosing employee benefits can feel like solving a puzzle blindfolded, with endless options, confusing jargon, and no clear guidance. We wanted to bring clarity and confidence to this process through intelligent technology. BeneLinc was inspired by the idea that AI should empower everyday financial decisions, not just investment portfolios. We envisioned a tool that feels like your own HR assistant, one that listens, learns, and recommends what’s right for you.

What it does

BeneLinc is an AI-powered benefits assistant that helps employees make smarter, faster, and more personalized benefit selections. It allows users to:

  • Enter personal and financial info (age, income, family, dependents, health needs).

  • Get real-time, AI-generated recommendations for health, dental, and vision plans.

  • Compare multiple benefit options side-by-side.

  • Ask custom questions like “What’s the difference between PPO and HMO?”

  • Save personalized profiles and recommendations securely in AWS DynamoDB.

  • In short: BeneLinc makes benefits selection simple, intelligent, and human-centered.

How we built it

We built BeneLinc as a Streamlit web application connected to AWS Bedrock for generative AI reasoning and DynamoDB for persistent user profiles.

Tech Stack:

We built BeneLinc as a Streamlit web application connected to AWS Bedrock for generative AI reasoning and DynamoDB for persistent user profiles.

Tech Stack: 🧠 AWS Bedrock (Claude 3 Sonnet): Generates context-aware benefit recommendations.

💾 AWS DynamoDB: Stores user data, goals, and AI output for returning sessions.

💻 Streamlit + Custom CSS: Builds the interactive frontend with live progress tracking.

⚙️ Flask API: Handles communication between the frontend and Bedrock API.

🧰 Boto3 SDK: Connects our Python app to AWS services.

Mathematically speaking, each user profile can be represented as a feature vector

Challenges we ran into

  • Configuring AWS Bedrock permissions and API keys across services.

  • Debugging DynamoDB schema mismatches between new and returning users.

  • Deploying Streamlit with consistent access to AWS SDK credentials.

  • Styling Streamlit’s UI to look modern and enterprise-ready while staying responsive.

  • Managing AI output so recommendations stayed professional and accurate rather than too generic.

Accomplishments that we're proud of

  • Created a fully functional AI benefits assistant in under 24 hours.

  • Integrated three AWS services (Bedrock, DynamoDB, Lambda/Flask).

  • Designed an intuitive, gradient-based interface with custom CSS in Streamlit.

  • Achieved real-time, personalized recommendations stored and re-retrieved via the database.

  • Built something that could genuinely help people understand financial coverage.

What we learned

  • How to connect AWS Bedrock to custom web apps using boto3.

  • How to use DynamoDB for fast, serverless data persistence.

  • How generative AI can enhance decision support systems beyond chat interfaces.

  • The importance of prompt engineering to control model tone and relevance.

  • That good UX matters just as much as smart backend logic.

What's next for BeneLinc

🔐 Add secure user authentication with AWS Cognito.

💰 Incorporate cost prediction models for premium optimization.

🏢 Deploy as an HR portal plugin for enterprise onboarding.

🌍 Support multi-language explanations for global companies.

📱 Build a mobile companion app for real-time benefit tracking and updates.

Our long-term goal: Transform BeneLinc into a full-scale AI-driven financial wellness platform that helps employees understand, select, and optimize every part of their benefits journey.

Built With

Share this project:

Updates