Inspiration
Everyone loves a great deal, but traditional discounts are often static and don't reward customers for shopping together. We wanted to rethink how merchants attract customers and how shoppers unlock savings.
That inspired CrowdRewards — Shop Together, Save Together, an AI-powered group-buying platform that turns individual purchases into collective savings.
Our vision was simple: What if customers could purchase a product at full price and automatically receive money back when enough people joined the deal?
By combining AI-powered campaign creation with PayPal payments and refunds, we created a shopping experience where merchants grow their customer base and customers benefit from the power of the crowd.
What it does
CrowdRewards enables merchants to create goal-based promotional campaigns where customers unlock discounts collectively.
Here's how it works:
- AI-powered campaign creation: Merchants describe their product, pricing, and business goals. AI helps generate compelling promotional campaigns.
- Set a crowd goal: Merchants define the number of customers needed to unlock a discount.
- Shop with PayPal: Customers join a campaign and pay the full purchase price securely through PayPal.
- Unlock collective savings: As more customers participate, the campaign progresses toward its target.
- Automatic refunds: Once the campaign reaches its goal, eligible customers receive partial refunds through PayPal, reflecting the promised discount.
For example, a coffee shop could offer a $10 product with a 50% discount if three customers participate. Everyone initially pays $10, but once the goal is reached, each eligible customer receives a $5 refund.
The result: Customers save together, merchants gain more sales, and PayPal powers the entire transaction lifecycle.
How we built it
We built CrowdRewards as a full-stack application using modern web technologies, AI, PayPal APIs, and AWS infrastructure.
Frontend
- React and Vite for a responsive customer and merchant experience.
- Campaign discovery, participation tracking, and merchant campaign management.
Backend
- Java 21 and Spring Boot for REST APIs and business logic.
- SQLite for storing campaigns, participation records, payments, and refund information.
- Campaign lifecycle management, payment tracking, and refund processing.
AI integration
- OpenAI API to assist merchants with campaign creation, promotional messaging, and campaign recommendations.
PayPal integration
- PayPal Sandbox REST APIs for creating checkout orders, capturing payments, and processing partial refunds.
- Payment and refund tracking linked to individual campaign participants.
Cloud infrastructure
- AWS EC2 running the application.
- Nginx for serving the frontend and reverse-proxying backend requests.
- HTTPS with Let's Encrypt and DuckDNS.
- Terraform for infrastructure provisioning.
- AWS Systems Manager and S3 for automated deployments directly from a local development machine.
We also implemented deployment health checks and rollback capabilities to make our demo environment more reliable.
Challenges we ran into
1. Designing a trustworthy group-buying payment model
One of our biggest challenges was deciding how to handle payments before a campaign reached its goal. We chose a full-price checkout followed by a partial refund when the target is achieved, creating a straightforward customer experience.
2. Integrating PayPal's payment lifecycle
Creating checkout orders, redirecting customers for approval, capturing payments, and issuing partial refunds required careful handling of PayPal API responses and transaction states.
3. Preventing incorrect or duplicate refunds
Refund processing must account for failed requests, repeated API calls, and customers who have already received their discounts. Designing a reliable, idempotent refund workflow was an important engineering challenge.
4. Connecting AI to real business value
We wanted AI to do more than generate generic marketing text. The challenge was making it useful for merchants by helping them design attractive campaigns with clear pricing, discount structures, and participation goals.
5. Deploying a complete application on a limited budget
We needed a working public demo without expensive cloud infrastructure. We used a lightweight EC2 architecture, SQLite, Nginx, Terraform, and automated deployment scripts to keep the solution practical and cost-conscious.
Accomplishments that we're proud of
- Built an end-to-end group-buying prototype that connects campaign creation, customer participation, PayPal checkout, and refund processing.
- Integrated AI into the merchant workflow, making campaign creation faster and more accessible.
- Designed a win-win business model where merchants can drive additional sales while customers unlock collective discounts.
- Implemented PayPal Sandbox integration for payments and partial refunds.
- Deployed CrowdRewards to AWS with HTTPS and a custom DuckDNS domain.
- Created a repeatable deployment pipeline using local builds, S3, AWS Systems Manager, health checks, and rollback support.
- Delivered a working full-stack application rather than just a concept or design mockup.
Most importantly, we're proud that CrowdRewards demonstrates a practical way to combine AI and digital payments to create a new commerce experience.
What we learned
Building CrowdRewards reinforced several important lessons.
Payments require more than a successful checkout. Refunds, transaction states, duplicate requests, and failure recovery are equally important parts of a reliable payment experience.
AI is most valuable when embedded in a real workflow. Rather than building another chatbot, we used AI to help merchants solve a specific business problem: creating effective promotional campaigns.
Simple infrastructure can support ambitious ideas. A carefully designed application running on a single EC2 instance can be enough to validate a product concept.
Great commerce experiences align incentives. Customers want savings, merchants want growth, and payment platforms benefit from increased transaction activity. CrowdRewards brings these interests together.
What's next for Crowd Reward
Our hackathon prototype is just the beginning. We envision CrowdRewards evolving into a scalable AI-powered group-commerce platform.
1. Smarter AI campaign optimization
Use historical campaign performance and conversion data to recommend optimal discount percentages, participation goals, and campaign durations.
2. Real-time crowd progress
Introduce live participation updates, countdown timers, and notifications when campaigns are close to unlocking discounts.
3. Social sharing and referral rewards
Allow customers to invite friends and accelerate campaign completion through social networks and referral incentives.
4. Merchant analytics dashboard
Provide merchants with insights into campaign performance, customer acquisition, revenue, and promotional return on investment.
5. Production-grade payment reliability
Strengthen PayPal webhook handling, idempotency, refund reconciliation, authentication, and fraud prevention before supporting real transactions.
6. Scale beyond the hackathon
Expand CrowdRewards into a multi-merchant platform with secure onboarding, configurable campaigns, and production-ready infrastructure.
Our long-term vision is to make group-powered savings a natural part of online shopping — with AI helping merchants create better offers and PayPal making the rewards possible.
Log in or sign up for Devpost to join the conversation.