Weekend Productivity Challenge: ROJAI – AI Listing Assistant
Inspired by my e-commerce business, I built ROJAI with Amazon Bedrock and AWS serverless services to automate marketplace listing creation.
Solving a Real Business Problem
When I received the announcement for the AWS Builder Weekend Productivity Challenge, I didn't want to build another AI demo. I wanted to build something I would genuinely use after the challenge ended.
Over the past few months, I've been building RojKilim , an online store specializing in authentic handmade kilims and artisan textiles from Eastern Turkey. As I prepared products for my store, I expected sourcing inventory, photographing products, and marketing them to be the most time-consuming parts of the business.
What surprised me most was that the biggest productivity bottleneck wasn't sourcing products or marketing them. It was creating marketplace listings.
Every product required writing an engaging title, a detailed description, SEO keywords, marketplace tags, and adapting the content for different selling platforms. Although each rug was unique, I found myself repeating the same writing process over and over again.
After creating dozens of listings, I realized I was spending more time writing product descriptions than growing my business.
Instead of accepting that repetitive workflow, I decided to build a solution.
That solution became ROJAI, an AI-powered marketplace assistant designed to help online sellers generate professional marketplace listings in seconds using Amazon Bedrock.
What is ROJAI?
ROJAI is an AI-powered productivity tool that helps e-commerce sellers create professional marketplace listings in seconds.
Instead of starting with a blank page, sellers simply provide:
- Product name
- Brand
- Product category
- Product description
- Target marketplace
- Preferred language
- Writing tone
ROJAI then uses Amazon Bedrock to generate:
- SEO-optimized product title
- Product description
- Key selling points
- Search keywords
- Product tags
The generated content can be reviewed, personalized, and published directly to marketplaces such as Shopify, Etsy, Amazon, and others.
Rather than replacing the seller's creativity, ROJAI removes repetitive writing so small business owners can spend more time sourcing products, serving customers, and growing their business.
Why ROJAI Is a Productivity Tool
Running a small business means wearing many hats. While building RojKilim, I found myself constantly switching between photographing rugs, editing product images, updating Shopify, improving SEO, answering customer questions, and managing inventory.
Creating marketplace listings became one of the most repetitive parts of that workflow.
ROJAI reduces that process from several minutes to just a few seconds by generating a professional first draft that sellers can review and customize before publishing.
Instead of repeatedly writing similar descriptions, I can focus on finding unique products, improving my store, and providing a better customer experience.
Building ROJAI
From the beginning, I wanted ROJAI to be more than a weekend prototype. My goal was to build a production-ready serverless application using AWS services while keeping the architecture simple, scalable, and cost-effective.
The frontend was built with:
- React
- TypeScript
- Vite
- AWS Amplify
The backend uses:
- Amazon API Gateway
- AWS Lambda
- Amazon Bedrock
When a seller submits product information, the frontend sends the request to Amazon API Gateway, which invokes an AWS Lambda function. The function validates the input, prepares an optimized prompt, sends it to Amazon Bedrock, and returns a marketplace-ready listing to the frontend within seconds.
To keep the infrastructure consistent and repeatable, I used AWS Cloud Development Kit (AWS CDK) to provision the AWS resources as Infrastructure as Code.
Throughout development, I also used Kiro as my AI development companion. It helped me explore architectural ideas, troubleshoot deployment issues, refactor code, and iterate faster while I remained responsible for the application's design, implementation, and testing.
AWS Architecture
ROJAI is built on a fully serverless AWS architecture.
The application uses:
- AWS Amplify for frontend hosting and continuous deployment
- Amazon API Gateway to securely expose the backend API
- AWS Lambda to process requests and communicate with Amazon Bedrock
- Amazon Bedrock to generate marketplace listings
- AWS Identity and Access Management (IAM) to manage permissions
- Amazon CloudWatch for logging and monitoring
- AWS Cloud Development Kit (AWS CDK) to deploy the infrastructure as code
Using these managed services allowed me to focus on building the application instead of managing servers.
Challenges and Lessons Learned
The most challenging part of building ROJAI wasn't writing the application itself. It was integrating multiple AWS services into a seamless workflow.
Some of the biggest challenges included:
- Configuring CORS between Amplify and API Gateway
- Managing environment variables across deployments
- Integrating Amazon Bedrock into the backend workflow
- Refining prompts to generate consistent marketplace listings
- Testing communication between multiple AWS services
Working through these challenges gave me a much deeper understanding of serverless application development, cloud architecture, and the importance of testing every component as part of a complete system.
Before building ROJAI, I assumed the hardest part would be integrating an AI model into the application.
Instead, I discovered that building a reliable AI-powered application depends just as much on architecture, deployment, security, monitoring, and user experience as it does on the model itself.
This project gave me practical experience with:
- Building a production-ready serverless application
- Deploying infrastructure using AWS CDK
- Integrating Amazon Bedrock into a real business workflow
- Troubleshooting distributed cloud applications
- Using Amazon CloudWatch for monitoring and debugging
- Improving prompt engineering for consistent AI output
Perhaps the biggest lesson was realizing that the best software ideas often come from solving problems you experience yourself.
ROJAI wasn't created simply to demonstrate generative AI. It was built to solve a repetitive task I encounter while running my own business.
What's Next for ROJAI
Although ROJAI was created during a weekend challenge, I see it as the beginning of a much larger project.
My roadmap includes:
- AI-powered product image analysis
- Marketplace-specific listing optimization
- Multi-language content generation
- Direct Shopify integration
- Batch listing generation
- Brand voice customization
- Listing quality scoring
Because I actively use RojKilim, every new feature will be driven by real-world experience rather than assumptions. My goal is to continue improving ROJAI into a practical assistant that saves small business owners time every time they publish a product.
Final Thoughts
Building RojKilim showed me that creating marketplace listings could consume more time than actually growing the business.
Building ROJAI allowed me to turn that daily frustration into a practical solution.
By combining Amazon Bedrock, AWS Lambda, Amazon API Gateway, AWS Amplify, AWS CDK, and Kiro, I built an application that automates one of the most repetitive tasks faced by online sellers.
This challenge gave me the opportunity to learn modern AWS AI services while building something that I genuinely use in my own business.
ROJAI began as a weekend challenge, but it has become a project I plan to continue developing long after the competition ends. My hope is that it will eventually help other small business owners spend less time writing listings and more time growing their businesses.
Live Demo
Application
https://develop.d1hu0f4tukm87q.amplifyapp.com/
GitHub Repository
https://github.com/arslakin/RojAI
Project Inspiration
https://rojkilim.com/
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article