Our AI-Assisted Projects
Showcasing the potential of AI-human collaboration in software development
Project Showcase
These projects represent our exploration into AI-assisted development, demonstrating the capabilities and lessons learned through practical applications.
Messaging Integration
Our first foray into AI-assisted development, focused on messaging system integration.
Technologies Used:
Timeline:
Completed in 10 hours
Project Details
This integration project involved creating a system to connect multiple messaging platforms within a Java application. It was our initial experiment with AI-assisted coding, and we approached it with minimal constraints on the AI.
The Challenge
The team needed to implement a sophisticated messaging system integration but lacked specific expertise in certain messaging technologies. We wanted to test whether AI could help bridge this knowledge gap efficiently.
Our Approach
We allowed the AI to take a leading role in this project, providing high-level requirements but giving it considerable freedom in implementation choices. We observed the AI's decision-making and intervened only when necessary.
Results
The project was completed in just 10 hours—a fraction of the time it would have taken using traditional development methods. The AI produced competent, working code that successfully integrated the messaging systems.
Key Lessons
- AI can dramatically accelerate development in unfamiliar technology areas
- The AI introduced features (like Lombok) that our team wasn't familiar with
- Without proper supervision, we risked losing control over architectural decisions
- Active human engagement is crucial - AI cannot create perfect code alone, and engineer commitment is needed to help AI create good solutions
This project set the foundation for our approach to AI-assisted development, highlighting both the enormous potential and the importance of maintaining human oversight.
Smartmark
IGCSE grading web application with AI-driven design and implementation.
Technologies Used:
Timeline:
Completed in 3 weeks (part-time)
Project Details
Smartmark is a web application designed to assist with IGCSE grading, featuring a modern interface and robust functionality. This project represented a more structured approach to AI-assisted development.
The Challenge
We needed to create a web application for IGCSE grading, but our team had limited experience in frontend development and design. We wanted to see if AI could help create a professional, user-friendly application while maintaining higher quality control than in our first experiment.
Our Approach
Based on lessons from our first project, we implemented a more supervised approach:
- Clear project requirements and boundaries were established upfront
- Regular review checkpoints were scheduled throughout development
- We deliberately slowed the pace to ensure we understood each implementation step and to avoid AI making architectural decisions
- Design decisions were explicitly approved before implementation
- We performed constant refactoring to ensure code quality
- We continuously tested new code as it was generated to catch bugs early
Results
The project was completed in 3 weeks with part-time supervision from a single engineer. The resulting application features:
- A beautiful, responsive user interface
- Intuitive grading workflows
- Client-side storage and service worker implementation
- Seamless deployment to Cloudflare Pages
- Well-structured, maintainable code
Key Lessons
This project demonstrated that with proper supervision, AI can deliver professional-quality applications even in areas where the team lacks specific expertise. The controlled approach allowed us to leverage AI's capabilities while maintaining project ownership and ensuring every aspect met our standards.
We also found that it is still not possible for non-technical people to create projects of this caliber, as debugging skills and good programming expertise are needed to effectively instruct the AI.
Smartmark represents our evolved understanding of effective human-AI collaboration in software development.
Sherlock Chrome
Chrome extension enabling AI interaction with web page content.
Technologies Used:
Timeline:
Completed in 1 day
Project Details
Sherlock Chrome is a browser extension that enables users to chat with an AI that can read and analyze the content of web pages. This project represents perhaps the most dramatic example of AI-assisted learning and development in our portfolio.
The Challenge
We wanted to create a Chrome extension that would allow users to interact with AI while browsing, with the AI able to understand and reference page content. None of our team had prior experience with browser extension development, making this an ideal test case for AI-assisted learning.
Our Approach
We applied the lessons from our previous projects:
- Clear requirements were established
- We maintained active supervision of the AI's work
- We focused on understanding the implementation as it was developed
- We asked the AI to explain unfamiliar concepts and technologies
Results
The entire extension was completed in a single day, despite our team having zero prior experience with Chrome extension development. The extension successfully:
- Integrates with the Chrome browser environment
- Extracts content from web pages
- Provides a user-friendly chat interface
- Connects with AI services to process and respond to queries
Key Lessons
This project demonstrated the power of AI as a learning accelerator. By watching the AI implement the extension and asking questions about unfamiliar concepts, our engineers gained practical knowledge of extension development in hours rather than weeks.
The experience highlighted how AI can serve as both a development tool and a teaching assistant, rapidly expanding our team's capabilities into new domains.
Our Innovation Process
How we approach AI-assisted project development
Commitment for AI-generated code success
We commit fully to the process, ensuring we don't give up early or blame AI for incompetence, but rather work collaboratively to achieve success.
Clear Requirements
We establish detailed requirements and boundaries for the project, ensuring both human and AI team members understand the goals.
Iterative Development
We work in short cycles with regular review points, maintaining continuous oversight of AI contributions while learning from its approach.
Knowledge Capture
Throughout the process, we document insights, techniques, and lessons learned to build our collective knowledge for future projects.
Interested in Our Approach?
Learn more about how we're integrating AI into software development to achieve remarkable results.