Before
They had an idea for civic transparency but no ready system to hold it. Candidate records span local, state and federal offices, and statements and votes live in different places. Backend schema design, JSON logic and integration were things they still had to learn.
“What did a candidate say, and what did they do?”
In their own words
The turning point
In the program they treated AI as a technical learning partner rather than a prompt box. ChatGPT helped them understand relational schema principles, clarify JSON concepts and troubleshoot structural issues. They decided the data had to come first, and built in layers.
“This project represents applied AI-assisted system design rather than simple prompting.”
In their own words
After
They architected an Airtable backend with linked tables for candidates, offices, campaign statements, voting records and an issue linkage layer. On top of it they built a Lovable prototype with dummy data where you can search a candidate and see a visible Consistency Score. A roadmap for automated data ingestion and syncing, including n8n, is now in progress.
What it lets people do
The platform lets people look up a candidate and see, in a neutral format, where campaign statements align with or diverge from recorded votes. It avoids interpretive fields and focuses on structured historical data, so the record speaks for itself.
What they learned
They learned that data architecture must precede automation, and that neutral systems require structural discipline. They also saw how much AI can speed up the learning curve of backend design, and how prototyping with dummy data allows safe validation before going live.
“Data architecture must precede automation.”
In their own words
Said & Done: Civic Consistency Platform was built during AI Mastery.
Read the full submission
The Automation Architect
Said & Done: Civic Consistency Platform – Data Architecture & Automation Design
Achievements
• Architected a relational backend data model in Airtable to support multi-level civic candidate records (local, state, federal).
• Designed structured data relationships between candidates, offices, campaign statements, and voting records.
• Built a working Lovable front-end prototype connected to structured dummy data demonstrating candidate lookup and consistency logic.
• Developed a non-interpretive scoring framework comparing stated positions with recorded votes.
• Created a scalable architecture roadmap for automated data ingestion and backend-to-frontend syncing.
• Applied AI-assisted schema design and troubleshooting to accelerate backend architecture learning and execution.
Tools
• Airtable (relational backend database architecture)
• Lovable (front-end prototype interface)
• ChatGPT (schema modeling, JSON logic guidance, integration tutoring, structured troubleshooting)
• Manual dummy data population for prototype validation
Methodology
Set & Done was architected as a data-first civic infrastructure prototype.
The objective was to design a system capable of answering a simple but structurally complex question:
What did a candidate say, and what did they do?
To achieve this, the architecture was developed in layers.
1. Relational Data Modeling (Airtable)
The backend schema was designed to support:
• Candidate Table
• Office Table (local, state, federal distinctions)
• Campaign Statements Table
• Voting Records Table
• Issue Linkage Layer (for mapping statements to votes)
This structure allows separation of entities while maintaining relational integrity.
The design intentionally avoids interpretive fields and instead focuses on structured historical data.
2. Consistency Logic Framework
A structured scoring model was designed to:
• Compare campaign statements to recorded votes
• Flag alignment or divergence
• Present results in a neutral, non-editorial format
The goal was not political analysis but data transparency.
3. Front-End Prototype (Lovable)
A functional interface was built in Lovable using dummy data to:
• Enable candidate search functionality
• Display structured profile information
• Present a visible “Consistency Score”
• Validate backend schema usability
Search example: “Jane”
This allowed rapid proof-of-concept testing before automation.
4. AI-Assisted Architecture Development
AI (ChatGPT) functioned as a technical learning partner during development.
It was used to:
• Understand relational schema design principles
• Clarify JSON formatting concepts
• Explore backend-to-frontend integration logic
• Troubleshoot structural build issues
• Conceptualize automation workflows
This project represents applied AI-assisted system design rather than simple prompting.
5. Automation Roadmap (In Progress)
The next development phases include:
• Automated candidate and vote data ingestion
• Backend syncing between Airtable and Lovable
• Workflow automation exploration (n8n)
• Reduction of manual data population
• API-level integration for scalable civic data retrieval
While automation is not yet fully implemented, the system has been architected to support it.
Learnings
• Data architecture must precede automation.
• Relational integrity is critical in civic transparency tools.
• Neutral systems require structural discipline.
• AI can significantly accelerate backend design learning curves.
• Prototyping with dummy data enables safe validation before live integration.
• Automation design is iterative and requires staged implementation.
This project represents applied architectural thinking in civic technology, with a clear path toward scalable automation.
Relevant URLs
• Lovable Prototype (View Only – Search “Jane” to test example profile)
• Airtable Backend Schema (View Only)
• Project Walkthrough Video (View Only – Demonstrating architecture and logic)
Written 15 February 2026· Submitted by the storyteller
Every story is written by a Mindvalley student or by someone our teachers have worked with. How we collect and check stories
