The idea came to me when designing, and redesigning and redesigning the Airtable base for my dad's car business.
I have desined the schema. Then tried to actually run through the flows and see how it would work. Then I realized some flaws in the schema and redesigned it. Then I tried to run through the flows again. And again. And again.
What if an LLM were to generate plausible fake usage? It would be a great way to test the design of the schema. It would help uncover flaws faster and more efficiently.
Product - Airtable extension that generates plausible fake usage given a base.