Meet your knowledge base: a table called pets. Each row is one pet (a record), each column is one fact about it (a field). Tap a plain-English question below and watch the AI turn it into SQL β then run it right here on your table. π
Ask your data a question
Tip: tap a question chip, or try typing something like βhow many pets do I have?β
π This is called text-to-SQL β a tiny slice of what grown-ups call RAG (looking things up in stored data). The AI doesn't memorize your pets. It queries the table every time, so the answer is always current β and you can store way more than fits inside the AI's head.
Here's the big idea.
A knowledge base is just a table β rows are records, columns are fields. SQL is the language for asking the table a question (SELECT β¦ FROM β¦ WHERE β¦).
And because AI can turn your plain English into SQL, your agent can look things up instead of guessing from memory β so it remembers far more than fits in its head, and its answers stay correct. ποΈβ¨