How does iDBQuery decide which source or table to use for a question?

When you've connected multiple databases, spreadsheets, and documents, someone has to decide where an answer should come from. In iDBQuery, that's automatic.

• A unified map. Every connected source is introspected into one live queryable model, so iDBQuery knows what data exists and where.
• Intent-based routing. For each question it matches your meaning to the relevant tables and sources — customer records here, transactions there, a contract in a PDF — guided by the semantic layer and glossary.
• Cross-source when needed. If the answer requires several sources, it queries each and joins the results, rather than making you export and stitch data yourself.

This is the practical payoff of federation: you ask a business question without knowing or caring which system holds which field. "Compare pipeline in the CRM to invoiced revenue in the finance database" just works. And because iDBQuery names the sources it drew from and cites figures back to their exact rows, you can always see that it reached into the correct systems — so the convenience of not tracking where data lives never costs you the ability to verify it did.