I used to think AI would struggle with complex data analysis because the hardest part was actually expressing the query.
I was wrong.
It turns out AI is totally capable of navigating schemas, inferring relationships, and translating intent into analysis.
Before MCP: find the right table, understand the schema, write the query, know what to exclude.
After MCP: start with a question. The agent handles documentation search, schema exploration, and query generation behind the scenes.
Product, compliance, research, strategy. No SQL required.