At least 95% of the challenge in querying data doesn't come from the difficulty of writing SQL statements IMO, but from the complexity/brokenness of the data itself. Especially with the move to unstructured data streams, most data warehouses seem to have become pretty bad when it comes to extracting truth out of the information within. Many business users know some SQL but blanch at creating coherent reports from what's available to them. What BI people are paid for seems to be having knowledge about the problems particular to a domain.
If your data is relatively clean and follows a good model, this would be a great way to help someone join, group, and aggregate their data without knowing SQL. I think a lot of people would use Tableau for this if it didn't also have a steep learning curve / high expense. However as soon as you get into data where you're writing case statements, coalescing null fields, matching on different data types, decoding, partitioning over streaming data etc., it doesn't help someone without knowledge of the caveats within the data sources themselves. Show me someone who doesn't know SQL using this to produce insight out of compromised data and I will be impressed.