Years of modeling decisions live inside your Tableau workbooks and Power BI reports. Getting that into TextQL used to mean re-exporting to something generic first, which is exactly the part that threw the decisions away.
It's worth knowing this works differently from a connector. A connector keeps Ana supplied with live rows from the source. An attachment tells her how your team already chose to shape them — she's reading the file itself, not a line back to the system that made it.
The native formats now go in as they are:
- Tableau:
.twb,.twbx - Power BI:
.pbit,.pbix - ThoughtSpot:
.tml
Drop one into a conversation and Ana reads it where it lands. The text-based formats, .twb and .tml, she reads straight through. The packaged ones — .twbx, .pbix, .pbit — are zip containers, so she unpacks them and works through the contents herself.
The two go together. Connect the source when you want Ana in the live data; attach the file when you want her to start from the thinking already built on top of it, instead of from nothing.
