Most location spreadsheets that have been in use for a few years end up with one column holding everything. A cell reads "Kitchen, Modern, Pool, Victorian, Fireplace", and those five words belong to four different fields.
A plain import puts that whole string in one place, where it is useless for filtering. Mixed tags analysis reads the cell and files each word where it belongs.
What goes where
Each tag is sorted into one of four fields:
- Type: house, office, warehouse, restaurant
- Rooms: kitchen, bedroom, lobby, conference room
- Features: pool, fireplace, balcony, parking
- Architectural style: Victorian, modern, art deco, craftsman
So the example above splits into a room, two styles that need a decision between them, and a feature. Which is the point: the split is what makes "show me houses with a pool" a filter rather than a text search.
Turning it on
- Start an import from Locations, then Batch Import Locations
- In the Smart Mapping step, switch on Assign mixed tags column with AI
- Point it at the column that holds the mixed content, often named something like "Content/Type" or "Tags"
- Check the preview, which shows the tags already distributed on the sample cards
- Import
Where it earns its keep, and where it does not
This is aimed at Airtable and legacy sheets, where several attributes were combined in one field, the column names follow nobody's convention, and there are too many rows to sort by hand.
It is worth being clear about what it is doing. The model reads short, contextless words and guesses, with a confidence score attached to each guess. It is dependable where the vocabulary is concrete: kitchen is a room, pool is a feature, warehouse is a type. It is weaker on words that could be two things at once, and on style labels generally, where two scouts would also disagree.
Which means the confidence scores are the useful part. Skim the low ones in the preview and fix them there. Anything you correct in the mapping step stays corrected for the whole file, rather than needing fixing row by row afterwards.