Smart collections (self-updating semantic folders)

Categories describe how you filed documents. A smart collection describes what a reader needs: "everything a new contractor needs before their first day on site", written in plain language, published once — and it stays current on its own as matching documents arrive. SharePoint sells this idea; FileDeck’s AI tier includes it.

FileDeck AI tier — Documents → Collections. Requires semantic search to be configured.

Creating one

1. Under Documents → Collections, give it a name, a plain-language query, and a maximum size (20 by default).
2. Save — the collection resolves immediately and shows how many documents it holds.
3. Put its shortcode anywhere: [filedeck_collection id="123"] (an optional layout attribute overrides the default layout).

How it stays current

The collection refreshes when you save it, once a day automatically, and on demand with the Refresh button. Matching is by meaning, not keywords — the same semantic engine as the search box.

Rendering runs through the normal library pipeline, so layouts, padlocks and access rules behave exactly as in any other library — and a visitor never sees a document their permissions wouldn’t show them anywhere else. The resolution itself also runs as an anonymous visitor, so a collection can never bake a restricted document into a public page.

One honest note on cost: a page view never calls the AI provider. Embedding calls happen only at refresh time, and only for the query (one short string), so a popular collection page costs the same as any library page.