Custom fields accumulate. A CSV import three months ago created Email_2, an enrichment writes kdm_contact, and somewhere along the way you decided the one you actually use should be called work_email. Until now there was no honest way to tidy that up. Renaming a field left every stored value stranded under the old name — invisible on the lead page, missing from your templates, gone. Deleting one hid the values without removing them. And the lead page showed whatever it showed, with no way to tell a field your workspace knows about from a leftover key nobody meant to keep.

The lead page now tells you which is which. Custom Fields lists the fields your workspace has defined, each with a proper name instead of the raw key and a small (i) you can click for the description you wrote — the same description the AI reads when it personalises a message. Anything else stored on that lead appears below under Other fields, so a stray Email_2 is visible rather than silently riding along. While editing, each row has a bin icon that removes the field from that lead for good.

Renaming a field now moves the data with it. Pick Rename, type the new key, and every lead's value is remapped in the background — you get a progress bar, not a spinner. Before it starts, we show you every filter, view, automation and sequence step that still mentions the old name, because those keep pointing at it and will stop working until you update them. If that list is longer than you expected, cancel; nothing has happened yet.

Delete now means two different things, and both are offered. Delete removes the definition and keeps the values on your leads, where they show up under Other fields — the right choice when you are reorganising and want the data back later. Delete and Purge also strips the key from every lead. That one is irreversible, and it says so.

The Lead Custom Fields page itself has caught up with the rest of the app: search, sortable columns, the same column-header filter menus as the lead list, and multi-select for turning "Include in AI context" on or off across several fields at once or changing their type in one go. Alternative keys — the column names a CSV import should map into a field — are editable there now too, rather than only through the API.

All of it is in the API: rename, purge, references, bulk-toggle and bulk-set-type on context-fields/, with the rename and purge returning a job you can poll.