Operating data modeling & product attributes

With write access, a connected assistant can build and maintain your custom tables in Data Modeling, and manage your Product & variant attribute definitions and the values stored on products and variants. This page explains what each covers and the plan requirements that apply.

Data modeling

The assistant can do most of what the Data Modeling canvas does, short of arranging the canvas itself:

Create a custom table — give it a name and description. Logistified automatically makes it available as a linkable attribute on products or variants.
Rename, redescribe, or deactivate a custom table.
Add, edit, or remove columns on a table (or on the built-in variant table) — set the column’s type, and optionally give it a constraint role so purchase-order constraints can use it.
Fill in row values — one row and column at a time, or in bulk: the same value can be applied to every row a filter or a list of products/variants matches in a single request.
Add or delete rows in a custom table.
Connect tables together — create or edit a relation between two tables (for example, linking a “Packaging profile” table to your products).

Product & variant attributes

The assistant can manage the simpler, single-field attributes configured under Settings -> Products -> Attribute Definitions:

List and read attribute definitions and their current values across products and variants, including a full spreadsheet-style read of the attributes grid.
Create, rename, or delete an attribute definition — text, number, money, date, yes/no, list, or a link to a Data Modeling table row.
Edit attribute values — on one product or variant, or across many at once (for example, “set case size to 12 for every variant in this collection”).

Plan requirements

Capability	Plan required
Data modeling — read tables, rows, and values	Any plan
Data modeling — create/edit tables, columns, relations, rows, and values	Elevate
Product & variant attributes — read definitions and values	Elevate
Product & variant attributes — create/edit/delete definitions and values	Elevate

Guardrails

Managed attributes stay managed. A table created in Data Modeling automatically gets a matching attribute so products/variants can point to one of its rows. That attribute can’t be renamed or deleted from the attributes side — the assistant is told to change the table itself in Data Modeling instead.
One active field per role. Only one active attribute or Data Modeling column can carry a given constraint role (Product MOQ, Unit CBM, and so on) at a time. Assigning a role that’s already taken is refused.
Deleting a column can be soft or hard. By default, removing a Data Modeling column just deactivates it and keeps the values already stored; the assistant can also ask for a full delete, which removes the column and every value stored under it.
Concurrent edits are caught, not silently overwritten. If someone is editing the same table or the same attribute definitions in the app at the same time, the assistant’s save is rejected once and automatically retried against the latest version; a second collision is reported so you can check what changed.
Deactivating a table doesn’t hide its attribute automatically. If you deactivate a custom table, the attribute it published stays visible in Product Settings until you hide it there yourself.