Skip to main content
A Data Store table holds structured data — rows and columns — an agent can look things up in, distinct from the documents in a Doc Store. It’s optional; add one only when the agent needs to query structured data.
A Data Store table with typed columns, primary key, and null-ability per field

​
Create a table

  1. In the Data Store area (or the “Create a Data Store table” step), choose to add a table.
  2. Table name — lowercase with underscores, e.g. client_management.
  3. Columns — define your fields, e.g. client_name, industry, primary_contact, arr, status. Standard columns (id, created_at, updated_at, is_deleted) are added automatically.
  4. Scope — see below.
  5. Review the SQL preview if shown, then create the table.

​
Choosing scope

Scope uses the same two axes as everything in Adopt: which clients can see the rows, and which agents can read them.
  • Global — org-wide; every workstream and agent reads it. Use for reference data that isn’t client-specific (e.g. a table of tax rates).
  • Workstream-based (row-level security) — each client sees only its own rows. The safe default for client data.
  • Workstream + Agent — per client, and only readable by specifically tagged agents. The tightest option; use when even within a client, only one agent should read the table.
If you’re unsure, choose Workstream-based. It’s the default the Quick path applies.

​
Populating it

Once created, the table can be filled via import or the API, and kept current the same way. (Verify the current import method in the app.)

​
Letting an agent use it

Creating a table doesn’t automatically let an agent read it — the agent needs the Query Data Store skill enabled and the appropriate access. See Instructions and skills and Access scoping.

​
Next steps

  1. Doc Stores and Data Stores
  2. Access scoping
  3. Schedule and publish