Records, pipelines, and apps
Connector results and agent outputs become shared work when a playbook publishes selected fields into a list. A list defines records and fields. A pipeline defines the stages and transitions of a record. An app renders those records for the team and can let a person act on them.
Two paths from the same data
Section titled “Two paths from the same data”After mapping the organization’s apps to an ontology, a team can use the mapped data in either or both ways.
| Path | Reads | Result |
|---|---|---|
| Custom app | Authorized connector data or shared lists | A custom view, such as a chart of prospect counts by stage. |
| Agent | Mapped records and other permitted evidence | A declared write back to an existing app, or publication to the shared list that a custom app reads. |
A custom app can plot the current prospect stages while an agent uses a playbook to qualify a prospect. The agent can update an authorized CRM row and publish the accepted outcome to the shared list. The chart and the agent use the same meaning of prospect and stage. The connector and list bindings determine which records each path can access. Protected writes follow their own approval rules.
Follow one record
Section titled “Follow one record”- The research playbook reads evidence about Northwind from a connector and maps approved fields for publication.
- The outreach playbook sends the reviewed message and publishes the outcome. The prospect record moves to Contacted.
- Pilot prospects displays the record from the same list. With live sync, open views receive the change. A configured stage transition can start another permitted playbook.
A publish step declares which fields may leave the private run. A pipeline transition governs a change of state, while a connector operation governs an external effect. Both can affect what a person sees in the app.
Built-in views and custom apps
Section titled “Built-in views and custom apps”A list-backed grid or board is a declarative view. Its definition names the list, fields, and optional pipeline. A custom app package has its own manifest and interface. A workspace installation binds the package to permitted resources and connector accounts. A Studio list-backed view travels with the team definition. A custom app installation stays with the workspace. The package can be versioned independently of the team that installs it.
The example files show a Studio list, pipeline, and board together. See lists, pipelines, and apps, custom app packages, and live sync. list-builder, pipeline-builder, and connector-app-builder cover authoring the corresponding resources.
Example two views over one meeting list
Section titled “Example two views over one meeting list”Use the same identity in each view
Section titled “Use the same identity in each view”One team view shows meeting rows and another groups those same records by pipeline stage. A playbook maps provider results into allowed list columns. Use a stable source identifier to find an existing record, then advance the stage through an allowed transition when the relevant outcome has been accepted.
Calendar event ID → meeting record → table and pipeline boardReview outcome → allowed stage transition → refresh both viewsChoose the publication destination
Section titled “Choose the publication destination”A structured output can publish to a list and refresh a Studio view. A compiled app can read or write through its installation’s authorized bindings. The playbook and the app each use the appropriate approval path for a governed external effect.
See record publication, stage transitions, and compiled interfaces.