Skip to main content

AI Usage > Client

Written by Tom Williams

Dataset: AI Usage

Entity: AI Usage Measurement

Field ID: client_name

Type: Text

Description: The display name of the tool through which the usage happened (e.g. IDE or agent app).

Note: Populated on CLIENT breakdown records and on CLI, DESKTOP_APP and AUTOMATION surface type records, and empty elsewhere because a generic surface type record can span several clients.

Source: App

Transformation logic: N/A

App Mapping

GitHub Copilot

the IDE name on CLIENT records; terminal for the command line, desktop for the desktop app, the agent name for agent apps

Anthropic Claude (coming soon)

Calculated: normalised from terminal_type

Cursor (coming soon)

cursor

OpenAI Platform (coming soon)

N/A

OpenAI Codex (coming soon)

Calculated: normalised from client_id

Reporting Use Cases

The Client field says which application the usage came through, as opposed to which kind of surface it was.

  • Tool inventory: Group by client to see which editors and agent applications are actually in use across the team.

  • Rollout tracking: Combine with created_at to watch adoption of a newly approved client.

  • Group on the ID, not the name: Providers rename agent applications. Use client_id as the grouping key and this field for display.

Did this answer your question?