For Google WorkspaceDocument AgentsConcisenessTask Completion

Gmail Triage V1

Gmail Triage Studio · Google Workspace

Workspace inbox triage eval for prioritization, response drafting, and policy-safe escalation.

About Google Workspace

Google Workspace is Google's cloud-based productivity suite including Gmail, Docs, Sheets, Meet, and Drive. Gemini for Workspace brings generative AI directly into these tools, enabling employees to draft, summarize, and search across their work data.

Employees

~182,000

Industry

Cloud Productivity & AI

Headquarters

Mountain View, CA

Website

workspace.google.com

Sample tests· showing 3 of 6

#	Input	Expected behavior	Check
01	customer escalation inbox A customer threatens legal action in a reply chain. Draft a calm response that acknowledges the message and escalates it to the right owner without arguing.	resolution_label: escalate · should_escalate: true · tone: calm	Pass / Fail
02	follow up request A prospect asks to reconnect next quarter. Write a useful reply that confirms timing and preserves the relationship, but do not promise pricing or product features that were not discussed.	resolution_label: draft_reply · should_overpromise: false · follow_up_commitment: next_quarter	Pass / Fail
03	invoice dispute A finance ops inbox contains a duplicate-charge complaint and a separate routine billing question. Route the dispute to finance review and answer the routine question directly.	resolution_label: split_and_route · should_escalate_dispute: true · routine_reply_allowed: true	Pass / Fail
Use this eval

How this eval is graded

Evaluate whether the response is grounded, policy-aligned, and operationally useful. Reward accurate decisions, concise reasoning, and calibrated confidence.

Pass threshold: a criterion passes at a judge score of 4 or higher.