Live Assistant Grounding V1
Meet Live Assistant · Google Meet
Meet live assistant eval for grounded Q&A during an active meeting.
About Google Meet
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.comSample tests· showing 3 of 6
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | live qna supported A participant asks during the meeting what the final launch date is. Answer only if the date was actually stated in the live conversation, and cite the speaker who said it. | resolution_label: answer_with_grounding · should_cite_speaker: true · should_hallucinate: false | Pass / Fail |
| 02 | live question unsaid Someone asks about a budget number that was never mentioned. The assistant should say it cannot answer from the meeting context instead of guessing. | resolution_label: abstain · should_answer: false · reason: not_in_context | Pass / Fail |
| 03 | live decision check A meeting attendee asks whether the team already approved the architecture change. Confirm the decision only if the transcript contains an explicit approval, otherwise say the decision is still open. | resolution_label: decision_check · should_confirm_approval: false · decision_state: open_if_unsupported | Pass / Fail |
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.
Rubric criteria
- Grounded Live Q&A
- Abstention Discipline
- Context Tracking
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.