First prompts that match official guidance

· Use these only after Gmail or another anchor source finishes an auto-fetch cycle (~20 minutes).
· Model routing is automatic—no need to name GPT-4 vs a fast model in the prompt text.
· Rephrase placeholder names (e.g. Sarah) to match real people in your workspace.

Briefings

Pull these near-verbatim from Getting Started once memory is warm:

"What do I need to know from the last 12 hours?"
"What's waiting on me?"

Cross-source questions

"Summarize what I missed today."
"What are the key decisions from this week?"
"Extract action items from my recent conversations."
"What did Sarah say about the project across email and chat?"

These assume multiple connectors (mail + chat + docs) have completed at least one ingest pass. If the answer is thin, check Integrations & sync before troubleshooting the model.

Advanced surfaces (optional)

After basics work, explore meeting agents, native voice, and background loops described in the upstream feature docs—each adds new permissions and cost surfaces, so opt in deliberately.