Verified against Gemini in Gmail · 2026-08-04
Walk into a meeting already briefed, using only what's actually in your Workspace
A prompt that chains Gemini's @-mentions across Gmail and Drive to assemble a meeting brief grounded in real threads and files, instead of generic prep questions that could apply to any meeting regardless of its actual history.
The prompt
Ready to copy — highlighted parts are example details you can swap.
I have a meeting about the Q4 vendor contract renewal with the vendor's account manager and my VP of Ops coming up on this Thursday. Pull in and read through: @Gmail the thread with the vendor about "renewal terms" from the last two months @Drive the "Vendor Contract Draft v3" doc and the pricing comparison sheet Build a one-page brief: 1. Where things left off — the last concrete decision, question, or commitment from the most recent thread or doc, with its date, so I know exactly how current this is. 2. Anything promised to the vendor's account manager and my VP of Ops that hasn't been delivered yet, or vice versa, based only on what's actually in the threads and docs above — not on an assumption about what a relationship at this stage would typically involve. 3. Open questions that are still unanswered across everything you read. Don't invent a resolution if you didn't actually find one, even if the thread seems to be trending toward an answer. 4. 4 sharp questions I should ask in this meeting, each one tied to a specific gap you found above — not a generic meeting question that any prep-for-a-meeting prompt would produce regardless of the actual situation. If you can't find enough in the sources above to answer part of this, say plainly which part is missing rather than filling it with a plausible guess dressed up as a finding. IF GMAIL AND DRIVE DISAGREE If the most recent Gmail thread says something different from what the current version of the Drive document says (a price mentioned in an email that doesn't match the number in the pricing sheet), flag that discrepancy explicitly as its own item — don't quietly trust whichever source you read second, and don't pick the one that seems more "official" without saying you had to choose. TIME SENSITIVITY OF WHAT YOU FOUND Note how recent the most current relevant thread or file update actually is. A brief built from a document last touched two months ago, with no more recent communication on the topic, should say so — the absence of recent activity is itself a fact worth surfacing, since it might mean the topic has gone quiet for a reason worth asking about in the meeting. KEEP IT ONE PAGE Resist the pull to include everything you found just because you found it. If a piece of context doesn't change what I should say or ask in this specific meeting, leave it out rather than padding the brief with background that's technically true but not actionable here. TONE OF THE RELATIONSHIP If the tone across the retrieved threads has shifted noticeably over time (increasingly formal, increasingly terse, a warmth that's cooled), note that in one line — it's context I'd want walking in, even though it's not a decision or a commitment in the usual sense.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Chaining @Gmail and @Drive mentions in one prompt grounds the brief in content you actually have — the real thread history and the real draft document — rather than the model falling back on generic meeting-prep boilerplate like 'ask about timeline and budget,' which it will produce fluently and confidently even with zero relevant context to actually back it up. The instruction to name what's missing rather than fill a gap with a plausible guess addresses a specific risk of workspace-grounded assistants: partial retrieval — a thread that didn't fully load, a doc it couldn't open — can look identical to complete retrieval in the output unless the model is told to distinguish 'I found nothing on this' from 'here's what I found,' and those two states produce very different confidence someone should actually place in the brief. Tying each prep question to a specific gap, rather than asking for prep questions in the abstract, is what keeps the output from being four questions any prep-for-a-meeting prompt would produce regardless of the actual situation — a sharp question has to reference something the sources actually revealed as unresolved, which a generic template question structurally cannot do. Requiring the 'last concrete decision' to come with its own date matters because a Gmail thread and a Drive doc can easily be out of sync with each other — the doc might reflect an older draft than the most recent email exchange, and stamping the date on the most current finding tells you which source to trust if the two disagree. The explicit instruction to flag it, rather than silently resolve it, when Gmail and Drive actually contradict each other addresses the same underlying risk directly: a workspace assistant reading two systems that were never guaranteed to stay in sync has no principled way to know which one is stale on its own, so presenting the conflict rather than picking a side keeps that judgment where it belongs — with the person walking into the meeting who has context the retrieved sources don't. Noting how recent the most current activity actually is turns silence into a signal rather than an absence: a topic with no communication in two months might mean it's settled, stalled, or simply forgotten, and any of those is more useful to walk in knowing than a brief that reads identically whether the last update was yesterday or last quarter.
Verified against
Gemini in Gmail Gemini 3 Pro · 2026-08-04
Changelog
- 2026-08-04 — Initial publish, verified against Gemini in Gmail on Gemini 3 Pro, chained with a Drive file reference.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
