Agent Library · By platform · Gmail · Drive · Sheets · Meet
G
Everything the company knows is in Drive. Unfindable.
Google Workspace is where most growing companies actually live — and where the least structure exists. Drive has no schema, Gmail has no memory, and Sheets has whatever the last person built. That looseness is exactly what an agent layer is good at absorbing.
Drive has no index
Every document the company has ever produced, in folders whose logic left with the person who made them.
Gmail is the queue and the archive
Work arrives, gets handled, and disappears into a mailbox only one person can search.
Sheets reporting is fragile
Business-critical reporting held together by formulas one person understands and nobody documented.
The agents
6 agents. Each one earns its autonomy.
Agent
What it does
Autonomy
Human gate
Gmail Triage & Drafts
Sorted by what genuinely needs you, with replies drafted for the routine majority.
Notify
Client escalations always human
Drive Answers
Ask a question, get an answer sourced from your own documents with links back to the file and the line.
Auto
—
Sheets Reporter
Recurring reports assembled from source data on a schedule, with the anomalies called out.
Notify
—
Meet-to-Action
Calls end with owners, deadlines and drafted follow-ups written straight back into your tools.
Notify
—
Doc Drafter
Proposals, SOPs and client documents drafted in your templates from your own precedents.
Review
—
Calendar Coordinator
Scheduling, conflict flagging and pre-meeting briefs attached automatically.
Auto
—
Autonomy key — Auto: executes above 95% confidence. Notify: executes at 85–95% with notification and undo. Review: always queued for a human. Flag only: never acts, alerts a person.
What changes
Three shifts, not a feature list.
Drive becomes answerable
The archive turns into something the whole company can query, with citations.
One inbox stops being one person's bottleneck
What was locked in an individual mailbox becomes routed, drafted and visible.
Reporting stops depending on one spreadsheet author
The assembly moves into a system. The judgement stays with your team.
The eval bench
What these agents have to pass before they touch real work.
- Retrieval accuracy on 100 questions with verified answers inside your own Drive
- Citation integrity: every claim links to a retrievable file, or the agent declines to answer
- Triage accuracy against 200 historic emails with known correct owners
Written against your data, not ours. The bench is built during Blueprint and runs on every deploy from then on.
Related
Where this connects.
See it on your Gmail.
A Blueprint maps which of these agents pays off first in your exact setup — and which to leave alone.
Apply for a Blueprint