Agent Library · By platform · Outlook · Teams · Excel · SharePoint
M
The inbox becomes a queue with drafts attached.
Microsoft 365 is where most of your company's working day physically happens, and where most of its institutional memory is trapped. Agents here do the two things that matter: they stop you writing from zero, and they make what you already wrote findable.
The inbox is the real job
Senior people spend their morning sorting, forwarding and writing the same replies they wrote last week.
Meetings produce discussion, not motion
Decisions are made and follow-through leaks because nobody owns writing it down.
SharePoint is a write-only archive
Everything is in there. None of it is retrievable when you need it.
The agents
6 agents. Each one earns its autonomy.
Agent
What it does
Autonomy
Human gate
Inbox Triage & Drafts
Outlook sorted by what genuinely needs you; replies drafted and waiting for the ones that do.
Notify
Escalations always human
Meeting-to-Action
Teams calls end with owners, deadlines and follow-up drafts, not with "someone should send notes".
Notify
—
Excel Report Assembly
Recurring reports built from source files automatically. You review the numbers, not the formulas.
Review
—
SharePoint Answers
Questions answered from your own policies, precedents and past work — with sources cited.
Auto
—
Document Drafter
First drafts of proposals, letters and SOPs in Word, in your templates and your voice.
Review
—
Calendar Coordinator
Scheduling handled, conflicts flagged, prep notes attached before every meeting.
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.
You stop writing from zero
Everything routine arrives pre-drafted. Your job becomes editing and judgement.
Meetings produce motion
Every discussion converts to owned actions automatically, and follow-through stops leaking.
Institutional memory on tap
What the company knows becomes something anyone can ask, with citations.
The eval bench
What these agents have to pass before they touch real work.
- Triage accuracy against 200 historic emails with known correct handling
- Draft send-rate tracked as unedited / minor-edit / rewritten, weekly
- Every SharePoint answer cites a retrievable source or returns "I don't know"
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 Outlook.
A Blueprint maps which of these agents pays off first in your exact setup — and which to leave alone.
Apply for a Blueprint