The living playbook

What actually gets adopted — ranked by results

This is the moat. Every piece of feedback folds in here, so the highest-performing GenAI plays rise to the top automatically. Copy the code in an afternoon; you can't copy the accumulated adoption data that makes these rankings true.

All entries below are examples until you rate and adopt them — then real stats appear.

Turn meeting notes / transcripts into action items

example

Copilot in Microsoft Teams

Example play · not yet measured — rate & adopt it to start tracking.

View prompt

From the meeting notes below, produce: (1) a 3-bullet summary, (2) a table of action items with owner and due date, (3) any decisions made. Flag anything ambiguous.

Notes:
[paste notes]

Find the right policy / document across the company

example

Microsoft 365 Copilot (work)

Example play · not yet measured — rate & adopt it to start tracking.

View prompt

Find the current company policy on [topic]. Give me the answer in 3 bullets, the source document name, and the last-updated date. If there are conflicting versions, say so.

Summarise long email threads before replying

example

Copilot in Outlook

Example play · not yet measured — rate & adopt it to start tracking.

View prompt

Summarise this email thread in 4 bullets: what's been decided, what's open, who is waiting on me, and a suggested next reply in my voice (concise, professional).

Thread:
[paste thread]

Draft a first-version slide deck from a brief or doc

example

Copilot in PowerPoint

Example play · not yet measured — rate & adopt it to start tracking.

View prompt

Turn the brief below into a 6-slide outline: title, problem, solution, proof, plan, ask. For each slide give a headline and 3 bullets. Keep it executive-ready.

Brief:
[paste brief]

Analyse a spreadsheet / explain or build a formula

example

Copilot in Excel

Example play · not yet measured — rate & adopt it to start tracking.

View prompt

Look at the data described below. Suggest the 3 most useful insights, the chart that best shows each, and the exact Excel formula to compute them. Explain each formula in one line.

Data:
[describe columns / paste sample]

Automate a repeatable, multi-step process

example

Copilot Studio agent

Example play · not yet measured — rate & adopt it to start tracking.

View prompt

Map this recurring process into steps a low-code agent could run: trigger, inputs, the decision rules, and the output. Flag the one step that still needs a human.

Process:
[describe the routine]

How it grows: when you rate and adopt a use case on the home page, it folds into this playbook — increasing that play's adoption rate and rating, and re-ranking the library. New tasks then retrieve from here, so recommendations get better with every use.