Productivity use cases
Where the hours actually come from: the four workplace patterns behind almost every real AI win, and how to spot them in your own week.
Ask people where AI saves time and you'll hear vague answers — 'emails, I guess?' The real wins cluster into four patterns. Learn to recognize the pattern and you'll spot opportunities everywhere; miss the pattern and you'll use AI as a novelty.
Pattern 1 — First drafts of anything
The blank page is where professionals lose hours. AI turns 'write from nothing' into 'edit something' — a fundamentally easier task. Emails, agendas, project briefs, job posts, performance-review bullet points, proposal skeletons. Rule of thumb: if the final version will be 60% AI draft + 40% your edits, you've cut the task in half.
Pattern 2 — Compression: long thing → short thing
Meeting transcripts → action items. A 30-page report → one-page brief. Forty customer reviews → five themes with example quotes. Compression is the model's safest skill (the facts come from your input) and often the biggest single time-saver for managers.
Pattern 3 — Translation between audiences
Not languages — registers. Engineer's update → executive summary. Legal clause → plain English for the sales team. Your frustrated bullet points → a diplomatic escalation email. Same facts, different audience; the model is exceptional at this.
Pattern 4 — Structure from mess
A rambling email thread → a table of who-owes-what-by-when. Free-text survey answers → categories with counts. Receipts pasted as text → an expense line list. Any time you're manually re-typing information that already exists in messy form, that's the pattern.
Here is my calendar and to-do list for last week: [paste it — strip names and sensitive entries first; "meeting re: employee issue" works fine]. Identify every task that fits one of these patterns: first-drafting, compression (long→short), audience translation, or structuring messy information. For each match, estimate minutes it took me and describe in one sentence how an AI assistant would help. Rank by time saved.
This prompt is the whole lesson turned into a tool. Sam ran it and found 6.5 hours across nine tasks — the weekly ops summary alone was 90 minutes of compression + translation.
'Making decisions' is not one of the four patterns. AI compresses the inputs to your decision and drafts the communication of it — the deciding stays with you. Teams that blur this line are how AI gets banned at companies.