You already use ChatGPT daily. Yet the week still fills up: proposals eat the evening, content stays at "I'll post this week," replies get uneven, quotes start from a blank page. More prompts don't fix that — cadence does. Here's the weekly OS I run solo.
Slot A — Client proposal (~45 min)
Give the model a spine: restate the client's problem in their words; the outcome in one measurable sentence; in scope / out of scope / assumptions; timeline by phase with a deliverable each; price + payment split. Prompt shape: role = senior consultant writing to a busy non-technical client; constraints = under 400 words, no jargon, one price, explicit out-of-scope list. The win isn't typing speed, it's never re-deciding the structure.
Slot B — Content batching (60-75 min weekly)
Dump 10 raw ideas from real work. Pick 3. Write each as claim + proof in one line. Let the model expand the middle; rewrite the first and last paragraph by hand. Then schedule. If it isn't claim + proof, it isn't a post yet.
Slot C — Inquiry replies + scope fences (30-40 min)
- Good fit: confirm the problem, propose a 20-minute call, name your price range up front.
- Unclear: ask exactly two questions, no more.
- Not a fit: decline in three lines, refer someone.
Fence I reuse: "That's outside this phase — happy to quote it as a separate add-on."
Slot D — Quote + simple SOW (40-50 min)
Deliverables, revision count, change-order trigger, payment schedule, validity date. Template it once; only numbers change.
Slot E — Weekly review (25-35 min)
What shipped? What slipped, and why? What are next week's five slots? Lock the calendar before closing the laptop.
An LLM removes typing friction, not decision friction. Fixed slots remove the deciding.
The packaged version
Every slot as SOP + prompts + filled example + checklist: Solo Operator AI Weekly System — $17
Narrower pieces: AI Workflow Kit — $12 and Quote, Contract & Tax Prep Kit — $15 (not legal advice). I'm the maker; questions about running slot B or C on a real week are welcome in the comments.
Top comments (0)