Stop treating ChatGPT like a Google search bar if you actually

I've spent a lot of time iterating on how to automate the boring parts of my week, and the biggest win has been my Friday status update. Instead of staring at a blank doc for an hour trying to remember what I actually did on Tuesday, I feed my rough notes into a specific prompt. The key is to tell the AI exactly who the audience is (my manager) and what the tone should be (confident but concise).

If you want to try this, here is the exact prompt I use. I've designed it to be a practical tutorial for anyone who needs to turn a mess of bullet points into a professional summary.

You are an expert Executive Assistant and Communications Specialist. Your goal is to transform my raw, messy weekly notes into a polished, professional status update for my manager.

**Context:**
- Tone: Professional, proactive, and concise.
- Format: Bullet points.
- Audience: A busy manager who cares about outcomes, blockers, and next steps, not every tiny detail.

**Structure of the output:**
1. **Key Achievements:** High-level wins from this week. Focus on the "so what?" (e.g., instead of "fixed bug X," use "fixed bug X, which reduces page load time by 20%").
2. **In Progress:** Current projects and their status.
3. **Blockers/Needs:** Anything stopping me from moving forward that requires manager intervention.
4. **Plan for Next Week:** Top 3 priorities for the coming days.

**Input Data:**
[Insert your rough notes/bullet points here]

The reason this works where others fail is the "so what?" instruction. By forcing the LLM to look for the impact of the action rather than just repeating the action, the output actually looks like it was written by a high-performing employee rather than a bot.

For those starting from scratch with prompt engineering, remember that the more "scaffolding" you provide, the less the AI has to hallucinate or guess. I've found that adding a "Format" section is the fastest way to stop the AI from rambling. When you define the exact headers you want, you save yourself from having to manually edit the response.

This has basically become a step-by-step system for me every Friday afternoon. It turns 60 minutes of dread into about 5 minutes of reviewing and sending. It's a simple shift, but it's the difference between using an AI as a toy and using it as a productivity engine.