The Prompt That Changed How I Write Every Single Document at Work

I work as a project manager, which means I spend a significant portion of my working life writing things — status reports, meeting summaries, escalation emails, change requests, stakeholder updates. None of it is glamorous writing. All of it has to be clear, professional, and fast.

I had been using AI tools inconsistently for about eight months, getting mediocre results, before I stumbled onto a prompt structure that changed everything. Not gradually — immediately and dramatically. I want to share it because I spent far too long getting below-average results from tools that were capable of much better.

The Problem With How Most People Prompt AI

The typical approach to AI writing assistance goes something like this: you need to write something, you open the AI tool, you type a vague description of what you need, and you get a generic draft that sort of addresses your situation but requires so much editing that you start to wonder if AI is actually saving you any time at all.

The problem is not the AI. The problem is that you have given the AI insufficient information to produce something specific and useful. You have essentially hired a very capable person and then refused to brief them properly.

The analogy I use: if you asked a new colleague to write a status report by saying “write a status report about the project,” you would get something generic. If you said “write a status report for my VP on the infrastructure migration — we are two weeks behind schedule because of a vendor delay, the revised completion date is March 15th, the main risks are now in the integration phase, and the tone needs to be honest about the delay but confident about our recovery plan” — you get something completely different and actually useful.

The Prompt Structure That Changed Everything

Here is the structure I now use for virtually every professional writing task. I call it the ACTOR framework: Audience, Context, Task, Output requirements, Restrictions.

Audience: Who will read this? Their role, what they care about, their level of technical knowledge, their relationship to the subject.

Context: What situation is this writing responding to? What background does the reader have? What has happened recently that is relevant?

Task: Specifically what am I asking the AI to produce? Not “write an email” but “write a 200-word email that does X, Y, and Z.”

Output requirements: Length, format, tone, structure. Should it have headers? Be bullet points or prose? Sound formal or approachable?

Restrictions: What should it not do? What to avoid, what not to include, what tone to stay away from.

What This Looks Like in Practice

Before the ACTOR framework, I might have typed: “Write an email to my client about a project delay.”

After the framework, the same request looks like this: “Audience: Sarah Chen, Senior Director at a financial services firm, technically sophisticated, values directness, has a good working relationship with me after 2 years but this is the first significant problem we have had. Context: Our software implementation is delayed 3 weeks because a third-party API we depend on had unexpected downtime that our testing did not catch. Task: Write an email informing her of the delay, explaining the cause honestly without excessive technical detail, telling her the revised timeline, and describing what we are doing to prevent this from happening again. Output requirements: Under 200 words, conversational but professional, no bullet points, should read like a confident professional taking ownership. Restrictions: Do not be defensive, do not over-apologize, do not use passive voice to obscure responsibility.”

The difference in output quality is not incremental — it is categorical. The second prompt produces something I can send with minimal editing. The first produces something I essentially have to rewrite.

Applying It to Different Document Types

For reports: The context section becomes crucial. Who commissioned the report? What decision is it informing? What does the reader already know and what do they need to learn?

For proposals: The audience section does a lot of work. What does this person or committee value? What objections are they likely to have? What would make them say yes?

For difficult emails: The restrictions section becomes essential. Do not sound passive aggressive. Do not over-apologize. Do not hedge to the point where the message is unclear.

For executive summaries: The output requirements section is key. Specifying “open with the key finding or recommendation before providing supporting context” produces a very different document than the default structure AI tends to use.

The One Thing That Still Requires Human Work

Even with a perfect prompt, AI does not know the interpersonal dynamics of your workplace. It does not know that your VP needs to feel like he was consulted before receiving news, not after. That organizational and relational knowledge is what you bring. The AI provides craft. You provide context. The combination is better than either alone.

How Long This Takes Once You Know the Framework

Writing out a full ACTOR prompt takes about 3 to 5 minutes. The resulting draft takes maybe 5 minutes to review and edit. For something that previously took me 20 to 30 minutes to write from scratch, I have cut the time to roughly 10 minutes while improving quality. Across the volume of professional writing I do in a week, that adds up to hours.

Frequently Asked Questions

Does this work with any AI tool? Yes. Claude, ChatGPT, and Gemini all respond significantly better to this structure than to vague prompts. Claude tends to handle nuanced tone instructions particularly well.

Do I really need all five elements every time? For simple tasks you can shorten it. For anything with significant stakes or complexity, all five elements produce meaningfully better results.

What if the AI still produces something I do not like? Give it specific feedback. “The tone is too formal — make it sound like a message between colleagues who know each other well.” Iteration with specific feedback is faster than starting over.

Conclusion

The difference between mediocre and excellent AI writing output is almost entirely in the quality of the prompt. The ACTOR framework — Audience, Context, Task, Output requirements, Restrictions — takes a few extra minutes to construct and consistently produces drafts that require significantly less revision. If you have been using AI for work writing and feeling underwhelmed, try this structure on your next document. The improvement is immediate.