You asked AI to help you prepared for your preformance review. It gave you a list of tips that looked like they came from a 2011 HR handbook. Speak confidently. Highlight your achievements. Be specific.
You already knew that. You have been doing this for twenty years.
You closed the tab. Made a mental note that AI was not really for you. Moved on.
Here is what actually happened. The tool was not broken. The instruction was. And the fix is not complicated. It is just specific. One you see it, you cannot unsee it.
Why AI Keeps Giving You Generic Answers
AI does not know you. It has not idea how long you have been in your industry, what your manager values, how political or organization is, or what "good" sounds like in your world. Without that contexct, it pattern-matches to the most common version of your request. The average answer for the average person. Written for nobody in particular.
Think about it this way. If a brand new junior colleague walked into your office and said "help me prepare for something," you would ask twenty questions before you said a word. AI skips that conversation entirely unless you have it yourself, upfront, in the prompt.
This is exactly where your experience becomes the advantage. You know things AI does not. Your job is to tell it.
The Four-Part Framework for AI Prompts That Actually Work
A prompt that works has four parts. None of them require technical knowledge. All of them matter.
Tell AI who it is talking to. Not just your job title. Your context, your industry, your specific situation. The more specific, the better the output.
Part Two: TASK
Tell AI exactly what you need. Not "help me prepare." Describe the kind of preparation, the type of conversation, and the outcome you are working toward.
Part Three: TONE
Tell AI how the output should sound. Warm? Direct? Confident but not aggressive? It will match whatever you describe. If you do not specify, it will guess.
Part Four: CONSTRAINT
Tell AI what to avoid. What the limits are. What you do not want. GThis is the part that prevents the output from sounding like everyone else.
Here is what that looks like in practice.
Before and After: The Same Request, Two Very Different Results
The prompt that gets nothing useful:
"Help me prepare for my performance review next week. I want to make sure I come across well."
What comes back: a generic checklist. Speak clearly. Use data. Prepare examples. Nothing you did not already know.
" I am a marketing director with 18 years of experience in financial services [ROLE]. I have a performance review next week with a manager that tends to focus on metrics over relationships [TASK CONTEXT]. I have had a stronger year commercially but also led my team through a difficult restructure. I want to prepare a clear confident narrative that acknowledges the hard year without apologizing for it, and positions me for a promotion conversation in Q1 [TASK]. Write me the talking points I should walk in with, in a direct and professional tone [TONE]. No generic advice. No corporate language. No bullet-pointed list of tips [CONSTRAINT]."
What comes back: a narrative. One you can actually walk into the room with.
Sample AI Output:
"This year asked a lot of you and your team, and you delivered on both fronts. Commercially, you hit your targets. Operationally, you held the team together through a restructure that most organizations get wrong. That is not one achievement. That is two. The question for Q! is not whether you are ready for the next level. It is how we make the case together."
This is not a generic template. This is a narrative. The difference is not the tool. It is the instruction.
The One Compoenent Most Professionas Skip
The role component. Almost everyone leaves it out. This is where your twenty years of hard-won professional knowledge stops being background noise and becomes a direct input into the output. When you tell AI who you are, what you know, and what the stakes are, it stops writing for the average person and starts writing for you.
Experience + AI = Indispensable. Not the tool itself. The way you direct it. Your experience is what makes the output worth anything at all.
Experience + AI = Indispensable. Not the tool itself. The way you direct it. Your experience is what makes the output worth anything at all.
A Ready-to-Use Prompt: The Salary Conversation
Same four parts. Higher stakes. Take this directly to ChatGPT or Copilot right now.
Copy and Use This Prompt
"I am a [your role] with [X] years of experience in [your industry]. I am preparing to ask for a salary increase of [X%] in a conversation with my manager, who is [choose: data-driven and focused on ROI/ cautious and needs reassurance/ supportive but constrained by budget]. My strongest case is [ a result you delivered/ a responsibility you have taken on/ a gap you have filled]. I want to make the ask clearly and confidently without over-explaining or sounding apologetic. Write me the opening two minutes of that conversation in a warm, direct, professional tone. No hedging. No qualifiers. Just the ask and the case for it."
Paste that in. Read what comes back. It will not be perfect on the first attempt. That is completely normal. You are the editior here, not the recipient.
If the tone is off, tell it. Reply in the same chat: "Make this less formal" or "cut the second paragraph" or "this sounds to aggressive, soften it." It will adjust. Push back as many times as you need to until it sounds like you. Because it should.
You Are Already Qualified to Lead This
You walk into rooms for decades with expertise, judgment, and the context that nobody else had. AI just needed to know that. Now it does.
Effective prompting is not a technical skill. It is a communication skill. And if you have spent twenty years briefing colleagues, managing up, and translating complex situations into clear decisions, you are already good at it. You just did not know it applied here too.
Start with one prompt today. The performance review template above, or the salary conversation. Fill in your details. See what comes back. Adjust until it's "good" for your expectations. That is the whole process.
That is what Experience + AI= Indispensable actually looks like in practice. Not in theory. Right now, in your next meeting, your next conversation, your next high-stakes moment.
Frequently Asked Questions
Why do AI prompts give Generic answers?
AI tools like ChatGPT and Copilot pattern-match to the most common version of your request. Without context about who you are, your industry, your tone, or your specific situation, the output defaults to a generic answer written for nobody in particular. The fix is to provide that context upfront in your prompt.
What makes an AI prompt actually work?
An effective AI prompt has four components: Role (who you are and your professional context), Task (exactly what you need), Tone (how the output should sound), and Constraint (what to avoid). Together these four elements give AI the information it needs to produce useful, specific output rather than generic advice.
How do I write a better AI prompt for work?
Start by telling AI who you are: your role, your industry, and the specific situation. Then describe exactly what you need, how it should sound, and what to avoid. The more professional context you provide, the more targeted and useful the output will be. Your experience is not background noise. It is a direct input that improves every result.
Do I need technical skills to write good AI prompts?
No. Effective prompting is not a technical skill. It is a communication skill. You are telling AI what you need, who you are, and how the result should sound. If you can brief a colleague, you can write a prompt that works.
What is the difference between a good and bad AI prompt?
A weak prompt gives AI no context: "Help mem prepare for my performance review." A strong prompt gives AI your role, your situation, your desired outcome, the tone you want, and what to avoid. The output difference is significant. One gives you a generic checklist. That other gives you a narrative you can walk into the room with.
Can I use this prompting approach in ChatGPT and Copilot?
Yes. The four-part framework works across all major AI tools including ChatGPT, Microsoft Copilot, and others. The principle is the same regardless of platform: give AI your professional context, your specific task, your desired tone, and your constraints.
Continued Reading