The Prompt Vault: Seven Prompts I Use Every Single Week That Quietly Run My Business

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

I want to start with a small confession. Most prompt guides on the internet are useless to me.

Not because the prompts are bad. Some of them are clever. Some of them produce decent output. But almost none of them survive contact with a real workweek. They look great in a tutorial, then break the moment you try to actually rely on them for something that matters.

The prompts I am sharing today are different. These are the seven I have used, refined, and reused over the last 18 months across hundreds of real tasks. Each one solves a specific recurring problem. Each one has been tested in production. Each one is short enough to memorize or save in a snippet manager.

If you only steal one thing from this newsletter all year, make it this article.

I run all seven of these through both Claude and ChatGPT depending on the use case. I will tell you which one I prefer for each prompt. Both work, and the access through Galaxy.ai makes it easy to swap between them.

Here we go.

Prompt One: The Decision Filter

This is the prompt I run when I am stuck on a decision and cannot tell if I am overthinking it or undercooking it. I use this maybe three times a week.

"You are a senior advisor with 20 years of experience helping small business owners make high stakes decisions. I am about to make the following decision: [INSERT DECISION]. Here is the context I am working with: [INSERT 3 TO 5 BULLET POINTS]. Here is my current leaning: [INSERT YOUR INSTINCT]. Do three things. First, identify the strongest argument against my current leaning. Second, identify the most likely failure mode if I proceed with my current leaning. Third, ask me one clarifying question that would change your analysis if answered differently. Be direct. Do not hedge."

What this prompt does is force the AI to disagree with you. Most people use AI as an echo chamber. They describe their plan, the AI says "great plan," and they feel validated. This prompt does the opposite. It asks the AI to argue against you, surface the failure mode, and probe the assumption you did not realize you were making.

I prefer Claude for this one. It tends to be slightly less agreeable and produces sharper counterarguments.

Prompt Two: The Email Triage

This is the prompt I run on emails that require careful responses. Not the routine ones. The ones where the wording matters.

"You are an executive communications specialist. I need to respond to the following email: [PASTE EMAIL]. The relationship context is: [INSERT 1 TO 2 SENTENCES]. The outcome I want from this exchange is: [INSERT GOAL]. Draft three different replies. The first should be direct and brief. The second should be warm and relationship building. The third should be firm but professional. For each reply, list the strongest argument for why that approach is correct in this situation. Then recommend which of the three I should send and explain your reasoning in two sentences."

The unlock here is the three options framing. You do not get a single reply that you have to either accept or reject. You get three real choices and a recommendation. You can pick the best one or hybrid two of them in 30 seconds.

I prefer ChatGPT for this one. It tends to nail the tonal range slightly better.

Prompt Three: The Meeting Prep

I run this one before any meeting that matters. Sales calls, partnership discussions, internal alignment meetings, anything where the stakes are higher than routine.

"You are a strategic advisor preparing me for an important meeting. Here is the context. Meeting type: [INSERT TYPE]. Other party: [INSERT WHO]. My desired outcome: [INSERT GOAL]. Their likely desired outcome: [INSERT GUESS]. Generate the following. First, three questions I should ask early to surface their real priorities. Second, two objections they are likely to raise and a one sentence response to each. Third, one specific opening line that signals competence without being aggressive. Fourth, the single most important thing I should listen for during the conversation."

This prompt prepares you in roughly three minutes for any meeting. The questions, the objections, the opening, and the listening focus. By the time you walk into the room, you are not winging it.

I use Fathom to record the actual meeting, which means I can review afterward how well the prep predicted the actual conversation. Over time, you get sharper at the prep because you can see what worked.

I prefer Claude for this one. It produces more layered strategic thinking.

Prompt Four: The Content Refiner

I use this prompt at least once a day on every piece of content I publish. It is the difference between content that sounds like a chatbot and content that sounds like a human being.

"Read the following draft and identify three specific weaknesses. Look for sentences that feel generic, paragraphs that lose energy, and any place where I am hedging instead of taking a position. Do not rewrite. Just identify the weaknesses. Then suggest one specific change for each weakness in the form of a sentence I could substitute. Draft: [PASTE DRAFT]."

Notice what this prompt does not do. It does not ask the AI to rewrite the content. That is the trap most people fall into. They run a piece through ChatGPT for "polish," and what comes out the other side is identical to every other piece of polished AI content on the internet. Bland, smooth, and forgettable.

This prompt asks the AI to be a sharp editor instead. It surfaces the weak spots. You make the changes yourself, in your voice. The output stays human.

I prefer Claude for this one. It is more willing to be critical without being dismissive.

Prompt Five: The Research Synthesizer

When I need to get up to speed on a topic quickly, I run this prompt.

"I need to understand [TOPIC] well enough to have an intelligent conversation with [TYPE OF EXPERT]. I have 20 minutes to learn this. Give me the following. First, the three most important concepts in this domain explained in one sentence each. Second, the two most common misconceptions someone like me would have. Third, three intelligent questions I could ask an expert that would signal I have done my homework. Fourth, the one thing in this domain that has changed in the last two years that even some experts have not fully internalized."

This prompt produces a 20 minute education on almost any topic. The misconceptions section is the most valuable part. It surfaces the assumptions you would otherwise carry into the conversation and embarrass yourself with.

I split this one between Claude and ChatGPT depending on the topic. Claude tends to be better for nuanced or human centered topics. ChatGPT tends to be better for technical or data heavy topics.

Prompt Six: The Pricing Sanity Check

This is one of my favorite prompts and most owners do not have it in their library. I run this any time I am about to send a proposal, set a price, or negotiate a contract.

"You are an experienced pricing consultant. I am about to charge [AMOUNT] for [DESCRIBE OFFER]. The customer is [DESCRIBE CUSTOMER PROFILE]. The market context is [DESCRIBE COMPETITIVE LANDSCAPE]. Tell me three things. First, is this price too low, too high, or roughly correct, and why? Second, what is one specific change to the offer that would justify a 30 percent higher price? Third, what is the most common reason I am likely to undercharge in this situation, and how do I avoid it?"

This prompt has saved me real money. Not 5 percent here or there. Significant amounts. Owners chronically underprice because they are afraid to hear no. This prompt confronts that pattern directly and surfaces the moves you can make to charge more without losing the deal.

I prefer ChatGPT for this one. It tends to be more grounded in market realities.

Prompt Seven: The Weekly Reset

The last prompt is the one I run every Sunday evening. It is the closest thing I have to a personal coaching session.

"You are my executive coach. Below is a list of what I committed to this week and what actually happened. Review the list and identify the following. First, the single biggest gap between what I said I would do and what I actually did. Second, the one pattern that shows up in my misses that I keep ignoring. Third, the highest leverage commitment I should make for next week, given what I learned this week. Be honest. Do not flatter me. Here is the week in review: [PASTE LIST]."

This prompt produces a 10 minute coaching session every Sunday. Over time, the patterns become impossible to ignore. You stop making the same mistakes because the AI keeps surfacing them and asking you to do something about it.

I have been running this prompt for 18 months. It has changed how I plan, how I commit, and how I show up. Of all the prompts in this article, this is the one with the deepest compounding effect.

I prefer Claude for this one. It is willing to be honest without being harsh.

The Real Lesson

These seven prompts are not magic. They are not even particularly clever. What they have in common is that they all force structured thinking around a recurring decision point in my week. Decisions, communication, meeting prep, content quality, research, pricing, and weekly review. That is most of the cognitive load of running a business.

By offloading the structure to a prompt, I free up my attention for the parts of the work that actually require my judgment. I am not asking the AI to do my thinking. I am asking it to organize the thinking I would otherwise be too lazy to do well.

How To Make A Prompt Yours

One thing I have learned the hard way. A prompt you copy from someone else will work about 70 percent as well as a prompt you have personalized.

The personalization is not complicated. You take the base prompt, run it three or four times on real situations from your business, and edit the prompt every time the output is not quite what you wanted. After about a week of use, your version of the prompt will look noticeably different from the one I shared. That is the point.

The other adjustment that matters is voice calibration. Every prompt I shared today is voice neutral on purpose. You should add one or two lines that anchor the output to your voice. Something like, "Write in a direct, conversational tone. Avoid corporate jargon. Use short sentences. Lead with the conclusion, then explain." That single addition makes the output feel like you wrote it instead of like a chatbot wrote it.

The third move is context loading. The more relevant context you give the AI up front, the better the output gets. If you are going to use the meeting prep prompt regularly, save a one paragraph description of your business, your typical client, and your standard offer. Paste that in every time before the prompt itself. The output sharpens dramatically.

A Word On Where To Run These

I am asked at least once a week which AI tool I recommend for running prompts like these. The honest answer is that the differences between the major models are smaller than the differences in how you use them. A good prompt run on a mediocre model still beats a bad prompt run on the best model.

That said, my workflow looks like this. I keep Claude open for anything that requires nuance, voice work, or strategic thinking. I keep ChatGPT open for anything that requires fast generation, structured output, or technical work. I keep Grok handy for anything where I want a slightly different perspective or want to cross check the other two.

Running all three through Galaxy.ai means I have one subscription instead of three and I can pivot between models without losing context. That alone saves me an hour a week in tab switching and login friction.

If you want the full vault, including the 32 prompts I have refined for client work across automation, content, sales, hiring, and operations, that is the core of the AI Workflow Blueprint. Reply to this email with the word BLUEPRINT and I will send it your way.

If you want me to walk through customizing these prompts for your specific business, the AI Business Accelerator includes the implementation calls where we do exactly that. Reply with ACCELERATOR for the details.

Save these seven. Use them this week. Watch how much faster you move on the decisions that used to take you a whole afternoon.

See you tomorrow on The Tool Review.

Jordan Hale

The AI Newsroom

Wispr Flow recognizes variable names, file references, and framework syntax mid-dictation. Speak your prompt, get developer-ready text for GitHub, Jira, or your editor. No mangled syntax. Ever.

Try Wispr Flow free