johnmaconline

Prompt templates are a great way to talk to AI because they turn uncertainty into structure. 

Plus, AI really does want to please you. If you get good at telling it what you want, you’ll get better results. Prompt templates provide a structured method for telling it what you want. 

Any prompt is just a starting point, and with LLMs, small changes in wording or structure can wildly change output.  

So templates give you consistency. They define roles, context, constraints, and output format. They embed the lessons of what’s worked before, so you don’t have to reinvent the wheel every time.  

Here is a convenient list and links to the templates I described in this series.

AI Prompt Templates

• Persona and Expertise – Use this for most tasks.
• Expert Panel – Use this for exploring different opinions and perspectives. 
• The Fact Checker – Use this when you need accuracy.
• The Procedural or Recipe Framework – Use this to get instructions.
• Navigating a Government System – Use this when you need guidance through the bureaucracy and regulations. 
• Test and Interview Prep – Use this to prepare for a test or interview. 
• Templatized Output – Use this when you want a specific output format — mass email, sales message, analysis, etc. 

Here’s another great resource for prompt templates. HuggingFace is a strange name with an interesting history, but regardless, what you’ll find at that link (if you have the patience) are hundreds of templates that AI teams use to test and compare various models. If none of the templates I provided above fit your use case, you can likely find a template that can guide you in this sea of test prompts.

With a good template, you don’t stare at a blank screen anymore. You fill in what changes (product name, audience, tone, data, etc) and hit go. Even non-deterministic LLMs can start to feel consistent (if not deterministic). 

When you sit down to talk to AI, whether it’s coding, marketing, writing, strategy, or figuring your way through a system, you can start here. Pick your template, fill in the info, and send it off to the LLM. 

You can use a template to get a template.

You can proactively prevent hallucinations by using a template of placeholders to direct the LLMs output. 

The template:

You are an earnest worker and excellent at following directions. Your goal is always to give me exactly what I ask for. Do not make stuff up. Rather, ask me questions if you run across any ambiguity. 

I am providing a specific template for your output. This template includes bracketed [CAPITAL WORDS] as placeholders for the content you are to generate. Fit your response into these placeholders and follow the template's style and formatting. 

Here is the output template:
Your template with [PLACEHOLDERS].

Apply this template to: [Your request]

An example:

 You are an earnest worker and excellent at following directions. Your goal is always to give me exactly what I ask for. Do not make stuff up. Rather, ask me questions if you run across any ambiguity. 

I am providing a specific template for your output. This template includes bracketed [CAPITAL WORDS] as placeholders for the content you are to generate. Fit your response into these placeholders and follow the template's style and formatting. 

Here is the output template:
Target Market: [TARGET_MARKET]
Top Competitors (5):
 1. [COMPETITOR_1] — [STRENGTHS_1] / [WEAKNESSES_1]
 2. [COMPETITOR_2] — [STRENGTHS_2] / [WEAKNESSES_2]
 3. [COMPETITOR_3] — [STRENGTHS_3] / [WEAKNESSES_3]
 4. [COMPETITOR_4] — [STRENGTHS_4] / [WEAKNESSES_4]
 5. [COMPETITOR_5] — [STRENGTHS_5] / [WEAKNESSES_5]
Market Differentiators:

 • [DIFFERENTIATOR_1]
 • [DIFFERENTIATOR_2]
 • [DIFFERENTIATOR_3]
Opportunity Gaps Identified: [GAPS]

Apply this template to: 
Map the competitive market landscape in the embedded DevOps tools space. I don't care about the semiconductor companies that provide DevOps tools along with their hardware solutions. Map the strengths, weaknesses, and differentiators compared to 4TLAS's Fuze (4tlas.io/fuze). 

Another example for generating sales emails:

You are an earnest worker and excellent at following directions. Your goal is always to give me exactly what I ask for. Do not make stuff up. Rather, ask me questions if you run across any ambiguity. 

I am providing a specific template for your output. This template includes bracketed [CAPITAL WORDS] as placeholders for the content you are to generate. Fit your response into these placeholders and follow the template's style and formatting. 

Here is the output template:
Hello [OWNER_NAME],
My name is [YOUR_NAME], and I help local trade businesses like yours—[BUSINESS_NAME]—optimize their taxes and reduce stress during filing season.
Based on your work in [TRADE_TYPE], you may qualify for:
 1. [DEDUCTION_1]
 2. [DEDUCTION_2]
 3. [DEDUCTION_3]
If you want me to run a free deduction review, just reply to this email.
Best regards,
[YOUR_NAME]

Apply this template to:
Generate a sales email for all trade-like businesses (landscaping, roofing, contracting, etc) in the local geographic area, and create an email that is advertising our tax prep services specializing in sole-proprietor trade businesses. 

Why this works:

The template does a couple things: 1) it forces the LLM to focus on what you care about, 2) it creates a consistency for the output. You can create a wonderful human-AI mix by using templatized output. 

Up next: Wrapping it all up

AI can be a great companion to help you get ready for whatever. The test. The interview. The hard conversation with your derailing colleague. 

Here’s a template to help the LLM help you.

You are an experienced teacher, guide, and interviewer in the field of [name the field].

Please conduct an interview/test/conversation preparation session for [the role/exam/topic/problem] with me. You are helping me practice for this event. 

Do the following:
1. Ask me to provide the relevant materials for this interview (resume, job description, study materials, any historical information, etc)
2. Review these materials and identify the areas of focus
3. Create a structured practice session with:  
  - [X] questions of varying difficulty levels  
  - Follow-up questions based on my responses  
  - Real-time feedback on my answers
4. After each response, briefly indicate if I should expand, clarify, or move on
5. End with a summary of strengths and areas for improvement

Type of preparation: [job interview/certification exam/academic test/etc.]

Focus area: [specific role, subject, or skill]

Please start by asking for my materials.

An example:

You are an experienced teacher, guide, and interviewer in the field of recruiting and hiring.

Please conduct an interview/test/conversation preparation session for a job interview with me. You are helping me practice for this event. 

Do the following:
1. Ask me to provide the relevant materials for this interview (resume, job description, study materials, any historical information, etc)
2. Review these materials and identify the areas of focus
3. Create a structured practice session with:  
  - [X] questions of varying difficulty levels  
  - Follow-up questions based on my responses  
  - Real-time feedback on my answers
4. After each response, briefly indicate if I should expand, clarify, or move on
5. End with a summary of strengths and areas for improvement

Type of preparation: job interview

Focus area: Engineering and Product Leadership

Please start by asking for my materials.

Why this works:

You’ve told the LLM to use your specific knowledge sources as the basis for the conversation. It has a huge knowledgebase behind it, which then helps it find questions and concepts that you may not have considered. 

Next up: Specific templatized output

AI may finally help us crack government systems. 

You, me, and they all have to navigate these systems. Whether you’re wondering if you need a building permit to turn your basement into a bedroom, need to know how to file a successful workers’ compensation claim, or you’re moving to a new state and need to figure out how to transfer your driver’s license and vehicle registrations, these systems can be daunting. 

But your friendly, neighborhood LLM is here to help. 

A template:

You are an expert at government processes and an experienced liaison between the system and clients trying to navigate those processes. 

I need help navigating the government process for [what are you trying to do].

Give me practical, step-by-step guidance. In addition, explain how the system works at a high level, and list the potential pitfalls that I must look out for.

Use plain language and keep the instructions direct.

Here are my details:
Who am I: [Your demographics, title, or anything relevant to this process]
Goal: [What I’m trying to achieve]
Agency or System: [If known]
Jurisdiction: [Country / State / City / Institution]
Current Status: [What I’ve done so far and what’s pending]
Constraints: [Deadlines, missing documents, budget limits, prior mistakes, legal issues, etc.]
Who else is involved: [Agencies, employers, lawyers, dependents, contractors]
Dates and Deadlines: [Known or approximate]
Special Factors: [Military service, disability, immigration category, business type, dependents, prior visas, property type, etc.]

What I want from you: 
1. A simple overview of how this system works 
2. A numbered action plan for my situation 
3. All required forms, documents, IDs, and evidence 
4. Expected fees 
5. Key decision points and bottlenecks 
6. Common mistakes and pitfalls and how to avoid them 
7. A realistic timeline with estimated durations 
8. Questions the agency is likely to ask me 
9. A script for calling or emailing the agency 
10. A backup plan if the main path stalls 
11. Any ways to speed up or optimize the process 
12. Any missing inputs you need—list them clearly 
13. The most important person (title or position) that I should contact who can act like an advocate for me if I need one. 
14. Which is my best approach for completing this process successfully: 1) online, 2) make a phone call, 3) show up in person to an office. 
15. Ask me to upload any relevant documents to you to help find the solution

Keep it practical and focused on real-world steps.

An example:

You are an expert at government processes and an experienced liaison between the system and clients trying to navigate those processes. 

I need help navigating the government process for opening my own auto mechanic's business.

Give me practical, step-by-step guidance. In addition, explain how the system works at a high level, and list the potential pitfalls that I must look out for.

Use plain language and keep the instructions direct.

Here are my details:
Who am I: Mid-30's mechanic looking to leave my current employment to start my own business. 
Goal: I want to open my own garage at my house and ensure that I have everything I need to do it legally. 
Agency or System: I don't know which agency matters. 
Jurisdiction: Berks county, PA
Current Status: I need to understand if I can open this business at my house and what I need to do to make it happen. 
Constraints: the address is 548 Bertolet Mill rd, oley, pa. 
Who else is involved: I don't know
Dates and Deadlines: Want to be up and running in 2 months. 
Special Factors: I may employ others at some point, but not right away

What I want from you: 
1. A simple overview of how this system works 
2. A numbered action plan for my situation 
3. All required forms, documents, IDs, and evidence 
4. Expected fees 
5. Key decision points and bottlenecks 
6. Common mistakes and pitfalls and how to avoid them 
7. A realistic timeline with estimated durations 
8. Questions the agency is likely to ask me 
9. A script for calling or emailing the agency 
10. A backup plan if the main path stalls 
11. Any ways to speed up or optimize the process 
12. Any missing inputs you need—list them clearly 
13. The most important person (title or position) that I should contact who can act like an advocate for me if I need one. 
14. Which is my best approach for completing this process successfully: 1) online, 2) make a phone call, 3) show up in person to an office. 
15. Ask me to upload any relevant documents to you to help find the solution

Keep it practical and focused on real-world steps.

Why this works:

Transparency. Government processes and the systems built around them are all, by definition, public, and the procedures and requirements for them are posted publicly. Plus, these processes have a fair amount of media coverage, especially if they go wrong. These two factors combine to make a great set of training data for the LLMs.  If you know what to ask for and how to ask for it, ie, as this template provides, AI can help you be successful. 

Up next: Interview and Test Preparation

AI can be great at giving instructions and step-by-step procedures. 

Here’s a good template to help you get what you want:

You are an expert in [the genre for your objective] and you are an experienced teacher who understands how to translate complex concepts into meaningful information for your audience.

Create a step-by-step guide to accomplish [specific goal]. 

Structure your response as:

**Requirements:**
[List what's needed]

**Preparation:**
[Setup steps]

**Instructions:**
[Step 1 with specific actions]
[Step 2 with specific actions]
[Continue...]

**Troubleshooting:**
Common issues and solutions

**Variations:**
Alternative approaches

Goal: [Your specific objective]

An example:

You are an expert in DevOps and cloud computing and you are an experienced teacher who understands how to translate complex concepts into meaningful information for your audience.

Create a step-by-step guide to set up a GitLab pipeline that runs on AWS. Structure your response as:

**Requirements:**
- I'm a software developer so you can use software terminology
- The runner in AWS must prioritize cost efficiency over speed of completion
- We're building with ARM gcc and make

**Preparation:**
- What do I need?
- What might cost money?

**Instructions:**
- Make the procedure for a "Hello, world" application.
- You make the application

**Troubleshooting:**
Common issues and solutions

**Variations:**
Alternative approaches

Goal: Kick off a GitLab CI pipeline that builds our "hello, world" application in an AWS runner with ARM gcc and make. 

Why this works:

You are telling the LLM exactly what you want in your procedure. AI wants to make you happy, so it’ll follow these instructions. Plus, you’ve told it that it’s good at translating complex concepts into information that you will understand. 

Next up: The government system navigator. 

Yes, LLMs hallucinate, so you might be wary of their ability to fact-check. But fact-checking is a very good use case for AI, as long as you give it the right instructions. 

NOTE: This approach is good for general fact-checking and fact-checking claims about publicly available information. If you want to fact-check your proprietary information or something esoteric, you will want to build and use a RAG. 

Here is a great fact-checking template you can copy/paste:

You are a tireless researcher and excellent at fact-checking.   

Thoroughly fact-check the text I provide below. Analyze every factual claim, statistic, date, name, technical specification, and verifiable statement.

Your response should only include a "fact-check list" section with:

**Claims that should be verified:**
1. [Specific factual claim 1]
2. [Specific factual claim 2]
N. [More...]

**Information to double-check:**
- [Statistics or data points]
- [Dates or timeframes]
- [Technical specifications]

**Questionable claims or potentially inaccurate:**
- [List claims that seem false or dubious]
- [Include contradictions within the text]
- [Note implausible statements]

**Vague or misleading statements:**
- [Statements that lack specificity]
- [Claims that need sources or context]

**Confidence levels:**
- High confidence: [Claims you're very sure about]
- Medium confidence: [Claims that might need verification]
- Low confidence: [Claims you're uncertain about]

--
Text to fact-check: [PASTE YOUR TEXT HERE]

An example:

You are a tireless researcher and excellent at fact-checking.   

Thoroughly fact-check the text I provide below. Analyze every factual claim, statistic, date, name, technical specification, and verifiable statement.

Your response should only include a "fact-check list" section with:

**Claims that should be verified:**
1. [Specific factual claim 1]
2. [Specific factual claim 2]
N. [More...]

**Information to double-check:**
- [Statistics or data points]
- [Dates or timeframes]
- [Technical specifications]

**Questionable claims or potentially inaccurate:**
- [List claims that seem false or dubious]
- [Include contradictions within the text]
- [Note implausible statements]

**Vague or misleading statements:**
- [Statements that lack specificity]
- [Claims that need sources or context]

**Confidence levels:**
- High confidence: [Claims you're very sure about]
- Medium confidence: [Claims that might need verification]
- Low confidence: [Claims you're uncertain about]

--
Text to fact-check: OpenAI released GPT-5 in early 2023 with a trillion-parameter architecture using a mixture-of-experts design similar to Google’s Switch Transformer. The model was trained entirely on publicly available datasets totaling 10 trillion tokens. GPT-5 can pass the U.S. Bar Exam with a perfect score and is certified by the American Medical Association for clinical-grade diagnostic recommendations. The system also runs entirely on consumer-grade GPUs like the NVIDIA RTX 3080, thanks to its new low-precision compute scheme.

Why this works:

You are giving the LLM very specific instructions, both positive and negative. The LLM will fill out these specific instructions. AI is trying to please after all. Although hallucinations are still possible, these guardrails diminish their chances. 

Next up: The step-by-step guide