How to 10x Your AI Prompting

a single context wrapper prompt for all your needs!

Sep 10, 2025

Welcome to another week of Teachnology. Context Engineering sounds complicated, but it’s not - in this post I’ll show you how to use AI + Context to craft the best prompts for getting stuff done. Adopting this approach for your organization (or your classroom) will result in consistent and outstanding AI responses that levels the playing field for all users. Also this week, more recommended reads at the bottom of the post. Enjoy!

My family did a local escape room over Christmas several years ago. We were pretty confident going in, knowing that the room was no match for our superior skills, intellect, and willpower. But nope, we needed help several times because we couldn’t find the hidden item or the secret doorway or solution to the puzzle. This is what AI prompting feels like for so many people - they don’t know the “secret sauce” to getting the results that others get through persistence and refinement. But first…

Rick’s AI Express

Of note in the past week:

Anthropic (Claude AI) secured $13 Billion US in funding, increasing its value to $183 Billion. Worldwide Venture Capital investment in AI companies accounted for 53% of all investment; US VC investment in AI was 64%. Of total investment dollars!
Last year it was determined that Google held an unfair monopoly on Search in the technology marketplace - like being the default on mobile phones. This week the same judge decided that Google did not have to sell off its Chrome browser, but had to share some of its data with competitors. This is a huge win for companies like Google and others in the Magnificent Seven, who already have monopolies.
OpenAI plans to compete directly with LinkedIn (owned by Microsoft) by creating its own jobs platform and certification programs. This represents a new phase in the AI talent acquisition market. Microsoft (perhaps in response?) decided to shift its Microsoft 365 AI needs from OpenAI to Anthropic. Your move, Sam…
Researchers at UCLA have developed a non-invasive brain-computer interface (BCI) that uses AI to help people control a computer or robotic arm with their mind. The technology decodes EEG signals to extract signals that correspond to movement intention. This is a direct competitor to Elon Musk’s Neuralink, and much less risky because no surgery is involved.
And lastly, OpenAI has confirmed that AI LLMs hallucinate because the system training rewards guessing more than admitting uncertainty. Okay, I think we all knew that, but the real question is “what are you gonna do about it?”

If my family had a map or a blueprint of the escape room, we might have found the way out on our own. We might have noticed missing books on the shelf, scratches on the floor, or answers to riddles and puzzles right in front of us. But we were blinded by our own overconfidence - I mean “How Hard Can This Be?”. Turns out, pretty hard…

AI prompting is a lot like that, and almost three years in there are those (perhaps your students, or your colleagues) who have mastered the skill of prompt iteration: carefully refining their initial request until they get the most useful or satisfying result. While many of us give up long before these AI power users, this has created a clear separation between those who can bend AI to their needs and those who continue to be frustrated or unimpressed with what AI offers.

Okay, let me get to the point with this post: Context Engineering allows everyone to get the same results as the most capable person in your classroom or organization. Great, but what is Context Engineering, why should you use it, and how do you get started?

What is Context Engineering?

In its simplest form, context engineering is the use of additional details to tell the AI model about its role and your goals, constraints, methodology, required output, tone, audience, and next steps when giving a prompt. It turns out that AI models perform much better when these considerations are explicitly stated.

Here’s an example of a prompt I use every week in Gemini to read my AI newsletters in Gmail, and to pick out several to use in my post. Summarize all of the emails in my gmail account that have the label “AI Newsletters”. I’m looking for content that is timely, insightful, and well written, with clear takeaways for readers. I will use some of these summaries in the AI Express section of my blog, and other summaries may become full posts at another time. Nothing wrong with that prompt, and I get pretty good results. However, compare that pretty good prompt with THIS PROMPT and you’ll see that there is a difference in the added context of the request - which means the results will be tailored to my needs.

If you’re thinking “that’s way too much work for me to do every time”, don’t worry; I’m going to show you how to get AI to do ALL of the heavy lifting. And in the end, you will have a SINGLE CONTEXT WRAPPER for any prompt that requires serious effort on your part. No more debating how to ask AI for help - this one really can do it all!

A Context Engineering (CE) prompt usually has a few predictable sections, kind of like scaffolding or a map that keeps the AI pointed in the right direction. The exact setup varies depending on the task, but here are the main sections you’ll typically see:

1. Role / Identity

Sets who the AI should act as, or what perspective it should adopt.
Example: “You are a middle school teacher preparing a digital citizenship talk.” Or “You are an HR professional, responsible for onboarding new employees”.
Why: Helps narrow the AI’s voice, priorities, and knowledge base.

2. Audience / Context

Defines who this is for and the situation.
Example: “Your audience is a group of 4th–6th grade students who already use social media.” Or “You are responsible for new employee workplace orientation - including tasks related to IT, payroll, policies, benefits, mentoring, and company culture.”
Why: Keeps responses age-appropriate, task focused, relevant, and tailored.

3. Task / Goal

States clearly what the AI needs to do.
Example: “Generate 5 key talking points with short, student-friendly examples.” Or “Generate an onboarding checklist for new employees with sign-offs for each major category.”
Why: Focuses the output so it doesn’t wander.

4. Inputs / Constraints

Lists the data, materials, or rules the AI must use or respect.
This is where you will want AI to ask questions about your basic prompt.
Example: “Use only positive examples, avoid scare tactics, keep sentences under 20 words.” Or “Ask the user for a list of specific items the new employee must complete within the first three days of employment.”
Why: Prevents generic answers and keeps the output practical.

5. Format / Output Requirements

Specifies the shape of the response.
Example: “Return as a numbered list with bolded headers and a one-sentence explanation each.” Or “Format as a table with one row per category, and each cell containing an action to complete.”
Why: Saves editing later by getting the right format upfront.

6. Extras (Optional)

Style or tone: “Conversational but not childish.” Or “Professional but friendly.”
Examples/analogies: “Use simple metaphors students will understand.”
Reflection/iteration: “At the end, suggest 2 follow-up questions I could ask”

The beauty of using AI to create this type of CE prompt is that you don’t have to provide all of the information; give what you have and ask AI to make sure it includes the sections listed above. If it doesn’t have enough information, it will ask for your input.

Why should your organization or classroom adopt Context Engineering?

That’s the big question: “Why should you use Context Engineering?” On the one hand, if you are already very good at refining your AI prompts to get the results you want, you might not see the need for Context Engineering. However, as someone who is pretty AI-savvy, this approach to prompting has quickly become my go-to tactic. Think of this as giving everyone the same easy-to-use, professional-grade tools to get the job done; Everyone can use the same prompt wrapper to analyze resumes or financial documents, to make existing assignments AI-resistant, to explain any topic at a grade level and analogy that is appropriate. No more guessing if AI is going to work this time; it will, because others are using the same approach successfully. And everyone can get the same excellent results that were formerly the domain of a few power users.

Here’s an example that I have previously used in ChatGPT: Analyze the attached assignment and suggest ways to make it AI-resistant. Usually I need to refine the prompt several times in order to get the result I want, but I’m used to the additional work of prompt iteration. What if we can make THIS PROMPT available to all teachers, and it will give more uniform and better results.

How to start using Context Engineering?

I suggest you ask yourself (or your friends, coworkers, employees, students) what regular tasks you need help with, and build a list of 5-10 specific use cases that AI can assist in. I’ll list examples from both the classroom and the corporate world below; however, you know your use cases best. The goal is to use a Context Engineering wrapper to develop each use case into a comprehensive prompt that will get the job done well, with high quality results, regardless of who is doing the prompting. Then you just reuse these prompts each time you have a question. Simple!

Teachers and Students:

Plan a unit based on your provincial / state curriculum guidelines. A typical prompt might be “help me create a unit for 8th graders in Ontario studying ecosystems”.
Review a specific concept or skill that a student needs help with. Don’t just give the answer, but make the student do most of the hard work in understanding and applying what has been taught. Give appropriate hints and review concepts they seem to have missed, but don’t give shortcuts.
Explore cross-curricular integrations for a specific lesson or unit. Maybe you’re wondering how to reinforce a topic in other subjects that you teach - for example, “We are studying the Romans in history; examine the role and impact of the Romans in other 7th grade subjects”
Assess student work in ways that are both appropriate and challenging for their grade level, age, diagnosed or undiagnosed learning exceptionality, and strengths.
Help a student edit their own work before handing it in. While a student might be tempted to ask AI to “turn this into an A+ essay”, shift the hard work back to the student - where it belongs.

Individuals and Businesses:

Compare end-of-year financial statements to last year’s, or monthly cash flow to last month’s.
Modify a résumé to align with job posting, or filter résumés to identify potential hires.
Ensure compliance with annual updates to employer and provincial/state policies.
New client or employee onboarding process.
Streamlining internal reports and meeting notes.

Here is the Prompt Wrapper that you can fill in - just replace [Task to be Accomplished] with your usual AI question. Then enter that entire prompt into your favourite AI chat window. Yes, you will be asked for some details that you might not have thought about when crafting your original prompt, but that ensures that the results will be tailored to your needs.

Instruction to AI:
The purpose of this prompt is to act as a wrapper, so that a generic chat question becomes finely tuned to the needs of the user or organization.
Please take the following request and turn it into a full Context Engineering (CE) prompt:

[Task to be Accomplished]

Steps for You (the AI):
Ask clarifying questions first so you fully understand the task before generating the CE prompt.
Once clarified, rewrite the task as a CE prompt using the following structure:

1.Role / Identity - Who you should act as in this task.
2.Goal / Task - What needs to be accomplished.
3.Context / Audience - Who the output is for, and the situation in which it will be used.
4.Constraints and Inputs - Resources, data, or limitations to use.
If inputs are missing, ask me to provide them.
If you must make assumptions, state them explicitly and mark them as provisional.
5.Format and Output Requirements - The structure, style, or format the response should take.
Maintain all six CE sections explicitly, regardless of model tendencies to shorten or merge them.
6.Extras - Tone, examples, analogies, follow-up questions, or refinements.
At the end of the CE prompt, ask me to confirm whether it meets my needs or requires adjustment.

How do we know if this Prompt Wrapper stuff works? Try it yourself. Do the following:

take your original prompt - without my wrapper - and enter it in your favourite AI chat window.
add your prompt to my prompt wrapper and enter it in a new chat window (use the same AI, so that we’re comparing apples and apples).
you decide which prompt gives better results for your use case.
try it with different AI models - I used it with Gemini, Claude, Perplexity, and ChatGPT. You might have noticed the line in the wrapper about “model tendencies”. This ensures that you will get similar results from any of the AI models.

If you have specific use cases you’d like to explore, please contact me. I’m currently using this prompt wrapper for all questions that requires a more detailed or nuanced response. The immediate benefit is that I don’t have to figure out a method for saving and organizing the hundreds of prompts that I used in the past; I have one prompt and I use it for almost all of my searches. How cool is that?

And with that, here are this week’s Recommended Reads - a list of the top articles I’ve read this week that you will inform, challenge, encourage, or inspire you. Enjoy!

Rick’s Teachnology Blog

Discussion about this post