Stop Re-Explaining Yourself to AI: Use a Context File for Better Training Design Outputs

If you use AI for course design, you have probably run into this: you open ChatGPT, Claude, or Copilot and ask for help designing a module. And every time, you get back something that looks like this:

Section 1: Introduction. Section 2: Key Concepts. Section 3: Review Questions.

Content dump, followed by a knowledge check. Every. Single. Time.

The content might be decent, but the structure defaults to lecture because that is what AI has seen the most of. For training developers who work with case studies, scenario-based learning, and experiential frameworks, this means spending as much time editing AI outputs as it would have taken to write the thing from scratch.

There is a straightforward fix for this: a context file. Here is how it works and how to set one up in whatever AI tool you are already using.

What Is a Context File (MD File)?

A context file — often saved as a markdown (.md) file — is a plain text document that tells AI exactly how you design training. Think of it as a standing briefing document: your methodology, your preferred formats, your audience, and your non-negotiables, all in one place.

You write it once. Then you upload or paste it at the start of any AI session. Instead of AI guessing at your design approach, it now has clear instructions from the very beginning. The result is a first draft that actually reflects how you work, not how a generic course template works.

What to Put in Your Context File

Your context file does not need to be long or perfectly formatted. The goal is to give AI the information it keeps getting wrong. For training developers, the highest-impact content to include is:

• Your design philosophy — the adult learning principles you apply to every course. If you use experiential learning frameworks, scenario-based design, or problem-first structures, describe them here.

• Your e-learning preferences — how you want modules structured, the role of branching scenarios, and — critically — what you never want AI to default to.

• Your classroom lesson plan format — AI defaults to a different format every time if you do not specify. Include your exact structure so you stop rebuilding it from scratch.

• Your audience profile — who your learners are, their experience level, and the operational context they work in. For law enforcement training, this context changes what counts as a useful example.

• Output expectations — what a good first draft looks like, so AI is not guessing.

You can build this file yourself, or let AI draft it for you. Open a new session and say: I want to create a context file for my training design work. Ask me questions about my methodology, my audience, and my preferences, then format it as a markdown file. Most tools will guide you through the questions and produce a usable first draft.

How to Use It in the Tools You Already Have

Claude

Upload your context file at the start of any conversation, or add it to a Claude Project so it loads automatically every time. Projects are free and persist across sessions — you can add multiple reference documents, so your design preferences are always in context without copying and pasting.

ChatGPT

Upload the file directly into any conversation. For a more permanent setup, paste the content into your Custom Instructions (Settings > Customize ChatGPT) so it applies to every session automatically. You can also build it into a Custom GPT designed specifically for training development.

Gemini

Attach your context file at the start of a conversation, or create a Gem — Gemini's version of a custom AI assistant — and add your file as a reference document so it is available every time you open that workspace.

Microsoft Copilot

Paste the content directly into a chat or upload a Word document version. For a more integrated setup, Copilot Studio lets you build a custom agent with your file as a persistent knowledge source. If your department uses SharePoint, storing the file there and referencing it in Copilot for Microsoft 365 keeps everything in one place.

Getting Started Without Overthinking It

You do not need to build the perfect file before you start. Focus on the two things causing you the most frustration with AI outputs — that is where you will get the fastest return.

1. Open a plain text editor — Notepad, VS Code, or even Word will do.

2. Write a clear description of how you want AI to handle your biggest frustration. Be specific about formats, structures, and what you do not want.

3. Save it as a .md or .txt file — the format matters less than the content.

4. Paste it at the start of your next AI session before you give your actual request.

The difference in output quality is immediate. When AI actually understands your design methodology, you stop functioning as a course editor and start working as a course developer again.

Want to Build These Skills With Your Team?

If you want your training team to start getting real results from AI — not just experimenting — Odin Training Solutions offers private 4-hour virtual workshops designed specifically for training developers. We work through your department's actual projects using multiple AI platforms, so participants leave with practical skills and working materials, not just theory.

Format: Private virtual sessions for up to 20 participants

Investment: $2,000 USD / $2,500 CDN per workshop

Reach out at kerry.avery@shaw.ca to discuss bringing this workshop to your team, or visit the workshops page on this site to learn more.