Course Home Lesson 9: Build Reusable Prompt Patterns, Not One-Off Requests

Lesson 9: Build Reusable Prompt Patterns, Not One-Off Requests

Lesson overview

The highest-value use of good prompting is not one well-crafted request. It is a library of well-crafted patterns you can use and adapt for recurring tasks. This lesson shows you how to identify those patterns, build them into reusable templates, and use them to improve both speed and consistency across your work.


What this means

A prompt pattern is a starter structure for a common task. It is not a rigid script — it is a framework with placeholders for the things that change from use to use.

A one-off prompt works once. A prompt pattern works every time you have that type of task.


Why it matters

Most people who use AI regularly are doing the same types of tasks repeatedly: writing meeting summaries, creating project updates, drafting emails, summarizing research, writing documentation, building outlines. Each of these tasks has consistent elements that do not change — and variable elements that do.

A prompt template captures the consistent elements in advance. You fill in the variables when you need it. The result is faster, more consistent output without the overhead of building a well-framed prompt from scratch every single time.

Teams that share prompt templates produce more consistent output across members. This matters when voice, quality, and framing need to be aligned.


What most people do wrong

Building prompts from scratch every time

If you are writing a new prompt for the same type of task every week, you are doing more work than necessary. The effort of crafting a good prompt should compound — not reset.

Saving finished outputs instead of the prompts that produced them

People often save good AI output. They rarely save the prompt that produced it. The prompt is more valuable: it is reusable.

Making templates too rigid

A template that tries to anticipate everything becomes a form to fill out. Good templates have enough structure to guide the output but enough flexibility to handle different situations.

Not iterating on templates

A first-version template is never perfect. After using it a few times, you will notice patterns in what needs to be adjusted. Updating the template based on those patterns is how it gets better.


How to build a prompt template

Step 1: Identify repeating tasks
Look at the last month of work where you used AI. What types of tasks came up more than twice? Those are template candidates.

Step 2: Identify what stays the same
What is consistent across every instance of that task? Audience? Format? Tone? Quality standard? Structure? These become the fixed parts of your template.

Step 3: Identify what changes
What is different each time? Topic? Specific content? Date? Participants? These become the variable placeholders.

Step 4: Write the template with clear placeholders
Use brackets or descriptive labels to mark where values get filled in. Keep the variable labels descriptive enough that anyone can use the template.

Step 5: Test it a few times and adjust
Run the template on two or three real tasks. Note what you still have to change. Adjust the template accordingly.


Example templates

Meeting summary template

Summarize the following meeting notes into a clear internal summary.

Context: [Brief description of what the meeting was about and who attended]
Audience: [Who will read this summary — e.g., team members who attended, stakeholders who did not]
Format: Decisions made, action items with owners, key open questions. Use a simple three-section structure.
Tone: Direct and factual. No narrative framing. Just what was decided and what happens next.
Length: Under 300 words.

Meeting notes: [Paste notes here]


Research summary template

Summarize the following research or source material for a specific purpose.

Context: [What the research is about and why it is being summarized]
Audience: [Who will use this summary and what will they do with it]
Focus: [What aspect is most relevant — e.g., risks, opportunities, key findings, competitive positioning]
Exclude: [What is out of scope for this summary]
Format: [Structured bullets / short paragraphs / executive summary style]
Length: [Word or sentence limit]

Source material: [Paste content here]


Rewrite / edit template

Rewrite the following draft to improve it.

What it is: [Brief description of what this piece is for]
Audience: [Who will read it]
Current problem: [What is wrong with the draft — e.g., too formal, padded, unclear, wrong structure]
Target tone: [How it should feel]
Constraints: [Length, format, what to keep, what to cut]

Draft: [Paste draft here]


Practical exercise

Identify one task you do repeatedly that involves AI. It should be something you have done at least three times in the past month.

Write a reusable prompt template for that task using the four-step process above. Include at least three fixed sections and at least two variable placeholders.

Run the template on one real task and note what you had to adjust. Update the template.


Reflection prompt

  1. What are the three most common tasks you use AI for? Do you have a reusable pattern for any of them?
  2. If you were to share your best AI prompt with a new teammate, what would you need to explain about how to use it?
  3. What would change about your workflow if you had templates ready for your five most common tasks?

Key takeaway

The best prompts compound. When you build reusable templates instead of one-off requests, you spend less time setting up and more time doing. Templates also make it easier to share strong AI workflows across a team.

Working Well With AI · Practical AI training for real work