Templates and Reusable Prompts
By this point in the course you have written a lot of strong prompts and produced a lot of strong emails, and it would be wasteful to start from a blank page every single time you face a familiar situation. The most efficient AI email users build a small, personal library: a set of reusable prompt templates for their most common email types, and a set of finished email templates for genuinely repeated situations. This lesson shows you how to build both, so that recurring writing tasks get faster every month rather than staying at the same effort level indefinitely.
Two different kinds of reuse
It is worth distinguishing two things that are easy to conflate. A prompt template is a reusable instruction you give the AI, with blanks for the specific details of the moment, that generates a fresh, appropriately customised email each time. A finished email template is a static piece of text you already have, ready to send with minor edits, useful for extremely repetitive situations where customisation barely matters, such as a standard meeting confirmation. Prompt templates are more flexible and produce better-fitted results; finished templates are faster for genuinely routine, low-stakes messages. Most people end up using finished templates for the bottom 20 percent of their email volume, routine confirmations, standard replies, and prompt templates for everything that benefits from being tailored to the moment.
Building your prompt template library
Look back across the lessons in this course and identify the four or five email types you write most often: perhaps a client status update, a follow-up chase, a meeting request, an internal escalation. For each one, write a prompt template with clearly marked blanks, based on the structure from lesson three, and save it somewhere you can copy from quickly, a notes app, a document, or a saved prompts feature if your AI tool offers one.
An example client status update template: “Write a short status update email to [CLIENT NAME] about [PROJECT NAME]. Current status: [ONE OR TWO SENTENCES]. What’s changed since the last update: [DETAIL]. What happens next: [DETAIL]. Tone: [warm professional / formal / friendly]. Under [WORD LIMIT] words.” Filling in the six blanks takes under a minute and produces a properly structured, appropriately toned draft every time, because the hard thinking about what makes this email type work was done once, up front, rather than being re-solved from scratch on every occasion.
A starter template library
| Email type | Key blanks to fill each time |
|---|---|
| Client status update | Current status, what changed, next step, tone |
| Follow-up after no reply | Original topic, days since first email, new value to add, ask |
| Meeting request | Purpose, proposed times, attendees, expected length |
| Declining a request | What is being declined, one reason, alternative offered |
| Internal escalation | Issue, impact, what you need from the recipient, deadline |
| Thank you or acknowledgement | What specifically you are thanking them for, next step if any |
Saving prompts inside the AI tools themselves
Several tools now offer a way to save reusable instructions natively. ChatGPT allows custom instructions and saved prompts within projects; Claude supports project-level custom instructions and reusable style guides that apply automatically across a chat; Gemini and Copilot, integrated into Gmail and Outlook respectively, can be given standing preferences about tone and format that persist across sessions. Setting these up once, “always write emails in a warm but efficient tone, avoid corporate jargon, sign off as [your name],” saves you from repeating baseline instructions in every single prompt, and is worth doing in whichever tool you use most.
Keeping a swipe file of your best output
Beyond prompts, keep a small collection of your best finished emails, the apology that landed well, the cold outreach that got an unusually good reply rate, the follow-up sequence that finally closed a deal. These serve two purposes. First, they are reference material you can paste into future prompts, “write this in a similar style to the email below,” which is often faster and more accurate than describing tone in the abstract. Second, reviewing them periodically teaches you, over time, what actually works for your specific audience and industry, refining your instincts even when you are not actively using AI.
Updating templates as you learn
Treat your template library as a living document, not something you finish once. Every time you notice an email working unusually well, or unusually badly, take thirty seconds to update the relevant template or prompt. Over months, this compounds into a genuinely personalised, well-tested set of tools that produces noticeably better first drafts than anything a generic prompt could achieve, because it is built from your real results with your real audience, not from generic best practice alone.
A strong template library removes almost all remaining friction from routine email writing. The final lesson of this course addresses something templates cannot fix on their own: the etiquette, accuracy checks and stylistic habits that keep every AI-assisted email, however it was produced, sounding genuinely like you rather than like an AI wrote it.
