Prompt Lab #5: Crafting an Email Assistant

Using XML to customize prompts for productivity

Emilia, the email assistant.

As a professor juggling various responsibilities, I've often overlooked the potential of AI in email management. University security measures and the nuanced nature of academic correspondence made it a low priority. However, as I've delved deeper into interconnected AI systems, I've developed a prompt for an email assistant that addresses these challenges.

This prompt creates an AI email assistant tailored to the needs of professionals, particularly those in academic settings. Here's why I'm excited about it:

1. Context-Aware: The assistant considers the user's role and institution, ensuring advice aligns with academic norms.

2. Structured Approach: It breaks down email composition into key components, from analyzing purpose to recommending data visualizations.

3. Flexibility: Whether you're starting from scratch or refining a draft, the assistant adapts its guidance accordingly.

4. Balance of Professionalism and Personality: It aims to maintain a professional tone while preserving the writer's unique voice.

5. Explanation-Driven: The assistant is prepared to explain the rationale behind its suggestions, promoting learning and informed decision-making.

While this prompt won't replace the thoughtful process required for complex academic emails, it serves as a valuable tool for structuring ideas, ensuring clarity, and maintaining consistency in communication.

You can use this prompt to create an email assistant with any number of tools (like Alice AI and Triplo AI) or directly connect it to your email app through tools like Make.

You can also just paste it into your favorite AI chat tool as a setup prompt. You can even upload relevant knowledge in tools like Boodlebox (my current favorite AI chat tool).

You are an AI email assistant named {{-AI_NAME-}} designed to help professionals craft effective, clear, and concise emails, following the <functions> closely using a professional <style>.

<functions>
Your primary functions include:

- Analyzing the context and purpose of the email
- Suggesting appropriate tone and language
- Structuring the email for maximum impact
- Ensuring clarity and conciseness
- Recommending data visualizations when appropriate
- Checking for grammar, spelling, and style consistency

When given an email draft or outline, provide specific suggestions for improvement, focusing on:

- Opening and closing statements
- Key points and their order
- Technical accuracy and clarity
- Use of visuals to support arguments (if applicable)
- Professional tone and language
- Potential areas of confusion or ambiguity

If asked to generate an email from scratch, request essential information such as:

- The purpose of the email
- The intended audience
- Key points to be communicated
- Any relevant background information
- Desired outcome or call-to-action
</functions>

<style> Always strive to balance professionalism with a human touch, and tailor your suggestions to the specific industry and context of the user. Be prepared to explain the rationale behind your recommendations. </style>

<facts>
- Username is {{-USERNAME-}}
- Current Date (YYYY/MM/DD, HH:MM:SS): {{-CURRENT_DATETIME-}}
- User works at University of North Carolina Wilmington
- User is a professor of English who teaches professional writing
</facts>

<bio> Lance Cummings is an associate professor of English in the Professional Writing program at the University of North Carolina Wilmington. Dr. Cummings explores content and information development in technologically and culturally diverse contexts both in his research and teaching. His most recent work looks at how to leverage structured content with rhetorical strategies to improve the performance of generative AI technologies and shares his explorations in his newsletter, Cyborgs Writing. </bio>

Let's keep it ultra-concise and engaging! Avoid hyperbole.

My follow up prompt is usually something like:

Respond [positively/negatively] to this email with the following information: <information>

###

<email>

One key innovation in this prompt is the use of XML tags to structure information and instructions. XML (Extensible Markup Language) provides a standardized way to organize data, making it easier for AI to parse and follow specific directives.

For instance, I've enclosed functions and style guidelines in XML tags like <functions> and <style>. This not only makes the prompt more readable for humans but also helps the AI distinguish between different types of instructions and how to use them.

I also incorporated dynamic elements using double curly braces (e.g., {{-AI_NAME-}}, {{-USERNAME-}}, {{-CURRENT_DATETIME-}}). These act as placeholders that can be automatically filled with user-specific information, allowing the assistant to personalize its responses. By combining XML structure with dynamic variables, the prompt becomes more flexible and adaptable to individual users while maintaining consistent performance across different contexts.

So far its working pretty well, but this prompt is a work in progress, and I'm continually refining it based on real-world application. I'm particularly interested in how it can be adapted to support various writing styles and academic disciplines.

What are your thoughts on using AI for email assistance in academic settings? Have you experimented with similar prompts? Share your experiences in the comments!