Module 4: Real-World Applications

Theory is only useful when it translates to action. This module walks through two of the most common and high-value use cases for prompt engineering in the workplace: automating communication and cleaning data.

Lesson 4.1 — Automating Email and Communication

Email is one of the highest-ROI applications of prompt engineering. The average professional spends 2–3 hours per day on email. AI can reduce that by 50–70% for routine communications.

Types of Emails AI Handles Well

• Follow-up emails after meetings or demos
• Client status updates and progress reports
• Internal announcements and team updates
• Proposal and quote cover letters
• Customer service responses to common queries
• Rejection or decline emails (handled diplomatically)

The Email Prompt Template

Use this template for any professional email:

“You are a [role]. Write a [type of email] to [recipient name/title]. The purpose is [specific goal]. Key points to include: [list 3–5 bullet points]. Tone: [formal/professional/warm]. Length: [word count or sentence count]. Do not include [anything to exclude].”

Case Study: Customer Service Response

A customer has complained that their order arrived 5 days late and one item was missing. They are frustrated and have threatened to leave a negative review.

Prompt: “You are a customer service manager. Write an empathetic response to a customer whose order arrived 5 days late and had one missing item. Acknowledge the problem, apologise sincerely, offer a replacement for the missing item and a 15% discount on their next order, and invite them to contact us directly if they have further concerns. Tone: warm, professional, and solution-focused. Length: under 150 words.”

Batch Processing: Responding to Multiple Emails

For high-volume email tasks, use few-shot prompting to process multiple emails in a single prompt. Provide 2–3 example responses, then paste the emails you want processed and ask the AI to follow the same pattern.

Lesson 4.2 — Data Cleaning and Formatting

Data cleaning is one of the most time-consuming tasks in any data-heavy role. AI can dramatically accelerate this work — especially for text-based cleaning tasks that would normally require manual review or complex formulas.

What AI Can Clean

• Inconsistent company names (“J. Smith & Co.”, “J Smith and Co”, “John Smith Co.” → standardise)
• Mixed date formats (“01/05/2025”, “May 1 2025”, “2025-05-01” → ISO format)
• Inconsistent address formats
• Duplicate entries with slight variations
• Extracting structured data from unstructured text (e.g., pulling names and amounts from invoice descriptions)

The Data Cleaning Prompt Template

“I have a list of [data type] with inconsistent formatting. Please standardise all entries to follow this format: [describe the target format]. Here is the data: [paste data]. Return only the cleaned data in the same order, one entry per line.”

Example: Standardising Company Names

“I have a list of company names with inconsistent formatting. Please standardise all entries to use the full legal name format (e.g., ‘John Smith Inc.’ not ‘J. Smith’ or ‘John Smith’). Return only the standardised names, one per line.

Data: J. Smith & Co., John Smith Inc, J Smith and Co., John Smith Incorporated, J.S. Co.”

Important Caveat

Always review AI-cleaned data before using it in production. AI is excellent at pattern-based cleaning but can make errors with ambiguous entries. Use it to do 80% of the work, then do a final human review of edge cases.