Workshop 1. Generative AI Foundations: Prompts, Privacy & Practical Applications
Welcome to this comprehensive course on generative AI for professionals. Over the next several weeks, you'll learn how to effectively utilize AI tools like ChatGPT in your workplace while maintaining proper privacy and security protocols.
This presentation will guide you through understanding AI fundamentals, crafting effective prompts, and implementing practical applications that can transform your workflow.
We'll begin with an orientation to AI concepts, explore prompt engineering techniques, and establish critical privacy foundations. By the end of this course, you'll be equipped with the knowledge and skills to leverage generative AI responsibly and productively in your professional environment.

by Haojun See

On The Ground - AI Adoption for Businesses and Individuals
What is Generative AI and How It Works
Content Creation
Generative AI tools can create various forms of content including text, images, and code based on your prompts and instructions.
Pattern Learning
These systems are trained on massive datasets including books, emails, and code repositories to recognize patterns in language and content.
Word Prediction
Large Language Models like ChatGPT function by predicting the next most likely word or token in a sequence, creating coherent outputs.
Pattern Prediction
AI doesn't truly "understand" content—it predicts plausible patterns based on its training data, which can sometimes lead to limitations.
Common Use Cases in Workplaces
Drafting Content
Create first drafts of emails, reports, policies, and social media posts to accelerate your writing process and overcome blank page syndrome.
Summarizing Information
Condense lengthy memos, meeting transcripts, and articles into concise, actionable summaries that capture key points efficiently.
Rewriting Content
Adjust the tone of communications, simplify complex legal text, or reformat content to better suit your audience and purpose.
Extracting Key Information
Identify and pull out action items, deadlines, and potential issues from lengthy documents to improve workflow management.
Data Privacy & Cybersecurity
No Sensitive Data
Never input confidential client information, personal data, or proprietary business information into public AI tools. This includes real names, strategy documents, and financial data.
Data Retention Risks
Be aware that many GPT tools may retain data for training purposes unless you've explicitly opted out. Always check the privacy policy of any AI tool before using it for work-related tasks.
Secure Alternatives
For sensitive work, prefer zero-data retention tools or self-hosted AI models that keep your data within your organization's security perimeter.
Prompting Is a Skill
What is a Prompt?
A prompt is essentially your instruction to the AI system. It's the input you provide to guide the AI in generating the output you desire. Like any communication skill, effective prompting can be learned and refined over time.
Characteristics of Good Prompts
The most effective prompts are clear, specific, and instructive. They leave little room for misinterpretation and guide the AI toward producing precisely what you need.
What GPT Needs to Know
For optimal results, your prompt should tell the AI who it should be (the role), what you want it to do (the task), and how to format or style the response (the output specifications).
Anatomy of a Good Prompt
Role
Define who the AI should act as (e.g., "You are a legal assistant" or "Act as a marketing specialist")
Task
Specify what you want the AI to do (e.g., "Summarize this text" or "Create a social media post")
Format
Indicate how you want the response structured (e.g., "in 3 bullet points" or "as a table")
Style
Clarify the tone, language level, or stylistic approach (e.g., "using plain language" or "in a professional tone")
Prompt Demo – Weak vs Strong
Weak Prompt
"Summarize this."
This prompt lacks specificity, format instructions, and style guidance. The AI has to guess what kind of summary you want, resulting in inconsistent and potentially unhelpful outputs.
Strong Prompt
"Summarize this in 3 bullet points, each under 12 words. Focus on deadlines and deliverables."
This prompt provides clear formatting instructions (3 bullet points), length constraints (under 12 words each), and content focus (deadlines and deliverables), leading to more useful results.
Discussion Points
What makes the second prompt more effective? Consider how specificity, format guidelines, and content focus all contribute to getting better AI-generated outputs.
Live Prompt Walkthrough
Input Selection
We'll start with a short legal paragraph or email as our source material. This could be a contract clause, client communication, or procedural document from your industry.
Three-Way Prompting
We'll prompt the AI in three different ways to demonstrate versatility: 1. Summarize - Condensing the key points 2. Rephrase - Changing the style or tone 3. Extract - Pulling out specific information
Output Comparison
We'll compare the results live, analyzing how different prompt approaches yield different results from the same input text, and discussing the strengths and weaknesses of each approach.
Understanding GPT Outputs
Pattern Imitation
GPT doesn't truly think or reason—it imitates patterns it has observed in its training data, making statistical predictions about what text should come next.
Hallucinations
The AI may generate plausible-sounding but false information, including fabricated dates, citations, or details that weren't in the source material.
Context Limitations
AI may miss important context or nuance, especially in specialized fields or when dealing with implicit information that humans would understand.
Human Review
Always fact-check and review AI-generated content, particularly for important communications or when accuracy is critical to your work.
Real-World Examples
Law
Contract clause simplification: Transform dense legal language into plain English explanations that clients can easily understand.
Deadline trackers: Extract all dates, deadlines, and filing requirements from lengthy legal documents into a concise timeline.
Small & Medium Enterprises
Standard Operating Procedure (SOP) generation: Create detailed, step-by-step instructions for common business processes.
Event summaries: Quickly generate recaps of client meetings or networking events with action items highlighted.
Client Communications
Follow-up emails: Draft personalized client follow-up messages based on meeting notes or discussion points.
Status updates: Create professional progress reports that highlight achievements and next steps in ongoing projects.
Quick Prompt Patterns to Master
"Summarize in..."
Ask the AI to condense information into a specified format: "Summarize this report in 3 paragraphs focusing on financial implications" or "Summarize this meeting transcript as 5 key decisions."
"Rewrite as..."
Request content to be restyled: "Rewrite this email in a more empathetic tone" or "Rewrite this technical explanation for a non-technical audience."
"Extract key points from..."
Pull specific information: "Extract all action items from this meeting transcript" or "Extract regulatory requirements from this policy document."
"Turn this into a checklist/FAQ..."
Convert formats: "Turn these project notes into a sequential checklist" or "Convert this product information into an FAQ format for customers."
Guided Prompting Exercise
Select a Work Document
Choose a relevant work document such as an email, paragraph from a report, or a standard operating procedure that you commonly use. For privacy reasons, use a sanitized version with no sensitive information.
Task 1: Summarize
Create a prompt asking the AI to summarize your document in exactly 3 bullet points. Try to be specific about what aspects should be emphasized in the summary.
Task 2: Rephrase
Ask the AI to rephrase your document in a friendlier, more conversational tone while maintaining all the key information and professional standards.
Task 3: Extract
Prompt the AI to extract all action steps, responsibilities, or deadlines from your document and present them in a clear, organized format.
Group Discussion Prompts

What worked well?
Share successful prompting techniques and outputs that exceeded expectations
What surprised you?
Discuss unexpected responses or capabilities you discovered
What would you change?
Explore prompt improvements and lessons learned
Use these discussion points to reflect on your experience with the guided prompting exercise. Share your screens to compare different approaches and results. Consider how you might apply these techniques to specific challenges in your role and industry.
Red Flags in Prompting
Requesting Definitive Expertise
Don't ask GPT to give legal advice, medical diagnoses, or other definitive expertise that requires professional certification. AI lacks the credentials, up-to-date knowledge, and professional liability to provide such guidance.
Vague Instructions
Always specify your desired format, tone, and key elements. Without clear specifications, you'll get inconsistent results that require more revision time. Be especially detailed about bullet points, section headers, or other structural elements.
Accepting First Outputs
When an output is unclear or off-target, don't just accept it or start over. Instead, reframe your request with more specific guidance. Iterative prompting often produces dramatically better results with minimal additional effort.
Data Risk Reminders
Never Paste Sensitive Content
Avoid inputting actual contracts, client names, financial data, or other sensitive information into public AI tools. This prevents both data leakage and potential breaches of confidentiality agreements.
Use Dummy Content
When working with templates or processes, substitute generic information or redacted sections. Replace real names with "Client A" or "[REDACTED]" and use placeholder values for sensitive figures.
Assume Outputs Could Be Reviewed
Always operate under the assumption that AI outputs might be seen by others or retained in some form. Don't let AI generate content that would be problematic if made public.
Final Practice Prompt
Task Description
For this final practice exercise, you will convert a client instruction email into an actionable checklist that can be used to track progress and responsibilities.
Select a typical email or message that contains multiple requests or instructions. This could be from a client, manager, or colleague. Remember to use a sanitized version with no sensitive information.
Example Prompt
"Turn this text into a 5-step checklist with responsible parties. Format each item as [Task] - [Person]. Add estimated time for each task in parentheses. Sort items by deadline, earliest first."
Notice how this prompt specifies:
  • The desired output format (checklist)
  • The number of items (5 steps)
  • Additional fields to include (responsible parties, time estimates)
  • Sorting criteria (by deadline)
Mini-Sprint Briefing
Apply
Use all three prompt styles (summarize, rewrite, extract) on a single work document
Analyze
Evaluate which prompt was most effective for your specific needs
Document
Capture before/after versions and your insights
Reflect
Write a one-page reflection on your experience and lessons learned
This mini-sprint will help you apply what you've learned in a practical context relevant to your work. The goal is not just to practice these techniques but to critically evaluate which approaches yield the best results for different types of content and purposes.
Recap and Next Steps
We've covered the fundamentals of generative AI, effective prompting techniques, and critical privacy considerations. These building blocks will serve as the foundation for next week's session, where we'll focus on finding high-value use cases in your specific work context and creating customized workflow integrations.
Reflection Questions (Optional Homework)
1
Efficiency Opportunities
What tasks can AI help you complete faster? Identify repetitive writing, research, or information processing activities in your workflow.
2
Human Advantage
What aspects of your job can AI never replace? Consider elements requiring judgment, creativity, empathy, or specialized expertise.
3
Automation Wishlist
What workflow do you wish you didn't have to do manually? Describe a time-consuming process that might benefit from AI assistance.
These reflection questions are designed to help you identify your highest-value opportunities for AI integration. By thinking deeply about your work processes, you can target areas where AI can have the most significant impact while preserving the uniquely human elements of your role.
Prompt Journal Template & Week Check-In
Daily Prompt Journal
To accelerate your learning, consider keeping a brief prompt journal with these fields:
  • Date: When you used the prompt
  • Task: What you were trying to accomplish
  • Prompt Used: Your exact instruction to the AI
  • Notes on Output: Quality assessment and observations
  • What to Improve: Ideas for better prompts next time
Just 5 minutes of daily journaling can dramatically improve your prompting skills through deliberate practice.
End-of-Week Check-In Questions
As you practice during the week, consider these reflection points:
  • "What's one AI win you had this week?"
  • "What do you want to try next with ChatGPT?"
  • "Where did you get stuck or feel limited?"
  • "How much time did AI help you save?"
These questions help track your progress and identify both successes and areas for improvement as you develop your AI skills.
Contact Us
We hope you found this workshop valuable as you begin your AI journey. We're excited to continue exploring practical applications in our next session.
Have questions or need assistance before then? Reach out anytime:
WhatsApp: wa.me/6598419481
Thank you for your participation and thoughtful engagement today!