What Is AI, Really?
Cut through the hype. How generative AI works, what it's good at, and where it falls short.
Module 1 Check-In: What Is AI, Really?
1. How does a large language model (LLM) like ChatGPT produce its responses?
2. Which of the following is an example of a task AI is genuinely good at today?
3. What is "hallucination" in the context of AI?
4. Generative AI is best described as a tool that:
What are you?
Explain what you are in one paragraph, as if you're talking to a curious 10-year-old. Then list three things you genuinely can't do well.
► TRY IT NOW: > > Open Claude (claude.ai) or ChatGPT > > Ask AI to explain what it is, in one paragraph, to a 10-year-old > > Then ask: "What are three things you can't do?" > > Notice what it says — and how confidently it says it
What are you?
Explain what you are in one paragraph, as if you're talking to a curious 10-year-old. Then list three things you genuinely can't do well.
What are the biggest AI trends?
What are the three biggest AI trends affecting [your industry] in the next 12 months? For each trend, give a specific example and one action a professional in this field should take in response.
► TRY IT NOW: > > Try the same question in two different AI tools (ChatGPT and Claude) > > Ask: "What are the three biggest AI trends?" > > Compare the responses — what's different? What's similar? > > Which felt more useful and why?
What are the biggest AI trends?
What are the three biggest AI trends affecting [your industry] in the next 12 months? For each trend, give a specific example and one action a professional in this field should take in response.
Practice Scenarios
You're a customer service rep. Ask AI to explain your company's return policy to a frustrated customer.
A customer is frustrated about a delayed order. Explain our return policy in plain English. Be empathetic but clear.
You're a manager. Ask AI to explain your team's quarterly goals to a new hire in their first week.
A new hire just joined. Explain our team's quarterly goals and top 3 priorities for them in their first 30 days.
The 4D Framework Overview
Your mental model for using AI intentionally - Delegation, Description, Discernment, Diligence.
Module 2 Check-In: The 4D Framework
1. What are the four D's in the 4D Framework?
2. The 4D Framework is primarily designed to help you:
3. Which D comes first in the practical workflow?
Delegation
Knowing what to hand off to AI and what to keep for yourself. The delegation tier list.
Module 3 Check-In: Delegation
1. Which task is the best candidate for full AI delegation?
2. A task is a poor candidate for AI delegation when it:
3. The "delegation tier list" concept means:
Write a status update email
Write a weekly status update email for my team. Project: [name]. This week we completed: [list]. Next week we're focused on: [list]. Any blockers: [list or "none"]. Audience: my manager. Tone: clear and direct. Max 200 words.
► TRY IT NOW: > > List 5 tasks you do regularly that involve writing, summarizing, or explaining > > Pick your top Tier 1 task and delegate it to AI right now > > Give it as much context as you can — don't overthink the prompt yet > > We'll improve the prompt in the next module
Write a status update email
Write a weekly status update email for my team. Project: [name]. This week we completed: [list]. Next week we're focused on: [list]. Any blockers: [list or "none"]. Audience: my manager. Tone: clear and direct. Max 200 words.
Give me 5 AI task ideas
I'm a [your role] at a [your industry/company type]. I spend [X hours/week] on [repetitive task]. Suggest 5 AI use cases ranked by time savings. For each: what to delegate, what context to include, and how to verify the output.
► TRY IT NOW: > > Pick three things from your typical workday > > Categorize each one: 'Give to AI', 'Do yourself', or 'Do with AI' > > For one 'Give to AI' task, write the pro-level prompt > > Notice how much context you need to add to make it useful
Give me 5 AI task ideas
I'm a [your role] at a [your industry/company type]. I spend [X hours/week] on [repetitive task]. Suggest 5 AI use cases ranked by time savings. For each: what to delegate, what context to include, and how to verify the output.
Practice Scenarios
You're a sales rep. Use AI to draft a personalized cold email to a prospect you researched on LinkedIn.
Draft a cold email to [prospect name], a VP of Marketing at [company]. They're a Series B SaaS company. Personalize it around their recent LinkedIn post about scaling teams.
You're an HR manager. Use AI to rewrite a generic job description for a senior role into one that attracts diverse candidates.
Rewrite this job description to attract diverse senior candidates. Remove biased language. Add inclusive framing. Keep the core requirements. [job description]
You're an operations lead. Use AI to turn meeting notes into an action-item summary with owners and deadlines.
Turn these meeting notes into an action-item summary with: task, owner, and deadline for each item. [meeting notes]
Description (Prompting)
Six prompting techniques that work. Writing prompts that get useful results the first time.
Module 4 Check-In: Description (Prompting)
1. Which of the following is the most effective prompt?
2. What is "role prompting"?
3. When an AI gives you a mediocre first response, the best next step is usually:
4. What does adding "context" to a prompt typically do?
Write an email about the project delay
Write a professional email to our client (mid-sized accounting firm) explaining we're delivering their Q4 report 3 days late due to a data integration issue. Tone: direct and accountable, not over-apologetic. Include: specific reason, revised delivery date of Friday Nov 8, and one concrete step we're taking to prevent recurrence. Max 150 words.
► TRY IT NOW: > > Take your prompt from Exercise 1 (the under-specified one) > > Rebuild it using all five elements: Task, Context, Audience, Tone, Constraints > > Run the new prompt and compare the outputs side by side > > Notice specifically what changed and why
Write an email about the project delay
Write a professional email to our client (mid-sized accounting firm) explaining we're delivering their Q4 report 3 days late due to a data integration issue. Tone: direct and accountable, not over-apologetic. Include: specific reason, revised delivery date of Friday Nov 8, and one concrete step we're taking to prevent recurrence. Max 150 words.
Rewrite this email
Edit this for a VP-level audience. Keep the key message. Tone: direct and confident. Remove anything that sounds defensive. Max 150 words. Add a clear subject line. [PASTE YOUR EMAIL HERE]
► TRY IT NOW: > > Pick any email you've sent this week > > Paste it into AI with the pro prompt template above > > Ask AI to rewrite it in three different tones: formal, conversational, persuasive > > Notice what changes and what stays the same
Rewrite this email
Edit this for a VP-level audience. Keep the key message. Tone: direct and confident. Remove anything that sounds defensive. Max 150 words. Add a clear subject line. [PASTE YOUR EMAIL HERE]
Practice Scenarios
You're a content writer. Ask AI to rewrite the same blog intro in three different tones: academic, conversational, and tabloid.
Rewrite this blog intro in three styles — academic, conversational, and tabloid — and explain what changed in each. [your intro]
You're a consultant. Use the 5-element prompt template to draft a client proposal section.
Write a 1-page client proposal section for [project name]. Client is a [type]. Budget: [range]. Timeline: [X weeks]. Tone: confident, not salesy. Include: scope, approach, and one relevant case study.
You're a recruiter. Ask AI to rewrite a screening email to candidates in a warm, human tone.
Rewrite this recruiter screening email in a warm, human tone. Keep it under 100 words. Make it feel like a real person wrote it. [current email]
Discernment
Evaluating AI output critically. Spotting confident mistakes. The description-discernment loop.
Module 5 Check-In: Discernment
1. AI output should always be:
2. The "description-discernment loop" refers to:
3. Which is the best indicator that an AI response needs more verification?
Tell me about AI regulations
Write a 200-word summary of current [your industry] regulations that affect [your role]. Include at least 3 specific statistics, dates, or regulatory references. I will be verifying each claim.
► TRY IT NOW: > > Give AI a task in your domain — something with specific facts required > > Ask it to include at least three specific data points > > Review the output with the editor's checklist > > Try to verify at least one claim
Tell me about AI regulations
Write a 200-word summary of current [your industry] regulations that affect [your role]. Include at least 3 specific statistics, dates, or regulatory references. I will be verifying each claim.
Summarize this article
Summarize [paste article URL or text] in 200 words for a [your role] audience. Highlight: (1) the main finding, (2) what it means for my work, (3) one thing I should verify before acting on it.
► TRY IT NOW: > > Find a Wikipedia article on a topic you know well > > Use the pro prompt to summarize it > > Read the summary carefully — find anything missing or wrong > > This is your discernment muscle — keep using it
Summarize this article
Summarize [paste article URL or text] in 200 words for a [your role] audience. Highlight: (1) the main finding, (2) what it means for my work, (3) one thing I should verify before acting on it.
Practice Scenarios
You're a market researcher. Ask AI to summarize a long industry report and flag claims that need verification.
Summarize this 10-page industry report in 250 words. Flag any specific statistics, company names, or dates as [VERIFY] — I'll check each one.
You're a financial analyst. Use AI to draft a brief on a company's earnings, then fact-check one claim.
Summarize [company name]'s most recent quarterly earnings in 200 words. Include revenue, key metrics, and one forward-looking statement. Flag any numbers that need verification.
Diligence
Using AI responsibly - privacy, attribution, transparency, and preserving your own skills.
Module 6 Check-In: Diligence
1. Which of the following should you NEVER paste into a public AI tool?
2. When you use AI to write something and publish it, best practice is to:
3. Using AI heavily for tasks you normally do yourself can create risk because:
4. The diligence principle in the 4D Framework primarily asks you to:
Write an AI policy for my team
I'm a [manager/director] at a [company type]. Help me write a 1-page AI usage policy for my team that covers: (1) what AI use is permitted and for what purposes, (2) what must never go into AI tools, (3) when and how to disclose AI use internally and externally. Tone: clear and professional. This is for [knowledge workers/sales team/operations].
► TRY IT NOW: > > Think of a task where you'd use AI at work > > Run it through three questions: (1) Sensitive data? (2) Disclosure needed? (3) Would the person affected be okay with how this was produced? > > Adjust your approach based on what you find
Write an AI policy for my team
I'm a [manager/director] at a [company type]. Help me write a 1-page AI usage policy for my team that covers: (1) what AI use is permitted and for what purposes, (2) what must never go into AI tools, (3) when and how to disclose AI use internally and externally. Tone: clear and professional. This is for [knowledge workers/sales team/operations].
Practice Scenarios
You're in legal. Apply the three ethics questions to using AI for contract review.
I want to use AI to help me review contracts for a small business client. Apply the three ethics questions: (1) sensitive data? (2) disclosure needed? (3) would the client be okay with this?
You're a marketing director. Evaluate using AI to generate customer personas for a campaign.
I want to use AI to generate customer personas for a B2B SaaS campaign. Evaluate: what data can I safely input, what should I keep private, and how should we disclose AI use in the final output?
Your 30-Day Plan
Putting it all together. Run the 4D Framework on real work tasks every day for 30 days.
Comprehensive Knowledge Check
10 questions covering all 7 modules. Score 70% or above to pass and earn your certificate.
1. What is the primary reason AI language models can "hallucinate"?
2. Which statement best describes the 4D Framework?
3. You need to draft a sensitive internal memo about an upcoming redundancy. What is the RIGHT approach?
4. What makes a prompt significantly more useful?
5. The description-discernment loop means you should:
6. Which task is best suited for AI delegation?
7. What is the correct response when AI produces specific statistics?
8. Role prompting is useful because:
9. Why does the 4D Framework include Diligence as its final D?
10. Which statement about generative AI is most accurate?
Your Certificate
Enter your name below to generate your certificate of completion.
Certificate of Completion
AI Basics Workshop
Awarded to
Completed on
Issued by Stephen Thorn, AI Trainer
www.stephenthorn.com
Take it further
Reference materials, guides, and next steps.
4D Framework Cheat Sheet
Quick-reference card for all four D's.
AI Tool Comparison Guide
ChatGPT vs Claude vs Gemini - which tool for which job.
AI Safety Guide for Offices
What to share, what to keep private.
Download Slide Template
Customise the presentation for your own team.
Get in Touch →
Questions about coaching, team training, or workshops.
📋 Premium Prompt Library - Your Course Cheat Sheet
15 professional prompts built on the 4D Framework. Copy, customize the [brackets], and use immediately.
🎯 Delegation Decisions
I need to decide whether to delegate [task description] to AI. Help me evaluate: (1) How clear-cut is the expected output? (2) What's the consequence if it's wrong? (3) Does it require personal context or accountability? Give me a delegation recommendation with your reasoning.Here are 5 tasks I do regularly at work: [list tasks]. Rank them from most to least suitable for AI delegation. For each, tell me what I'd gain and what I'd risk.I'm considering using AI for [task]. What's the minimum information I'd need to give AI to make it useful? What should I NOT include for privacy reasons?My team is skeptical about using AI for [type of work]. Help me build a quick business case: what's the time saving potential, what are the risks, and what guardrails would make it safe?
✍️ Description & Prompting Mastery
Act as a [role: e.g. senior copywriter / HR expert / financial analyst]. I need help with [task]. My audience is [describe audience]. The format should be [format]. Keep it under [word/page limit]. Here's my draft/context: [paste content]I asked AI to help with [task] and got a mediocre response. Here's what it gave me: [paste response]. What's missing? How should I rewrite my prompt to get a better result? Give me an improved prompt I can use right now.Give me [number] variations of [piece of content] - each with a different [tone/angle/length]. Label each clearly so I can pick the best one.I need to explain [complex topic] to [audience: e.g. my CEO / a new employee / a non-technical client]. Draft a [format: email/summary/talking points] that covers the key points without jargon.Review this and give me: (1) What's working well, (2) What's unclear or weak, (3) Three specific edits to improve it. Here's the content: [paste content]
🔍 Discernment & Verification
You just gave me [specific claim/statistic/recommendation]. How confident are you in this? What should I independently verify before using it? What sources would you check?I'm about to use this AI-generated content for [purpose/audience]. Play devil's advocate: what could be wrong, misleading, or missing? Flag anything I should double-check.Compare these two AI responses I got to the same question: [paste response A] vs [paste response B]. Which is more accurate and why? What would you add or correct?
🛡️ Diligence & Responsible Use
I want to use AI to help with [work task]. Before I start, help me think through: What data should I NOT share? What's my review process before using the output? How will I attribute it if I publish?Draft an AI usage policy for [team/department/project]. Include: what AI is approved for, what requires human review, what's off-limits, and how to handle attribution. Keep it practical - one page max.I've been using AI heavily for [type of work] lately. What skills or judgment might I be at risk of losing? What should I make sure I still practice myself to stay sharp?
Your AI Basics Prompt Guide
Copy-paste prompts for real work. Customize the bracketed sections.
✍️ Writing & Communication
Rewrite this in a [professional/casual/concise] tone: [paste text]Write a [email/memo/summary] about [topic] for [audience]. Keep it under [X] words.Give me 5 subject lines for an email about [topic]. Make them [direct/engaging/urgent].
🔍 Research & Summarizing
Summarize this in 5 bullet points, then give me the 3 most important takeaways: [paste text]What are the pros and cons of [topic/decision]? Give me a balanced view.Explain [concept] like I have no background in it, then give me a more technical explanation.
💡 Brainstorming & Planning
Give me 10 ideas for [goal/problem]. Be creative and practical.Help me make a plan to [goal] in [timeframe]. Break it into weekly steps.What am I missing? Here's my plan: [paste plan]
📋 Work Tasks
Review this for clarity and errors: [paste text]Write a meeting agenda for a [X]-minute meeting about [topic].Help me prepare for a conversation with [person/role] about [topic].
17 professional prompts built around the 4D Framework. Copy, paste, and customize the [bracketed] sections for your situation.
🎯 Delegation - Decide What to Hand Off
Task Assessment
Before I delegate this task, help me assess: [describe task]. Is it primarily a language task? What's the risk if you get it wrong? What context do you need from me? Weekly Prioritization
I'm trying to decide what to hand off to AI this week. Here are 5 tasks I regularly do: [list tasks]. Categorize each as 'delegate fully', 'do with AI', or 'keep human'. Explain your reasoning for each. Role-Based Triage
I'm overwhelmed with work. Help me identify which tasks in my role as [job title] are the best candidates for AI delegation. Consider: repetitive tasks, first drafts, research, summarization, and formatting. Risk Check
Is [specific task] a good task for AI? What can go wrong if I rely on AI output without verification? What should I always check before using AI output for this? ✍️ Description - Write Better Prompts
5-Element Prompt Template
You are a [role/expert]. I need you to [task]. Context: [background]. Audience: [who reads this]. Tone: [tone]. Format: [format]. Constraints: [any limits]. Prompt Improver
I wrote this prompt: [paste prompt]. Improve it by adding: (1) a clear role for you to play, (2) more specific context, (3) the format I want, and (4) at least one constraint. Show me the improved prompt. Draft Polish
My first draft of [type of content]: [paste draft]. Rewrite this to be more [concise/professional/engaging]. Keep the key points but improve the flow and remove any filler. Audience Translator
I need to communicate [complex topic] to [audience description]. They have [level of expertise] in this area. Write this for them in [format], and explain any jargon in plain language. Refine & Retry
I asked you [question/task] and your first response wasn't quite right because [explain issue]. Please try again with this additional context: [context]. Focus specifically on [what matters most]. 🔍 Discernment - Evaluate AI Output
Confidence Check
You just told me [specific claim]. How confident are you in this? What could be wrong? What should I verify independently before I act on this? Self-Audit
Review your previous response and flag: (1) any statistics or numbers, (2) any specific names, titles, or citations, (3) any claims I should verify before using this professionally. Rate your confidence in each item. Devil's Advocate
I'm going to use this AI-generated content in [context]. Play devil's advocate: what are three ways this content could be wrong, misleading, or incomplete? Critical Editor
Act as a critical editor. Review this text and identify: factual claims that need verification, assumptions that may not apply to my situation, and anything that sounds plausible but could be incorrect: [paste text] 🛡️ Diligence - Use AI Responsibly
Personal AI Policy
Help me write a personal AI usage policy for my role as [job title]. Include: what types of tasks I'll use AI for, what information I'll never input, how I'll disclose AI use to colleagues or clients, and how I'll verify AI output. Safe Workflow Design
I want to use AI for [task] but I'm concerned about [privacy/accuracy/attribution] because [reason]. Help me design a workflow that achieves my goal while managing this risk. AI Disclosure Statement
Draft a short AI disclosure statement I can include when I use AI assistance for [type of content/work]. It should be transparent and professional, not defensive or apologetic. Skill Preservation
I've been using AI heavily for [task] and I'm wondering if I'm losing my own [skill]. Help me design a practice to maintain my expertise while still getting the efficiency benefits of AI.