The recorded session
Watch the Learning Lab.
Follow the session, then use the playbook with your own work.
Watch replay on Vimeo ↗What to take into your work
- Find useful AI tasks in your own responsibilities.
- Give AI enough context to produce something you can use.
- Set boundaries and review the result before acting.
Use the learning
Your playbook
Choose a prompt, replace the bracketed inputs, and work through it in your approved AI tool. Review the result and adapt it to your situation.
This playbook accompanies the April 7, 2026 session. Tool names and setup steps reflect that session; available features may differ in your account.
Welcome to Your AI Toolkit
You've heard about AI from every direction - conferences, LinkedIn, your CEO, that one colleague who won't stop talking about ChatGPT. But between the hype and the fear, you're stuck in the same spot: "I know I should be using this, but I don't know where to start or what's allowed." This playbook gives you three AI prompts that cut through the noise: identify the specific tasks in YOUR role where AI saves real time, build sample prompts you can use today, and create a responsible use framework so you experiment with confidence.
- You don't need to use everything at once. Start with the one prompt that solves your most immediate problem.
- These prompts work across platforms: Copilot, ChatGPT, Gemini, Claude, or your company's internal AI tool.
- Each prompt includes a main version (for licensed tools with full context windows) and an alternate version (for free Copilot or tools with limited input).
How to Use This Playbook
- Start with Prompt #1. List your key responsibilities and let AI identify where you're spending time that AI could save.
- Pick your first use case. Use Prompt #2 to build a complete prompt kit for that specific task - ready to use tomorrow.
- Set your guardrails. Use Prompt #3 to create a personal AI usage framework so you know what's safe and what's not.
- Give AI context before asking for help. The more specific you are about your role, your industry, and your constraints, the better the suggestions.
- Review and refine. AI suggests where it can help. You decide where it actually fits into your workflow.
The AI Use Case Finder Framework
Not every task is a good AI task. Use this filter to identify where AI adds real value in your role.
AI Fit | Task Characteristics | Examples | Time Saved |
|---|---|---|---|
🟢 High Value | Repetitive, text-heavy, follows a pattern, takes 15+ minutes, doesn't require confidential data | Drafting emails, summarizing reports, building agendas, creating templates, preparing talking points | 2–5 hours/week |
🟡 Medium Value | Requires some judgment, needs your review, involves analysis or brainstorming | Brainstorming options, analyzing data patterns, creating first drafts of proposals, preparing for presentations | 1–2 hours/week |
🟠 Use With Caution | Involves confidential information, requires deep expertise, has legal/compliance implications | Performance reviews (use for structure, not content), financial analysis (verify all numbers), customer communications (review every word) | 30–60 minutes/week with heavy review |
🔴 Not for AI | Requires human judgment, empathy, relationship, confidential data, or legal accountability | Termination conversations, union negotiations, safety investigations, decisions with legal exposure | 0 - keep these human |
Key Insight: The leaders getting the most from AI aren't using it for the flashy stuff. They're using it for the boring stuff - the emails, summaries, agendas, and prep work that eat 10+ hours a week. That's where the time savings are. The 2-Hour Rule
- The goal: Find tasks where AI saves you at least 2 hours per week.
- The test: If you wouldn't save at least 15 minutes on a task, it's not worth the prompt setup.
- The compounding effect: 2 hours/week = 100+ hours/year = 2.5 weeks of your life back.
The AI Efficiency Ladder
Know where you are. Know where you're going.
Level | What It Looks Like | Time / Task |
|---|---|---|
1. On-the-Fly Prompting | Opening an AI chat, typing a question, getting an answer. Starting from scratch every time. | 5–10 minutes, inconsistent |
2. Repeatable Prompts | Saved templates you fill in. Same quality every time because same structure. | 2–3 minutes, consistent |
3. Custom AI Projects | AI pre-loaded with your context – your role, industry, frameworks. Works like a team member. | 1–2 minutes, personalized |
4. Automation & Agents | AI connected to your systems, triggered automatically. Runs without you. | 0 minutes, scales |
Your goal: Move from Level 1 to Level 2. The prompts in this playbook get you there.
The Prompt Structure
Every strategic prompt has four parts:
Part | Purpose | Example |
|---|---|---|
ROLE | Who should AI pretend to be? | "You are an executive productivity advisor for automotive leaders..." |
CONTEXT | What does AI need to know about you? | "I'm a plant manager at a Tier 1 supplier, managing 150 people across 3 shifts..." |
REQUEST | What specifically do you want? | "Identify 5-10 tasks where AI can save me 2+ hours per week..." |
QUESTION | Invite the AI to get as much information from you as possible. | "Please ask me any clarifying questions." |
Key Prompts from This Lab
This lab teaches three prompts that work as a sequence - Discover, Build, Guard.
Prompt #1: The AI Use Case Finder
What it's for: Taking your actual job responsibilities and identifying the specific tasks where AI saves real time - with a sample prompt and quality check for each one. When to use it: When you know you should be using AI but don't know where to start. When you want a personalized roadmap instead of generic advice.
Main Version (Licensed Copilot / ChatGPT / Gemini)
Use when you can provide full context about your role and responsibilities.
ROLE You are an executive productivity advisor who specializes in helping automotive industry leaders integrate AI into their daily work. You focus on practical, high-value use cases that save measurable time — not theoretical possibilities. You understand that leaders in manufacturing, dealerships, and automotive supply chains have unique constraints: confidentiality requirements, union environments, regulated processes, and very little patience for tools that don't immediately prove their value. CONTEXT I am a [ROLE] at an automotive [OEM / Tier 1 / Tier 2 / Dealership / Independent]. I want to identify specific, practical ways to use AI in my daily work that will save me real time every week. REQUEST Based on the responsibilities I describe, do the following: 1. AI USE CASE MAP Identify 5–10 high-value AI use cases specific to my role. For each: | Task | Current Time Spent | AI Can Help By... | Estimated Time Saved | AI Fit Level (🟢/🟡/🟠/🔴) | 2. TOP 3 QUICK WINS The three tasks I should automate FIRST — because they're easy, safe, and save the most time. Ranked by impact. 3. SAMPLE PROMPT FOR EACH For every use case, provide a ready-to-use prompt I can copy, paste, and modify. Include: - The full prompt text - Where to paste my specific information - What the output should look like - How long it should take 4. QUALITY CHECK FOR EACH What to review before using AI output: - What to verify (facts, names, numbers, tone) - What to never trust AI with (decisions, commitments, legal statements) - A simple "Is this ready?" checklist per use case 5. WHAT TO SKIP Tasks in my role where AI is NOT the right tool — and why. Be honest about the limitations. FORMAT - Organize by task category (Communication, Planning, Analysis, Documentation, People) - Include estimated weekly time savings for each - Start with the easiest, lowest-risk use cases QUESTION Before generating use cases, ask me: 1. What's your exact role and title? 2. List your key responsibilities (paste your job description or bullet points) 3. What takes up most of your time each week? 4. What's the most tedious part of your job? 5. What AI tools do you currently have access to? 6. Any restrictions I should know about? (data policies, union rules, industry regulations)
Alternate Version (Copilot Free / Limited Context)
I am a [ROLE] at an automotive [COMPANY TYPE]. Here are my key responsibilities: [PASTE BULLET LIST] My weekly headaches: [LIST YOUR MOST TIME-CONSUMING OR TEDIOUS TASKS] Identify 5-10 specific tasks where AI can save me at least 2 hours per week. For each: 1. What the task is 2. How AI helps (with a sample prompt I can copy and use) 3. How much time it saves 4. What I need to check before using the output 5. Rate it: green (safe, easy), yellow (needs review), orange (use cautiously), red (keep human) Start with the 3 easiest wins. Be specific to automotive. No generic suggestions.
Example
Q1 - Role and title: “Plant Manager at an automotive Tier 1 stamping and assembly supplier. 300 employees across two shifts.” Q2 - Key responsibilities: “Daily production meetings, weekly quality reviews, customer complaint responses, status reports to regional VP, shift scheduling, safety audit management, continuous improvement updates, supplier issue resolution, and performance conversations with direct reports.” Q3 - What takes most of your time: “Status emails to leadership (5 per week, 25–30 minutes each), meeting prep (6–8 meetings per week, 20–30 minutes each to prep), and monthly plant performance reports (2–3 hours each).” Q4 - Most tedious part: “Writing the same weekly status email every Friday. Production numbers, quality metrics, open issues, next week’s priorities. Same format, different data. Takes 60–90 minutes I don’t have.” Q5 - AI tools available: “ChatGPT (free), Microsoft Copilot through our M365 license.” Q6 - Restrictions: “Union environment. No employee names or personal data in public AI tools. Proprietary customer specs and pricing are off-limits. Enterprise Copilot is approved for internal data; ChatGPT is not.”
Prompt #2: The Prompt Kit Builder
What it's for: Taking a specific task you do regularly and building a complete, reusable prompt kit - including the prompt template, variations for different situations, and a quality checklist. When to use it: After identifying your top use cases with Prompt #1. When you want to move from Level 1 (on-the-fly prompting) to Level 2 (repeatable prompts).
Main Version (Licensed Copilot / ChatGPT / Gemini)
Use to build a complete prompt kit for one recurring task.
ROLE You are a prompt engineering coach for business leaders. You turn one-time prompts into reusable, reliable prompt templates that produce consistent results every time. You know that the difference between a leader who "tried AI once" and a leader who uses AI daily is having saved prompts that work. CONTEXT I have a task I do regularly that I want to build a complete prompt kit for. I want a template I can reuse every time, with variations for different situations, so I never start from scratch. REQUEST Based on the task I describe, create: 1. MASTER PROMPT TEMPLATE A complete, copy-paste-ready prompt with: - ROLE section (who AI should be) - CONTEXT section (with [FILL IN] placeholders for my specific information) - REQUEST section (exactly what I want AI to produce) - FORMAT section (how the output should be structured) - QUALITY CHECK (what to verify before using) 2. VARIATIONS 3–4 modified versions for different situations: - Quick version (when I have 2 minutes) - Detailed version (when I need comprehensive output) - Audience-specific version (when the output goes to different people) - Follow-up version (when I need to iterate on the first output) 3. EXAMPLE INPUT AND OUTPUT A realistic example showing: - What I would paste into the prompt - What good AI output looks like - What bad AI output looks like (so I know when to re-prompt) 4. PROMPT MAINTENANCE - When to update this prompt (new responsibilities, process changes) - How to improve it over time (what to adjust based on results) - Where to save it for quick access FORMAT - Make the master prompt copy-paste ready with clear [FILL IN] markers - Include estimated time from prompt to usable output - Design for someone who's used AI a few times but isn't an expert QUESTION Ask me: 1. What's the specific task? 2. How often do you do it? (daily, weekly, monthly) 3. What does the output need to look like? (email, document, talking points, analysis) 4. Who sees the output? (just you, your team, leadership, customers) 5. Show me an example of the task done well (optional)
Alternate Version (Copilot Free / Limited Context)
I do this task regularly: [DESCRIBE THE TASK] Frequency: [HOW OFTEN] The output goes to: [AUDIENCE] Build me a reusable prompt kit: 1. A master prompt template I can copy, paste, and fill in every time (with [FILL IN] placeholders) 2. A quick version for when I have 2 minutes 3. A detailed version for when I need comprehensive output 4. An example of what good output looks like vs. bad output Make the prompt specific to automotive industry context. Include a quality checklist so I know when the output is ready to use.
Example
Q1 - Specific task: “Weekly status emails to regional VP and corporate leadership. Every Friday I write a report covering: production numbers vs. target, quality metrics (scrap rate, PPM, customer complaints), open issues and containment actions, and next week’s priorities. Currently takes 60–90 minutes because I’m writing from scratch each time.” Q2 - Frequency: “Weekly. Every Friday by 3 PM. Non-negotiable.” Q3 - Output format: “Professional email, 3–4 paragraphs. Starts with a production summary, then quality highlights, then open issues with status, then next week’s priorities. Needs to be specific — actual numbers, not vague statements. My VP wants to scan it in 60 seconds and know exactly where we stand.” Q4 - Who sees it: “Regional VP, corporate quality director, and my plant leadership team (copied). Tone needs to be confident and direct — no hedging, no filler.” Q5 - Example of task done well: “A good status email from my plant: ‘Production hit 97.2% of target (462,000 parts shipped vs. 475,000 planned). Miss driven by 6-hour unplanned downtime on Line 4 — die wear. Replacement die installed, line back at rate. Scrap rate 2.1% vs. 1.8% target. Top contributor: Line 2 trimming defect (corrective action in place, monitoring through Friday). Three open customer issues — all containment verified, root cause pending on 2 of 3. Next week: OEM customer audit Wednesday, new product launch prep on Line 5.’”
Prompt #3: The Responsible AI Framework Builder
What it's for: Creating a personal (or team-level) AI usage framework that clarifies what's safe to use AI for, what requires caution, and what should stay human - so you experiment with confidence. When to use it: Before rolling out AI across your team. When you're not sure what your company allows. When you want clear guidelines instead of vague anxiety about "doing it wrong."
Main Version (Licensed Copilot / ChatGPT / Gemini)
Use to build a practical AI usage framework for yourself or your team.
ROLE You are a responsible AI advisor for automotive industry leaders. You help people create practical AI usage guidelines that protect the company while encouraging experimentation. You know that most leaders are paralyzed by not knowing what's "allowed" — and you replace that paralysis with a clear green/yellow/red framework. CONTEXT I want to create a simple AI usage framework for myself and my team. I need to know what's safe, what needs caution, and what's off-limits — based on my specific role, industry, and company environment. REQUEST Based on my role and environment, create: 1. THE GREEN/YELLOW/RED FRAMEWORK 🟢 GREEN — Safe to use AI freely: - Tasks, data types, and outputs that are low-risk - No approval needed. Just use it and review the output. 🟡 YELLOW — Use with caution: - Tasks where AI helps but output needs careful review - May involve internal information (not confidential) - Review before sharing externally 🔴 RED — Do not use AI: - Tasks involving confidential, personal, or legally sensitive data - Decisions with legal or safety implications - Customer-facing content without human review 2. DATA SAFETY CHECKLIST Before pasting anything into AI, ask: - [ ] Does this contain employee names or personal information? - [ ] Does this contain financial data, pricing, or proprietary information? - [ ] Does this contain customer information? - [ ] Could this cause harm if it appeared in a public data breach? - If any answer is "yes" — anonymize or don't paste. 3. THE "BEFORE YOU HIT SEND" CHECKLIST Before using AI output: - [ ] Did I verify all facts, names, and numbers? - [ ] Does the tone match my voice and our culture? - [ ] Would I be comfortable if my boss saw this was AI-assisted? - [ ] Have I checked for hallucinated information? - [ ] Is this appropriate for the audience? 4. TEAM GUIDELINES (if applicable) A simple one-page AI usage policy: - What we use AI for - What we don't use AI for - How we review AI output - How we learn and share what works 5. COMMON SCENARIOS For my specific role, provide green/yellow/red guidance for 10 common scenarios. QUESTION Ask me: 1. What's your role and industry? 2. What AI tools does your company provide or allow? 3. What types of data do you handle? (confidential, internal, public) 4. Does your company have existing AI policies? 5. What's your biggest fear about using AI at work?
Alternate Version (Copilot Free / Limited Context)
I am a [ROLE] at an automotive [COMPANY TYPE]. AI tools available: [LIST] Types of data I handle: [DESCRIBE] Company AI policy: [EXISTS / DOESN'T EXIST / NOT SURE] Create a simple AI usage framework: 1. GREEN tasks (safe, no approval needed) — list 5-7 examples specific to my role 2. YELLOW tasks (use with caution, review before sharing) — list 5-7 examples 3. RED tasks (do not use AI) — list 5-7 examples 4. A data safety checklist (5 questions to ask before pasting anything into AI) 5. A "before you hit send" checklist for reviewing AI output Be specific to automotive. Give me a framework I can print and pin next to my monitor.
Example
Q1 - Role and industry: “Service Director at an automotive Tier 1 supplier. I manage a 12-person engineering team that handles product development, APQP, and customer quality support.” Q2 - AI tools available: “Microsoft Copilot through our M365 enterprise license. Several engineers are also using free ChatGPT and Claude on their personal devices. One engineer pasted proprietary customer specs into a free AI tool last week — that’s what triggered this.” Q3 - Types of data handled: “Customer-specific part drawings and specs (proprietary), PPAP documentation, internal cost data and pricing, employee performance records, quality investigation reports with root cause data, and supplier audit findings.” Q4 - Existing AI policy: “No formal policy. IT sent a vague email saying ‘use AI responsibly’ but gave no specifics. Nobody knows what’s actually allowed.” Q5 - Biggest fear: “Someone pastes confidential customer data into a public AI tool and it ends up in a breach. Also worried about engineers trusting AI-generated calculations without verifying them — we can’t have AI hallucinating tolerance specs on a brake component.”
Prompt #4: The Meeting Prep Machine
What it's for: Taking any upcoming meeting and preparing a complete brief: agenda, talking points, anticipated questions, follow-up template - in under 5 minutes. When to use it: Before any important meeting, review, or presentation. The single highest-value daily AI use case for most leaders.
Main Version
ROLE You are an executive assistant who prepares leaders for meetings. You build concise meeting briefs that ensure the leader walks in prepared, focused, and ready to drive decisions. CONTEXT I have a meeting coming up and I need to be prepared. I'll give you the context — who's attending, what we're discussing, and what I need to accomplish. REQUEST Create a meeting prep package: 1. AGENDA (structured, timed) - Topic, time allocation, desired outcome for each item - Decision items clearly marked - Parking lot for items to defer 2. MY TALKING POINTS - 3–5 key points I need to make - Data or evidence to support each point - One question to ask that moves the discussion forward 3. ANTICIPATED QUESTIONS - 3–5 questions I might be asked - A concise, confident answer for each 4. FOLLOW-UP EMAIL TEMPLATE - Pre-written follow-up email with [FILL IN] placeholders for decisions and action items - Ready to send within 10 minutes of the meeting ending 5. ONE-PAGE BRIEF - Everything above condensed to a single page I can glance at during the meeting QUESTION Ask me about the meeting: Who, what, when, and what I need to accomplish.
Alternate Version (Copilot Free)
I have a meeting: With: [WHO] About: [TOPIC] I need to: [MY GOAL] Context: [BACKGROUND] Create: 1. A timed agenda with clear outcomes for each topic 2. My 3-5 key talking points with supporting evidence 3. 3-5 questions I might be asked with concise answers 4. A follow-up email template I can send after the meeting Keep the whole package to one page.
Example
Q1 -Who: “Quality review with our largest OEM customer. Their quality director, SQE lead, and purchasing manager will be there. From our side: me (plant manager), quality manager, and engineering lead.” Q2 - What: “Discussing 3 open PPAPs (two behind schedule), a containment issue on a brake bracket (field returns, 50 units affected), and a corrective action plan they rejected last week. They also want an update on our scrap reduction initiative.” Q3 - When: “Tomorrow at 10 AM. Two hours blocked. This is the meeting that determines whether they expand our business or start dual-sourcing.” Q4 - What I need to accomplish: “Get the corrective action accepted on second submission. Show progress on PPAP timelines. Demonstrate that containment is airtight on the brake bracket. Walk out with the relationship intact and no new escalations.”
Prompt #5: The Weekly AI Time Tracker
What it's for: Building a simple system to track where AI saves you time each week - so you can measure the impact, share results with your team, and build the case for broader adoption. When to use it: After your first week of using AI consistently. When your boss asks "Is AI actually saving you time?" and you want a real answer.
Main Version
ROLE You are a productivity measurement specialist. You help leaders track and quantify the time AI saves them, so they can make data-driven decisions about expanding AI use. CONTEXT I've started using AI for several tasks and I want to track how much time it's actually saving me. I need a simple system I'll actually maintain. REQUEST 1. WEEKLY TIME TRACKER A simple table I fill in each day: | Day | Task | Without AI (est. minutes) | With AI (actual minutes) | Time Saved | Notes | 2. MONTHLY SUMMARY TEMPLATE - Total hours saved - Top 3 time-saving use cases - What worked well - What didn't work (tasks where AI didn't help) - One new use case to try next month 3. EXECUTIVE SUMMARY A one-paragraph template I can share with my boss: "This month, I used AI for [X tasks], saving approximately [Y hours]. The highest-value use cases were [list]. I plan to expand into [next area] next month." 4. TEAM ADOPTION FRAMEWORK If I want to share this with my team: - How to introduce AI without mandating it - A 30-day "try one thing" challenge - How to collect and share team results QUESTION Ask me what tasks I've been using AI for so far and approximately how long they take with vs. without AI.
Alternate Version (Copilot Free)
I've started using AI for these tasks: [LIST TASKS] Create: 1. A simple weekly time tracker table (task, time without AI, time with AI, time saved) 2. A monthly summary template (total hours saved, top use cases, what to try next) 3. A one-paragraph executive summary I can share with my boss 4. A 30-day team challenge to help my team start experimenting Keep it simple enough that I'll actually fill it in every day.
Example
Here’s what I’ve been using AI for over the past 30 days:
- Weekly status emails to regional VP — used to take 60–90 minutes, now takes 15 minutes with a saved prompt. Five emails per week.
- Meeting agendas and talking points — used to take 20–30 minutes per meeting, now takes 5 minutes. Six meetings per week.
- Monthly plant performance report — used to take 2–3 hours, now takes 45 minutes with AI drafting the narrative sections.
- Corrective action drafts — used to take 90 minutes per 8D, now takes 30 minutes. About 2 per month. I want to track this systematically so I have real data when my boss asks if AI is worth expanding to the rest of the leadership team.”
Prompt #6: The Stress-Test Prompt
What it's for: Uploading a proposal, strategy, plan, or idea and having AI stress-test it - finding blind spots, risks, and weaknesses before your boss, your board, or your customer does. When to use it: Before any high-stakes presentation, proposal submission, or strategy review. Smart leaders don't just use AI to build faster - they use it to think better.
Main Version
I'm going to paste [proposal/strategy/idea]. Act as a friendly but ruthless reviewer. Step 1 — Steelman: First, restate my idea in the strongest possible form (3–5 bullets). Step 2 — Stress test: Identify the top 7 risks/blind spots across: strategy, execution, incentives, stakeholders, edge cases, metrics, and timeline. Step 3 — Fixes: For each issue, give: Impact if ignored Likelihood (low/med/high) Early warning signal One concrete mitigation Step 4 — Kill criteria: Tell me what would make you say "don't do this". Step 5 — Stronger version: Rewrite my proposal in a tightened form that addresses the biggest weaknesses.
Alternate Version (Copilot Free)
I'm going to paste my [proposal/strategy/plan]. Act as a devil's advocate. Do the following: 1. Restate my idea in the strongest form (3 bullets) 2. List the top 5 risks or blind spots I'm missing 3. For each risk: rate it (low/med/high) and give one concrete fix 4. Tell me: what would make this fail? 5. Rewrite the weakest section to make it stronger Be direct. I want to find the problems before someone else does.
Example
Here’s my proposal for reducing overtime costs on the stamping line: Problem: Overtime has averaged 18% over the past 6 months, costing $240K above budget. Root causes: unplanned die changes running 40% longer than standard, scheduling gaps during shift transitions, and reactive maintenance driving 3–4 unplanned stops per week. Proposed solution: (1) Implement predictive die maintenance using press tonnage monitoring — $45K capital investment, 3-month payback. (2) Restructure shift overlap from 15 to 30 minutes with a formal handoff checklist. (3) Move from reactive to condition-based maintenance for the 5 highest-downtime presses. Expected results: Reduce overtime to 10% within 90 days. Annual savings of $150K. Reduce unplanned stops from 3–4 per week to 1 or fewer. I’m presenting this to my VP next Tuesday. Stress-test it before I walk in.”
Quality Check - Before Trusting AI Output
- Did you verify the facts? AI can hallucinate statistics, names, dates, and even company policies. Never trust AI-generated facts without checking.
- Does this match your company's AI policy? If you're not sure what's allowed, ask IT or your manager before pasting sensitive information into any AI tool.
- Did you check for confidential data? Before pasting anything into AI, ask: "Would I be comfortable if this appeared in a public data breach?" If not, don't paste it.
- Does the output sound like you? AI-generated text often sounds generic. Edit it to match your voice, your culture, and your relationships.
- Are you using AI to assist, not replace? AI drafts. You decide. AI summarizes. You verify. AI brainstorms. You choose. Keep yourself in the loop.
Copilot Basics (If You Use Microsoft 365)
- Work Mode vs. Web Mode: Work mode searches your company's Microsoft 365 data (emails, files, Teams). Web mode uses the open internet. For identifying AI use cases, Web mode is great for brainstorming. For actually using AI in your workflow, Work mode can summarize your recent emails, prepare for meetings using your calendar, and draft responses using your email history.
- When to use each: Use Web mode with the prompts in this playbook for planning and setup. Once you have your use cases, try Work mode for the ones that involve your company data (email summaries, meeting prep, document analysis).
- How to save prompts: Copy your best prompts into a Word document, OneNote, or Notion page. The goal is to stop typing prompts from scratch and start using saved templates.
- Reminder: Your company's AI usage policy determines what data you can use with AI tools. When in doubt, use the Green/Yellow/Red framework from this playbook.
What to Try This Week
- Run Prompt #1. Paste your responsibilities and find your top 3 AI quick wins.
- Pick ONE task. The most tedious, repetitive task on your list. Build a reusable prompt for it using Prompt #2.
- Use it for real. Don't just test it — use the output in your actual work this week. Send the email AI helped you draft. Use the agenda AI built.
- Track the time. Note how long it took with AI vs. how long it usually takes. You'll want this data.
- Share with one colleague. "Hey, I tried this prompt and it saved me 30 minutes. Want to see it?" That's how AI adoption spreads.
Resources & Next Steps
Key Frameworks from This Lab
- The AI Use Case Finder Framework uses a Green/Yellow/Orange/Red filter to identify which tasks are high-value AI candidates and which should stay human.
- The 2-Hour Rule: Focus on tasks where AI saves at least 2 hours per week. That's 100+ hours per year.
- The AI Efficiency Ladder shows four levels from on-the-fly prompting to full automation. Most leaders should aim for Level 2: repeatable prompts.
- Responsible use isn't about restriction — it's about confidence. A clear framework lets you experiment without fear.
What We Didn't Cover (Yet)
- Building custom AI assistants (GPTs, Copilot agents) pre-loaded with your context.
- Connecting AI to your company systems for automated workflows.
- Advanced prompt engineering techniques for complex analysis.
- Building the business case for organizational AI adoption.
Next Steps
- Use Prompt #1 to find your top 3 quick wins. Start using one this week.
- Build a prompt kit for your most repetitive task using Prompt #2.
- Create your personal Green/Yellow/Red framework using Prompt #3. Pin it next to your monitor.
The Mental Model - Discover, Build, Guard
Today's session was built on one idea: the leaders getting the most from AI aren't doing anything fancy. They found 3–5 tasks that eat their time, built reusable prompts, and set clear guardrails. That's it. That's the whole strategy.
Step | Prompt | What It Does |
|---|---|---|
Discover | Prompt 1 — The AI Use Case Finder | Maps your responsibilities to high-value AI use cases with sample prompts and quality checks. |
Build | Prompt 2 — The Prompt Kit Builder | Turns one-time prompts into reusable templates you can use every day. |
Guard | Prompt 3 — Responsible AI Framework Builder | Creates a Green/Yellow/Red framework so you know what's safe and experiment with confidence. |
You don't need to become an AI expert. You need to find 3 tasks, build 3 prompts, and save 3 hours a week. Start there.
The 9 AI Skills That Set You Apart
You've been told AI writes emails and summarizes documents. That's the toothpick on the Swiss Army knife. Here are the 9 skills that separate people who use AI from people who get real results with AI. Skills covered in this lab:
- The Art of the Ask - The RCRQ Framework. The quality of your output depends on the quality of your input.
- Know What Good Looks Like - Develop your eye for quality. Give AI an example output.
- Push Back, Don't Settle - The first answer is never the final answer. Iterate.
- Set the Rules - Use system prompts and custom instructions to define how AI behaves.
- Build Your AI Toolkit - Save your best prompts. Build a personal library over time. Skills to explore next:
- Make AI Challenge You - Upload your proposal and ask AI to stress-test it. (See Prompt #6 above.)
- Feed It Smart - Learn to compress and structure the context you give AI. Better inputs = better outputs.
- Make AI Your Tutor "Teach me this in 5 minutes at a conversational level." AI adapts to your pace.
4. Create Your Digital Identity (Master Prompt)
Create a master prompt that tells AI who you are - your role, your style, your goals. Load it at the start of every session so AI already knows your context. No more starting from scratch every time.
Act as a strategic interviewer and AI architect. Your job is to extract everything needed to build a comprehensive Master Prompt that will serve as my persistent digital identity across AI platforms. The goal is to create a Master Prompt that: Encodes who I am professionally Clarifies what I’m building Captures how I think and make decisions Defines what I value and reject Establishes how I want AI to respond to me Reduces generic outputs and increases strategic alignment Process: Step 1: Interview Me Ask structured, high-signal questions in the following categories. Do not skip categories. Ask follow-up questions when answers are vague. Role & Professional Identity What do you do? Who do you serve? What problems do you solve? What industry context matters? Strategic Direction What are you building over the next 2–5 years? What constraints or realities shape your decisions? What are you optimizing for (money, scale, impact, freedom, etc.)? Decision-Making Style How do you prefer to receive advice? Do you want challenge, validation, or neutral analysis? How do you balance speed vs. depth? Standards & Taste What does “excellent” look like in your world? What frustrates you in AI outputs? What tone should be avoided? What structural preferences do you have (concise, long-form, bullets, narrative)? Communication Preferences Desired tone Level of technical detail Use of analogies or not Formatting preferences Constraints & Non-Negotiables Ethical boundaries Political sensitivities Industry compliance concerns Personal red lines Repeating Context What background information do you repeatedly have to explain to AI tools? What should always be assumed? Ask questions one category at a time. Wait for answers before moving forward. Step 2: Synthesize After the interview is complete: Synthesize all answers into a structured Master Prompt. Write it in instruction format, as if it will be pasted into the “System” or “Custom Instructions” section of an AI tool. Use clear, directive language. Eliminate redundancy. Preserve nuance. Step 3: Deliver Two Outputs Output A: A Clean, Copy-Paste Master Prompt (plain text, ready to use). Output B: A Short Explanation (under 300 words) explaining how to use this Master Prompt across different AI platforms (ChatGPT, Claude, Gemini, custom GPTs). Do not provide commentary outside these outputs. Begin the interview now.
The professionals who treat AI as a skill - who invest in learning how to use it well - will have an edge that compounds over time.
AI isn't the future. It's the present - and smart leaders are already using it:
- 65% of women hide their AI usage at work (McKinsey & Lean In, 2024). Don't hide it - lead with it.
- Find your top 3 time-wasters.
- Build one reusable prompt this week.
- You now have the tools to work smarter - not harder.
Related resources
Keep working with the material.
The prompts and exercises are included above. Save a useful prompt and return to it with your next task.