Level Up Your GenAI Game

A No-BS Guide to Better Prompts

Let's be honest: most of us are using AI like we're at a drive-through - throwing quick, vague requests and hoping for the best. "Write me a blog post!" "Make this better!" "Help me with marketing!"

Sound familiar?

Here's the thing: AI isn't a fast-food joint; it's more like a brilliant (but very literal) team member who needs proper briefing.

The better you communicate, the better your results.

After a year of daily AI wrangling, here's what I've learned about getting AI to actually do what you want.

1. 🥇 Make AI Your Default Buddy

It didn’t come naturally for me to start thinking AI-first (and I’m still working at it) so I have a very non-technical solution to help me remember to seek it out more.

I used to keep a sticky note on my monitor, “Will money solve this problem?” and now it has a neighbor that says "Will AI solve this problem?" Sounds simple, but it's revolutionized my workflow. Before diving into any task, I pause and consider if AI could help.

🧠 Some game-changing moments:

• Before writing a strategy doc, I ask AI to outline key points to consider

• Starting client meetings with "Let me share what AI suggests first" (they love this)

• Using AI as a sounding board for ideas (seriously, it's like having a 24/7 brainstorm partner or... sometimes… therapist lol)

✨ Pro Tip: Save your best prompts as text snippets. Any notes app works. When you nail a prompt that works well, save it for future use.

2. 🤠 Make AI Wear A Hat (The Persona Hack)

Before: "How should I price my consulting services?"

After: "You are a senior McKinsey pricing strategist with 20 years of experience in the tech consulting industry. How should I structure my AI consulting services pricing for enterprise clients? Consider market positioning, value-based pricing, and competitive dynamics."

The difference? Night and day. By giving AI a specific role, you tap into its ability to adopt different perspectives and expertise frameworks.

🎩 Great roles to try:

• Senior Data Analyst for number crunching

• Executive Communications Director for important messages

• Devil's Advocate for testing your ideas

• Career Coach for professional development

• Experienced Project Manager for planning

✨ Pro Tip: Create a "persona library" of roles that consistently give you good results. I keep a running list of the most effective personas for different tasks - my "McKinsey advisor" for strategy, "Senior Copy Editor" for writing, and "Data Visualization Expert" for presentations.

3. 🗝️ Context is Key

Before: "Write me a sales email."

After: "Write a sales email for our AI consulting services.

❓️ Context:

• Target audience: CTOs at mid-size tech companies (500-2000 employees)

• Pain point: Falling behind competitors in AI adoption

• Our differentiation: Focus on practical implementation vs theoretical strategy

• Tone: Professional but conversational

• Previous success: 40% increase in productivity for similar clients

• Call to action: Book a 30-minute discovery call"

The more context you provide, the more tailored and effective the output.

🤔 Think about:

• Who's this for?

• What's the background?

• What's the goal?

• What constraints exist?

• What's worked before?

✨ Pro Tip: Keep a basic context template in your notes app. Mine includes: target audience, business context, tone preferences, and key constraints. I copy-paste this structure whenever I'm writing a new prompt - it ensures I never forget crucial context.

4. 🫵 Get Specific About Outputs

Before: "Help me analyze our Q4 results."

After: "Create an analysis of our Q4 results in the following format:

• A bulleted executive summary (max 5 points)

• Three data visualizations showing key trends

• A section comparing results to Q3 and previous year

• Recommendations in a table with 'Quick Wins' and 'Long-term Strategies'

• Export the data in CSV format for further analysis"

🔮 Be crystal clear about:

• Format (bullets, paragraphs, tables)

• Length (word count, number of sections)

• Tone (formal, casual, technical)

• Special requirements (citations needed? data viz?)

• File formats or programming language needed (csv? react?)

✨ Pro Tip: Start complex requests with "Let's approach this step by step:" - I've found this simple phrase dramatically improves how AI structures its responses, especially for detailed analyses or reports.

5. 🪜 Break It Down (The Step-by-Step Method)

Before: "Help me plan a conference."

After: "Let's break this conference planning into steps. First, help me create:

1. A timeline template for a 2-day conference showing key milestones

2. Then we'll work on:

   • Speaker invitation email templates

   • Budget spreadsheet structure

   • Marketing timeline

   • Vendor evaluation criteria"

Complex tasks become manageable when broken down. Instead of one massive prompt, use multiple focused ones. This also helps you save your credits as you can QA (quality assurance) each step before the following builds upon it.

✨ Pro Tip: Before sending a big prompt, write down your "acceptance criteria" - what would make this output truly useful? Use these criteria to break down your request into smaller, more manageable chunks.

6. ♻️ Iterate Like a Pro

The secret sauce? Iteration. Your first prompt rarely gives you exactly what you need.

Here's my process:

1. Start with a basic prompt

2. Evaluate the response

3. Refine based on what's missing

4. Keep the good parts, fix the gaps

Example iteration:

1. First try: "Write a blog post about AI in HR"

2. Response: Too general

3. Second try: "Write a blog post about how HR teams are using AI for recruitment, focusing on practical examples and ROI data from 2023-2024"

4. Response: Better but needs structure

5. Final try: "Write a blog post about AI in HR recruitment. Include:

   • 3 real-world case studies

   • Specific tools being used

   • ROI data from 2023-2024

   • Common implementation challenges

   • Future predictions for 2025 Format as a 1000-word article with subheadings and a key takeaways section"

The Bottom Line

Effective prompt engineering isn't about memorizing frameworks or following rigid rules. It's about clear communication and iteration. 

Start with these basics, save your successful prompts, and keep refining your approach. The more you practice, the more natural it becomes.

And remember: when in doubt, be more specific. AI can handle complex instructions better than vague ones.

What's your favorite prompt hack? Drop me a note - I'm always collecting new tricks to share!