- Leading Disruption
- Posts
- My Every Day, Powered By AI
My Every Day, Powered By AI
Charlene Li
June 30, 2025
I often get asked the same question, whether I’m speaking at a conference or hosting my weekly livestream:
"How much of your presentation was made by AI?"
People are curious, and sometimes skeptical. They want to know if my slides are AI-generated, how I'm using the tools they’re hearing about everywhere, and most importantly, how much they can trust the content when AI is involved.
I get it: AI empowers everyone to create, code, and learn on our own terms. But without deeper insight into how to use and fact-check AI, things can get murky fast. Transparency into AI best practices is essential for building trust in this new landscape.
Using Multiple Models
Models are constantly changing, and new ones enter the market every day. Pay attention to the purpose of the model you’re considering using: who was it created for, and what was it created to do? This will inform your understanding of its strengths, weaknesses, and utilities. Being judicious about the models you choose to use for various tasks is the foundation of a successful, productive AI experience.
My go-to models include:
✍️ Claude, which is great for lyrical, creative writing. It understands my voice well and helps me communicate clearly.
♊ Gemini, which has a huge context window and can handle a ton of information at once.
💬 ChatGPT, which is my AI Swiss Army knife! It’s incredibly versatile, and knows my voice and workflow well from constant use.
…and many more!
Once you’ve chosen the right platform to work with, you can deepen your work by customizing models to meet your needs. Custom GPTs, Gemini Gems, Claude Projects and other custom models allow you to create AI tools that do exactly what you’re looking for and save you time. A handful of examples:
My BookWise Editor is packed with my chapters, interview transcripts, and podcasts. It knows how I think and write.
I uploaded pictures of my wardrobe to FashionistaAI, which gives me styling advice.
My Functional Medicine Advisor model helps me interpret labs—not replacing my doctor, but helping me be more informed.
I trained a Virtual Bookkeeper on QuickBooks and Gusto for financial questions. Here’s a screenshot of it helping me with end of year 401K contribution accounting. And yes, I have my accountant check everything!
Power Prompting: Your Foundation
Even if you’ve created a custom model, humans are an essential part of the equation, and the way you prompt AI matters. When prompting is done right, you get better results. When I use AI tools, I use a very specific prompt structure, followed by stacked prompts that allow me to refine the results.
Here’s how I structure my prompts:
1️⃣ Define a role. Tell your AI model how it should be thinking. For example, “You’re a book marketing expert.”
2️⃣ Provide detailed instructions. Walk your model through a detailed sequence of instructions, such as “Read my chapter, create an outline, and make recommendations for an improved flow.”
3️⃣ Add context. Give your AI partner the information it needs to produce a strong result. For example, tell it “I’m trying to improve the narrative flow of this chapter.”
4️⃣ Add style/tone notes. If you use an AI model with memory, it will get to know your voice over time. Otherwise, give it specific instructions so it can capture what you’re going for.
5️⃣ Ask for a specific output. Tell your AI model how you’d like to receive your results: as a list, a back-and-forth conversation, or a bullet point summary.
Most importantly, I end my prompt with an invitation: "Before you start, ask me any clarifying questions you may have."
💡This instantly transforms AI from a question/answer machine into a thought partner. Rather than prompt engineering, I think of it as conversation engineering.
Once my AI model understands what I want from it, I stack prompts slowly and strategically to refine the output:
Write an outline
Write a first draft
Improve clarity and flow
Add examples and storytelling
For highly complex tasks, I use chain-of-thought prompting, which breaks multi-step processes into smaller pieces, presented to AI in a very specific order. “Come up with insights based on these reports” becomes:
📖 Read these reports
🔁 Find commonalities
📃 Generate a summary
🤓 Pull out insights
The Transparency Imperative
When it comes to AI usage, don’t keep it a secret. The more transparent you are about using AI, the more people trust you.
The landscape of AI is rapidly changing, and trust in its adoption hinges on being realistic about how and when you use AI to support your work. AI isn’t a replacement for human work and ingenuity; it’s a partner, skilled agent, and tool for creativity.
The future belongs to leaders who can harness AI's power while maintaining human trust. That starts with pulling back the curtain. In a world where we're all figuring this out together, honesty isn't just the best policy; it's the only sustainable one.
💭 Your Turn
How are you using AI? What are your favorite hacks and best practices?
What I Can’t Stop Talking About
How AI helped me nail one of the most important speeches of my life: AI isn’t just robots crunching numbers and processing data. It can also be a powerful tool of connection in deeply emotional times. It didn’t write my words, but it organized the feelings in my heart when it mattered most.
AI won’t make you dumber, but bad habits will. A recent study shows that if you’re interested in outsourcing your critical thinking to AI, you’re in for a slippery slope. Trust in AI increases when users treat AI as a partner, not an oracle.
My Upcoming Appearances
Aug 29: Indy SHRM Annual Conference Keynote, Indianapolis, IN
Oct 15: Executive Women's Forum, Keynote, Denver, CO
Nov 13: Brilliance 2025, Celebrating Women Disrupting Healthcare Keynote, Chicago, IL
If you found this note helpful, please forward and share it with someone who needs the inspiration today. If you were forwarded this, please consider subscribing.