How to talk to AI
Artificial intelligence has been a fantastic personal and professional help in my life.
I use AI tools weekly in at least a dozen ways, big and small. They help me communicate more clearly, work more efficiently, make better decisions, and do things that would ld take staff time or skills I wouldn’t have access to otherwise.
Would you help me with something? I’m thinking about doing a Wave Report called “My AI Diary,” in which I document and share how I’ve used AI over the past week, including what worked well and what didn’t. Would that be helpful to you?
If a Wave Report on the AI tools I use and how it’s going is something you’d like to see, just contact me.
Back to today’s topic. Using AI tools, I estimate we add the equivalent of half a person of productivity to our team – and we run a relatively simple operation compared to most organizations.
If AI tools can be so beneficial, why don’t more people use AI tools to transform their work?
Overcomplicating AI
I think we make using AI too complicated, and it prevents people from benefiting from using tools. We have terms like “prompt engineering” that sound like you need a degree to understand.
Above: Industry buzzwords like “prompt engineering” are well-intended, but they can instead put concepts out of reach, keeping normal people from fully utilizing AI. (Credit: marketoonist)
According to Wikipedia, prompt engineering is “the process of structuring an instruction that can be interpreted and understood by a generative AI model.” (Wikipedia)
The problem with concepts like prompt engineering is they become yet another skill people feel they must master before they can use a tool like AI.
My advice is simple – talk to AI like you would any human.
How to Talk to AI
When people ask me how to use text-based tools like ChatGPT (or Copilot, Claude, or Gemini) to get the best output, I advise to keep it simple – talk to it like it’s a friend or co-worker.
In the real world, if you want feedback from a co-worker, you give them all the context they need to provide you with good feedback.
For example, let’s say you seek feedback on an email you’ve drafted. You might tell your co-worker who you wrote it for, what you are trying to convey, any concerns, and what you hope the reader will do with that information. You’d give them a copy of what you wrote, and ask a question, like “How can I make this more clear and effective?”
Like a co-worker, the more specific context you can provide an AI tool, the better. If you give a co-worker generic information, you will get generic responses. The same is true with AI. The more specific you can be about what you are asking for and what you want in return, the more helpful your AI co-worker will be.
If you aren’t sure what would be helpful context, just ask the AI directly, “What else would be helpful to know that would enable you to give me better advice?”
I recently asked ChatGPT to help me brainstorm ideas for symbols to use in the book I’m working on recurring giving in the subscription economy. (I’ve summarized below, but you can see the conversation for yourself here.)
ChatGPT responded with ten symbols – a helpful list that sparked some ideas.
In hindsight, I gave ChatGPT almost no context. It was the equivalent of me calling a co-worker and saying, “I need ten ideas,” but not telling that co-worker why or what it’s for.
I followed up, asking ChatGPT what would help it give me better advice:
ChatGPT helpfully responded:
With ChatGPT’s suggestions in mind, I responded, clarifying several points that I thought would be most helpful:
Sure enough, ChatGPT suggested a whole new set of ideas that were significantly higher quality and more relevant to my specific situation:
If you want to see the entire conversation and list of ideas, you can see the conversation for yourself here.
The bottom line is that by asking ChatGPT what context would be helpful for it to give me better advice, I got more valuable ideas from the tool.
💡 Takeaway: To get the best out of AI tools, talk to it just like you would a co-worker or a friend. This can help you get out of your head and not overcomplicate things.
A note about data privacy – we should always be conscious of what data we put into AI tools. It partly depends on the tool and whether they will use my data to train their model. If you are uncertain about privacy, I have two recommendations:
Most tools have a “data controls” setting or a “privacy mode”/“temporary chat” feature that will not use the data to train the model. Toggle this on.
As a matter of habit, I omit personal or private details if they aren’t important context. For example, I use a generic placeholder like “ABC,” or just initials instead of someone's real name. For example, “Help me draft an email reply to ABC.”
Guide: 12 Ways to Leverage AI in 2024
I wrote a guide highlighting twelve ways you can use AI in 2024 for yourself and your organization. It’s my gift to help leaders grow personally and organizationally.
Throughout this short 20-page booklet, I outline 12 practical ways to use AI and AI tools in your life and leadership with examples, suggestions, and example prompts.
You can download the guide at www.imago.consulting/ai.
💡 Takeaway: Leveraging AI to be more effective and efficient doesn’t have to be overly complicated. My guide 12 Ways to Leverage AI in 2024 is a place to start.
Until next week… Surfs Up! 🌊
- Dave
About the Author | Dave Raley
Consultant, speaker, and writer Dave Raley is the founder of Imago Consulting, a firm that helps non-profits and businesses create profitable growth through sustainable innovation. He’s the author of a weekly trendspotting report called The Wave Report, and the co-founder of the Purpose & Profit Podcast — a show about the ideas at the intersection of nonprofit causes and for-profit brands.
Want to receive insights like this weekly?
Every Friday, we send out The Wave Report, highlighting one trend or insight “wave,” from donor and consumer trends to innovation in different industries or new models for growth.
Subscribe today to receive free weekly insights in your inbox here: