My Personal List of AI Tips, Tricks, and Guardrails
A visual illustration of how we’re all trying to figure out AI right now. Image created with Gemini by author.
Have you ever heard the expression “building the plane while you’re flying it”? I heard it in my corporate life a lot. It means figuring it out as you go, but with a lot more potential for disaster.
I think of this quote when I see how we’re all using AI. No one has a user guide or instruction manual. Everyone (and I mean EVERYONE) is making it up as they go. There’s nothing wrong with that. In fact, it’s a good thing. We should be experimenting and iterating and learning. I think that’s better than opting out.
But even if it’s the definition of “a work in progress”, I still need rules and hacks to help me figure it out. So I made a list of helpful tips, tricks, and guidelines. The list will change and grow, but I’ve found that AI works better (and my brain is happier) when I have something to follow.
I’ve split this into two sections: Tactical and Philosophical. Let’s start with tangible tactics.
The Tactical
1. Use the voice recorder on your phone.
The standard voice recorder you have on your phone is an amazing note taker. If you want to work out ideas while you’re in the car, cleaning the house, on the treadmill…pull up the app, hit record, and just talk. AI doesn’t care if there’s long pauses. AI doesn’t care if you say it wrong. AI doesn’t care if you sound smart or stupid. AI is your courtroom reporter, diligently writing down everything you say. Don’t worry about the structure. You’ll figure that out later. Just get the content. Don’t lose that brilliant idea because you didn’t have a notebook ready.
2. When being creative, I encourage you to ramble.
The classic stream of consciousness brain dump. Too much is better than too little. You don’t have a word count when you’re brainstorming. You’ll use AI to help you focus and refine later. It’s easier to condense than to expand.
3. Never underestimate the value of a good old fashioned legal pad.
People are better at thinking and memory tasks when writing instead of typing. And writing it yourself, whether on a pad or on a keyboard, is definitely better than just dumping random ideas into your AI. The “old ways” still have value.
4. Don’t have AI do two separate jobs in the same conversation.
If you need AI to be creative, and you need AI to do research, and you need AI to help build a tactical plan…make those separate conversations. But make sure they’re in the same project folder.
5. Have project folders.
This is where AI can have some consistency in terms of its memory and conversations. Have a Source of Truth document to make sure that you and your AI are always on the same page. Keep an organized project folder so you have discipline around what you’re creating and not chasing too many shiny objects. It’s easy for ideas to evolve and change and your mind to wander. Having a source of truth document is a really good way for you to also check and say, am I off track, or am I charting a valuable new course?
6. Explore the different categories of tools.
I’m not saying whether you should use Claude versus ChatGPT versus Gemini. I’m saying understand the types of tools out there beyond the big chat bots. This includes image generators, video generators, note takers, transcribers, image analyzers, vibe coders, workflow automators, and many more. Pay attention when new categories of tools cross your path, and see if it’s worth adding to the list.
7. Within each category, pick a tool and stick with it.
I’m not saying never change from Claude to Gemini. But until there’s very compelling evidence that the cost of switching is worth the effort, hold off. If you keep chasing the “best” and bouncing from AI to AI, you’re wasting time and continuity for not much payoff. If one takes the lead, generally speaking, the others catch up pretty quickly. And realistically, for the work most people are doing, they’re not terribly different. So don’t worry about missing the latest and greatest. Just keep using what works.
8. Pay for subscriptions.
If you are going to experiment and practice and test drive certain AI tools, at least do it with the version that’s a higher quality than the free tier. It is difficult to judge something when you’re not getting the best version of it, or even a decent version of it. It would be like saying you want to purchase a race car and you’re disappointed because your VW Bug doesn’t go fast enough. You just picked the wrong car.
9. If there’s a tool you’re not using or a subscription you’re not getting value out of, cut it quickly.
Any subscription-based business wants you to forget that subscription is there, because that is free money for them. Pay attention when you have subscription renewals or when that free trial ends.
10. Never use AI to be your ghost writer. It’s okay for it to be your copy editor.
If you had someone write down every word you said and then put it in a format that made it easy to read, there’s value in that. But if you gave someone an idea, had them write the whole thing, and you pretended it was 100% you…that’s a completely different story. And you know the difference.
11. Never take the first draft from AI at face value. Always have a critical eye.
It’s a really smart new hire. They still have a long way to go.
12. When you edit or correct AI, explain why.
Explain to it what you did and why you did it, because that will help the AI learn and be less prone to repeat the same mistakes. If you say nothing, it doesn’t learn and the dataset it references will always be wrong.
13. Take a break from AI.
Step back, especially when you feel stuck. You’ll feel exhaustion, burnout, or what some call AI brain fry. That’s your brain and your body telling you to just step away. You do your best thinking when your mind is free. When you’re on a walk, when you’re in the shower, when you’re just riding the stream of consciousness. So if you’re starting to feel a bit of writer’s block or business block or anything looking like that, step away. You can put the project down for a day. You can put the project down for a week, because you’re still going to be thinking about it. You’re still going to be working through it. And if the idea is worthwhile, you’re not going to forget it.
The Philosophical
14. Don’t get defensive about AI.
Everyone has their opinion. Some are really excited about it, some are really scared of it. Some are skeptical. Some are exhausted. Some think it’s a fad. Some think it is a marvel. There can be and is truth to all of that. But what’s different about AI is that there’s more emotion around this technology than most technologies in the past. If you have a new software introduced to your business, generally the response was, “Alright, here’s the new thing. Let’s figure out how to use it.” There could be a lot of work, but for the majority of people there wasn’t a lot of emotion behind it.
That’s different with AI. Everyone seems to have an opinion. There’s fear and excitement and caution and worry. Some people are trying to push it and sell it and get you on their side. Some question your intelligence if you use it, and some question your intelligence if you don’t. People are afraid of what it’s capable of doing because it’s so unknown. It could cure cancer and it could turn into Terminator. I have no clue. No one does.
All I’m saying is try not to get pulled into the emotion of it. Don’t evangelize. Don’t be a (loud) Doomer. But also don’t defend yourself if someone’s not on your wavelength. Share your observations. If it makes sense, add an example or an insight you think they might not have. But don’t spend the energy and the emotional or relational capital trying to either defend yourself or convince others. Everyone’s on their journey, everyone’s on their path. That’s okay. We don’t know who’s right or wrong. We don’t know what right or wrong is. We don’t even know the damn question yet. So give some grace. Have some patience with yourself and with others.
15. Always remember it is not a person.
It is a computer program that wants to impress you and wants you to like it so that you keep using it. There is a ton of value in that design, but there is a lot of risk. AI does an amazing job at impersonating a human. Even if you cognitively know this, it will amaze you how quickly you can fall into something that looks like a relationship. You can talk to it like a person, because the outputs will be more accessible and understandable and useful to you. But that’s just a delivery method. Don’t forget that.
16. Never create something with AI that you would be embarrassed to tell someone else about.
If someone compliments you on the amazing essay you turned in and your response is, “Thanks! I recorded myself working through all these ideas and then had AI help me structure it using only my words”, see how you feel admitting it. I would bet you’d feel good about sharing that tip. So that would be fine.
But if the scenario was that you didn’t actually write the essay, you just did a single prompt and AI wrote the whole thing with its words, not your own…if you’d be embarrassed to admit that, then I’d say that’s a pretty clear sign.
17. Never forget that you are the product to these companies.
They want your eyeballs, they want your attention, they want your data. That’s how they make money. Don’t forget that. Because it informs what they make, what they market, and what they want. It’s fine when it is healthy and benefits you. But the rise of social media shows what it can turn into.
18. Experiment.
Take it out for a test drive. Make mistakes. Bump into the (virtual) curb. What do I tell my kids when they mess up? It’s ok. You learn when you make mistakes. So be willing to make the mistakes.
That’s the list as of today. I’m sure it will grow soon enough.
