[2026 - Week 7]
Welcome to Issue #7 of 3 Things AI - a weekly note where I share three practical or interesting ways I used AI this week.
Thanks for your patience as I missed a week with sickness roiling the family plus a week-long travel for work. But no slow down in the world of AI!
If you missed the last issue, I featured three uses of AI shared by friends and colleagues. One update: I speculated that McGee had used Gemini to build his personal CRM, but after publication confirmed that he used Claude Code. McGee shared a pro-tip: he switched between two of the models - Opus (built for more complex tasks) and Sonnet (designed for everyday tasks) - and let Opus write the instructions for Sonnet to follow, ad then let Sonnet do the mundane work. This minimized token use - as you have to pay extra to use more tokens than your plan allows.
On to this week’s three things AI!
I learned the importance of prompt engineering
A friend and former colleague asked me about using AI to research energy policy in certain states and regions. She was disappointed in the results from ChatGPT and felt like she should be getting better answers. I’m sure the original prompt was more detailed, but still something like “tell me about energy policy in New Jersey.”
This happens constantly. People try AI once, get mediocre results, and conclude it's not that helpful.
But here's what almost nobody does: ask the AI to write the prompt for you.
It’s sort of a meta concept, and to those who already do this, it’s obvious. But it works absurdly well.
Instead of jumping straight to your question, try something like this: "I need to research energy policy trends in New Jersey. Can you write me a master-level prompt that will get me the most useful information?"
The AI will write you something 10x better than what you'd write yourself. Then you just copy that prompt and feed it right back in.
With my friend's energy policy question, the AI suggested breaking it into components: current policies, recent changes, political landscape, implementation challenges. It recommended asking for comparisons and specific examples. It offered several formats for the results. It even asked her a few simple but clarifying questions. Stuff she wouldn't have thought to request.
The results from this approach were completely different: structured, detailed, actually useful. And if you pay for ChatGPT (or Claude or Gemini) you can use Research mode, which will enhance the results even further (note that in research mode the tool will often take 5 - 15 minutes to research and respond to your question).
This isn't some advanced technique. It's embarrassingly simple. But for some reason, people don't do it as often as they should. You don’t have to be an expert at crafting prompts or using these tools - you literally have the smartest assistant at your fingertips, so use it!
Try it this week. Whatever you're researching or working on, ask for a better prompt first. See what happens. You might be surprised how much better the AI gets when you let it teach you how to talk to it.
I asked AI to help me shop for clothes (and it actually worked)
Not sure if I should be embarrassed to admit this one but I’m putting it all out there for my readers. I rarely shop for new clothes, and tend to just wear stuff out before replacing it, and since my wife and I are high school sweethearts, I’ve also come to rely on her advice when it comes to choosing colors or styles (pretty sure I’m not alone on this one).
Well, enter AI.
I found myself alone at a store and in need of some new clothes. I found a few shirts I liked, took pictures of the different colors on the rack, and took a picture of myself trying one on. Fed them into ChatGPT and asked for suggestions on which colors or styles, and even sizes, were best.
Wow. Who knew!
I got recommendations to try different things that I (or my wife) would never have tried. And the language it used was pretty amazing. Some examples:
“Enough depth to contrast your skin.”
“Brightens your face more than navy/black.”
“Works under flannels, vests, quarter-zips”
“This size is giving you clean shoulder lines. No torso ballooning.”
Kind of hilarious but super practical. I walked out with some new shirts and now I’ve got my own personal assistant to help me whenever needed. Try it, you might be pleasantly surprised!
AI took center stage in my company’s strategy session
I participated in my company's 2026 planning session last week and something felt different.
AI wasn't a side topic. It wasn't an "innovation initiative" tucked into the last 20 minutes. It was front and center in our actual business strategy.
I can't share the specifics of what we're planning, but I can tell you this: it's happening. It feels like the shift everyone's been talking about is actually here. Companies are making real decisions about how AI changes what they do and how they do it.
What was most interesting to me was that my colleagues in the room weren't debating whether AI matters. They were figuring out governance, implementation, priorities. How fast to move. What to build first. Who owns what.
It's exciting. Because this is the beginning of something. You can feel it. One person on my team has been building internal tools and customer-facing dashboards with AI; solving problems and delivering features in days that used to take months to do. After an evening spent building with AI, he said “I've never been this excited to come to work."
How often do you hear that?
Years from now we'll look back at 2026 as when businesses actually started integrating AI into their core operations instead of just talking about it.
If your company hasn't had this conversation yet, they will. Get ready.
__________
If you have feedback, questions, or an AI tool you’re enjoying, just reply. I’m always curious what others are using.
I send this once a week. If it’s useful, you can subscribe here → https://3thingsai.beehiiv.com/subscribe