The Extremely Obvious Guide to Using AI (That Most Teams Will Heroically Ignore)
Every company is in an AI arms race right now, blasting out announcements like “✨AI-POWERED✨” even when what they actually built is… a glorified autocomplete or chat bot. Meanwhile, the truly obvious things—the boring, unsexy-but-critical things—are consistently being ignored.
So here’s a list of obvious approaches that everyone knows but almost nobody does. Even with AI. Especially with AI.
Obvious Approach #1: Talk to Your Customers
Yes, I know. Revolutionary. Truly groundbreaking stuff.
You’d think “talk to humans who use our product” would be a given, but somehow this step gets downgraded to:
- one rushed Zoom call,
- three sticky notes, and
- a product manager saying “I talked to my cousin and he said he’d totally use it.”
There is an art to interviewing, and I promise, asking “Would you use this?” is not it. That question is the equivalent of asking someone if they’d go to your next D&D game night—they will say yes out of politeness and then absolutely not show up.
Good research uncovers why customers hire your product. From there, you’ll clearly see where AI can help… or where it's just going to be an expensive gadget you talk about in your board meeting.
Two words: calculated risk.
(Or if you prefer corporate phrasing: “strategic bets with adult supervision.”)
Obvious Approach #2: Keep a pulse on your industry
Take a look around: everyone is building AI tools. Your competitors, your partners, your vibe coding dentist...interesting times for sure.
But don’t panic-build something because the VC group chat is vibing about AI. Yes, the guys with the 💰 are giving you That Look. No, it doesn’t mean you need to slap a chatbot on every screen.
Just stay aware. Notice what’s working, what’s hype, and who’s clearly winging it. (Spoiler: everyone.)
But seriously though—do Approach #1 first.
Kind-of-Obvious Approach #3: Understand the Limitations of Your Platform
If you’re a platform, you’ve probably got data. If you’ve got data, AI might be able to do something interesting with it.
The key question: does your data even make sense outside your own database?
If not, your AI dreams will look like this:
- Step 1: Build AI feature
- Step 2: AI feature realizes your data is spaghetti
- Step 3: AI politely hands that spaghetti back
- Step 4: 😭 (crying face)
I’m not super technical, but I do know that data needs to flow somewhat logically for AI to be of any use.
Kind-of-Obvious Approach #4: An Environment for Innovation
“Be innovative.” In many companies this translates to: “We want something new but also not too new and also zero risk and also done by Friday.”
Real innovation requires an environment where teams can experiment, fail, learn fast, and iterate without being immediately questioned about ROI on Day 2. I think of it like handing someone a box of LEGOs and saying, ‘Be creative—but only use the blue ones and don’t change the instructions.’
I won't go any deeper on this topic but a good reading up on innovation helps. There are hundreds of books out there. That said, even the best book (or consultant) can’t fix a culture that punishes experimentation.
Kind-of-Obvious Approach #5: Tie that Cool Idea Back to Business Impact
When you look at a shiny new idea and think “Ooh!”, take a breath and ask:
“By building or investing in ‘X’, will we improve ‘Y’, which will generate or improve ‘Z’ for the business, and ultimately benefit the customer in some meaningful, non-gimmicky way?”
If the answer is “we’re not sure but it would be very fun to demo,” that’s usually a no.
Unless you need morale. In which case… maybe.
AI isn’t magic, and it won’t fix broken processes, lack of research, or data chaos. But I have a hunch that if you take the obvious steps—actually take them—you can dramatically increase the chances that what you build will matter to someone.
And honestly? That’s more innovative than half the “innovation” announcements out there, no?
