There’s an old idea about getting wisdom from a multitude of counselors — you shouldn’t make an important decision off just one opinion, no matter how confident that opinion sounds. Turns out that’s just as true when the opinion is coming from a machine as when it’s coming from a person.
Different AI platforms pull from different information and reason a little differently. Ask two of them the same question and you can get two meaningfully different answers. The first few times that happened to me, it was mildly unsettling. The more I worked with it, though, the more I realized that disagreement isn’t a bug. It’s one of the most useful features available, if you know how to use it.
The Method
Here’s the workflow that’s become second nature: build something in one AI platform — a plan, a piece of copy, a whole system — and get it to a place you’re happy with. Then, instead of just running with it, hand that finished product to a different platform and ask it to find the weak spots.
Not “does this look good,” because everything looks good if you ask it that way. Ask what’s missing. What a skeptical customer or a competitor would push back on. What assumption the first tool made that might not actually hold up.
It works remarkably well. The second platform has no ego invested in the first draft, and because it’s reasoning from a different well of information, it catches things the first one missed entirely — not because the first one is bad, but because no single source, human or machine, sees everything.
This Isn’t a New Idea
It’s the same principle behind getting a second opinion from a doctor, or having a business partner review your numbers before you sign a contract. Multiple perspectives catch things one perspective misses. What’s new is how cheap and fast it’s become to actually get that second opinion — it used to cost a favor, a phone call, a delay. Now it costs a few minutes and a copy-paste.
That’s honestly one of the biggest shifts in how I run things day to day — not that AI gives me answers, but that it makes it trivially easy to get a second answer before I commit to anything that actually matters.
Match the Scrutiny to the Stakes
A caption that’s slightly off costs nothing. A pricing decision or anything legal-adjacent that’s slightly off can cost real money. Scale how many “opinions” you gather based on how much is riding on getting it right — and if two sources disagree, treat that as your cue to verify, not a chance to pick whichever answer you liked better.
This is adapted from Chapter 8 of Small Business, Smart Tools, which also covers what happens when nobody double-checks — including a payroll script that sent the same email nine times in one day.
