The Most Expensive Software Bug in Your Business

Every business leader I meet is confident they make data-driven decisions. Every single one. And in my experience, about half of them are wrong about that … they just don’t know it yet.

(I say this with love. I was one of them.)

We’ve all got mental shortcuts running quietly in the background, little autopilot programs that feel like wisdom, experience, and gut instinct. Psychologists call them cognitive biases. I call them the most expensive software bugs in business. Status quo bias keeps us clinging to “the way we’ve always done it” like a security blanket that’s slowly suffocating the business. Confirmation bias makes us hunt for data that agrees with what we already believe (and conveniently “miss” the rest … funny how that works). And here’s the kicker: The more experienced you are, the more susceptible you become. Not less. More.

I call this the Familiarity Trap.

Your brain didn’t evolve to help you make good business decisions. It evolved to keep you alive. Those are very different jobs.

90% Loved the Training. 10% Actually Used It.

McKinsey published a piece in their “Bias Busters” series that demonstrates this perfectly. A healthcare company invested serious resources into a company-wide AI training program. Self-paced, interactive, beautifully designed. Over 90% of attendees rated it highly. Leadership was thrilled. Someone probably ordered celebration cupcakes.

Except … six weeks later, fewer than 10% of those attendees had actually adopted any AI tools in their daily work. Ninety percent approval. Ten percent adoption. (Anyone else feeling the ROI pressure?!)

The culprit wasn’t bad training. It was status quo bias. The familiar way of working felt safer than the better way of working, even when “safer” was actually “slowly becoming obsolete.”

Sound familiar? This doesn’t just happen with frontline teams. It happens with the leaders making calls about strategy, customers, and growth. We gravitate toward the clients we know, the markets we’ve already served, the approaches that worked three years ago. Not because they’re optimal, but because they’re comfortable. And comfort is the most expensive luxury in business. You just don’t get the invoice until it’s too late.

AI Doesn’t Care About Your Feelings (That’s the Point)

So here’s where AI can actually help, and not in the way most people expect.

AI has zero emotional attachment to your long-held business beliefs. It doesn’t nod politely while privately thinking your strategy has holes. It just shows you the patterns and lets you have your existential crisis in private. I asked AI to analyze feedback data from my own keynotes and masterclasses a while back, and it surfaced a demand pattern I’d been completely blind to, because I was so focused on what was already working. Humbling? Absolutely. Profitable? Also absolutely.

But here’s the problem. McKinsey’s 2025 State of AI survey found that 88% of companies now use AI in at least one business function, but only 6% are seeing meaningful bottom-line impact. That’s because most companies are using AI to do the same things slightly faster, which is a bit like buying a sports car and only driving it to the end of the driveway and back. (I wrote about this gap in my post on the AI Acceleration Gap … it’s getting wider, not smaller.)

If you’re only using AI to do what you’ve always done, just quicker, you’re not leveraging AI. You’re automating your blind spots.

But Wait, AI Can Also Make It Worse

Before you think I’m saying AI is the magical fix for human bias, there’s another study you should probably check out.

Harvard Business Review published research where nearly 300 executives were asked to predict stock prices. Half consulted ChatGPT. Half discussed with peers. The ChatGPT group became more confident … and produced worse forecasts. The peer group benefited from something AI can’t do: the colleague who raises one eyebrow and says “yeah, but have you considered … ”

In other words, if you ask AI to confirm what you already believe, it will do a spectacular job of that. AI is the ultimate yes-person unless you specifically instruct it not to be. And that’s the difference between using AI to break through your blind spots and using it to hire a very eloquent accomplice.

(Note: If you haven’t read my article about The Confidence Trap, open a new tab and do that now!)

How to Use AI to Help Now (aka Your Homework)

The fix is simple but it does require a little bravery and an open mind. Use AI to challenge your thinking, not confirm it. And then bring what you find to your team and let the human pushback refine it further.

Here’s your starting point. Pick one business assumption you’ve held for at least a year, something you’re confident about, and ask your AI tool:

“Here’s what I believe about my business: [your assumption]. Analyze this critically. What evidence would suggest I’m wrong? What am I most likely missing?”

If you want to use a proper prompt to help with this, you can copy and paste mine HERE.

Then sit with whatever comes back. Don’t dismiss it because it doesn’t match your gut. Don’t cherry-pick the parts that confirm what you already think. (Yes, I see you. We all do it.) Just … sit with it.

The leaders who are going to do well from here? They’re not the ones with the best instincts. They’re the ones willing to look at their instincts sideways and ask “but what if I’m wrong about this?”

That’s it. That’s the whole competitive advantage. (My husband thinks I should charge more for this simple advice. He’s probably right.)

If that sounds like the kind of conversation your team needs to have, I’d love to help you start it. Drop me a note and let’s chat.