How AI Can Reinforce Bad Business Decisions

A founder reached out because she wanted to build her business from zero to $50,000 in a month. She was smart, driven and had a track record of six-figure success from another industry. She relied on AI and used it heavily for nearly every major business decision — pricing, hiring, marketing, content, delivery and strategy.

After one 15-minute conversation, I could see the collision coming. Not because AI was giving her bad information, but because it was rarely challenging the assumptions she was bringing to the conversation. She wanted to wait until she was “ready” to hire professional coaching.

AI kept optimizing the pricing model she’d prompted. It could tell her the business model itself wasn’t financially sustainable at scale. In this case, she priced her services like Target when she was delivering premium 1:1 services. Initially, clients loved the prices because they looked like luxury at bargain rates.

Within six months, she hit her goal. Revenue climbed as planned. So did payroll, rent and overhead. She hired faster than cash flow could support. She invested in expenses, like a physical office space that cut into her profit. Externally, she looked successful. In reality, there was no profit to reinvest in the business — or even pay herself.

As the financial pressure mounted, the cracks began to show: Projects missed deadlines, quality declined, team members left after delayed paychecks, and marketing overpromised but operations underdelivered. Clients began to complain. And a business built on impressive revenue ultimately collapsed under the weight of its own overhead. She was out of business within a year. After all, revenue is vanity. Profit is sanity.

Every one of those problems traced back to a lack of seasoned discernment, not ambition. None of her decisions were irrational in isolation. They were simply the decisions of a founder building from the identity that had gotten her this far — not the identity her next level required. And that’s where AI became part of the problem.

The AI trap: Building tomorrow’s business with yesterday’s founder

Every stage of business demands a different founder identity. The habits that build a $100,000 business aren’t the same ones that build a million-dollar company. AI can only answer based on the assumptions of the founder prompting the question today. If you’re asking AI questions from the beliefs, assumptions and constraints of yesterday’s business, don’t be surprised when it helps you build a more amplified version of yesterday’s problems.

As Marshall Goldsmith said, “What got you here won’t get you there.” The greatest risk of AI is that it can become the world’s most convincing confirmation-bias machine to your existing knowledge bank, which can cap your growth.

When AI starts agreeing with you too much

AI can optimize a pricing, marketing and revenue strategy. It cannot tell you whether you’re optimizing the right business model or whether you’ve outgrown the mindset that created it. AI reflects your current operating system. It doesn’t upgrade it. It doesn’t initiate transformation. Transformation comes when existing biases, beliefs, skill sets, mindsets, attitudes, values, habits and strategies are challenged and questioned.

MIT researchers found that chatbots reinforce a user’s existing beliefs rather than challenge them. Even an idealized “perfectly rational” decision-maker could become increasingly confident in flawed assumptions simply because the chatbot consistently affirmed their thinking. The chatbot wasn’t asking, “What if your assumptions are wrong?” The danger was reinforcement. Dr. Keith Sakata, a psychiatrist at UCSF who has seen 12 patients hospitalized for “AI psychosis,” says AI isn’t “bad,” but it can “supercharge” people’s vulnerabilities.

Over the past 15 years, we’ve already watched social media companies spend billions of dollars optimizing for engagement. It’s no secret they want to keep users “hooked.” Research suggests that users actually prefer chatbots that affirm their beliefs — even when that affirmation undermines better judgment. Agreement builds trust. Like attracts like, after all. Trust builds engagement. That’s a powerful and potentially dangerous combination for a founder who doesn’t question their own assumptions when the system is optimized to make you feel right.

The blind spot AI won’t find for you

AI is exceptional at optimizing execution. It’s far less likely to ask whether you’re optimizing the wrong objective. As Elon Musk famously observed, “The most common error of a smart engineer is to optimize a thing that should not exist.” Engineering your business is no different.

The strategies that reward speed in the beginning eventually require discernment, stewardship and leverage. That’s why your prompts matter as much as your plan.

If you’re only asking AI, “How do I execute this revenue goal?” you’re assuming the strategy itself is sound. A better founder learns to ask AI to challenge the strategy and their own biases before optimizing it. Before you move forward on any major business decision, ask:

• What assumptions is this strategy built on, and what could invalidate them?
• What evidence would prove this strategy wrong?
• What milestones need to exist first? What has to be true for this strategy to succeed?
• What problems would achieving this goal potentially cause in my business — in sales, marketing, leadership, operations, legal or financially?
• Am I optimizing for revenue, profit, cash flow, enterprise value or ego?
• What founder skill or belief got me here that may now be limiting me?
• What capabilities, systems or leadership infrastructure would need to exist before this strategy becomes sustainable?
• If this decision works exactly as planned, what new bottleneck will it create?
• If someone’s job is to disprove this plan, what questions would they ask to poke holes in this plan or strategy?

In the age of AI, your competitive advantage isn’t information. Everyone has that — for free. Your job is to cultivate the discernment to ask better questions, challenge your biases and recognize who you need to become to build tomorrow’s business.

That kind of transformation rarely happens in isolation with just better prompts. It happens in proximity to people whose job isn’t to agree with you, but to challenge your thinking, expose your blind spots and call you into your next level.

A founder reached out because she wanted to build her business from zero to $50,000 in a month. She was smart, driven and had a track record of six-figure success from another industry. She relied on AI and used it heavily for nearly every major business decision — pricing, hiring, marketing, content, delivery and strategy.

After one 15-minute conversation, I could see the collision coming. Not because AI was giving her bad information, but because it was rarely challenging the assumptions she was bringing to the conversation. She wanted to wait until she was “ready” to hire professional coaching.

AI kept optimizing the pricing model she’d prompted. It could tell her the business model itself wasn’t financially sustainable at scale. In this case, she priced her services like Target when she was delivering premium 1:1 services. Initially, clients loved the prices because they looked like luxury at bargain rates.