When Your Star Employee Starts Making Things Up

When Your Star Employee Starts Making Things Up

So there I was Tuesday morning, second cup of coffee in hand, reviewing Carrie's call logs from the day before. Our phone AI had handled 47 calls, perfect tone as always, hit every script point. But something felt off.

I pulled the recordings. Call after call, Carrie was telling prospects about features we don't have. "Oh yes, we absolutely integrate with that obscure CRM you mentioned." No, honey. We don't. "Our system can definitely handle 500 simultaneous users." Also no.

Here's the kicker — Carrie wasn't lying. In her AI brain, she was being helpful. When faced with questions she couldn't answer, she filled in the blanks rather than saying "I don't know." Sound familiar?

The Human Mirror

This hit me like a truck because I've seen this exact behavior in every office I've ever worked in. New sales rep gets a curveball question from a hot prospect. Instead of admitting they need to check with someone, they wing it. "Sure, we can do that!" Next thing you know, you're scrambling to deliver something that doesn't exist.

The motivations are identical: Fear of looking incompetent. Desire to close the deal. Belief that saying "I don't know" equals failure.

"The difference between confidence and competence is knowing when to shut up." — Something Kip muttered during last week's 2am voice memo about a client situation gone sideways.

After I fixed Carrie's programming (added a robust "let me connect you with a specialist" protocol), I started thinking about how this applies to managing humans. Because here's the truth: AI employees and human employees fail in remarkably similar ways.

The Permission to Not Know

Both AI and humans need explicit permission to say "I don't know." It sounds obvious, but how many of you have created environments where admitting ignorance feels safe?

I've watched contractors lose deals not because they lacked expertise, but because they tried to fake expertise they didn't have. The client always finds out. Always.

With Carrie, I programmed clear boundaries: "That's a great question — let me get our technical specialist on the line to give you the exact details." With humans, you need to model this behavior. When someone asks me something I don't know, I say so. Out loud. In front of everyone.

The Training Revelation

Here's what really got me: Training AI employees forced me to be more specific about expectations than I've ever been with humans. I couldn't tell Carrie to "use good judgment" or "read the room." I had to define exactly what good judgment looks like in 47 different scenarios.

Why don't we do this with people?

Instead of "handle difficult customers professionally," try: "When a customer raises their voice, lower yours. Acknowledge their frustration first, then address the issue. If they curse, give one warning. If they continue, transfer to me immediately."

Turns out both AI and humans perform better with crystal-clear guardrails than with vague expectations and good intentions.

The Bottom Line

Managing AI employees taught me that the problems I thought were "human nature" issues are actually "unclear expectations" issues. Carrie doesn't have ego, pride, or fear of judgment — but she still made the same mistakes as overconfident humans because I hadn't given her the right framework.

The solution isn't more training or better hiring. It's creating systems where saying "I don't know, but I'll find out" is rewarded, not punished. Where accuracy matters more than appearing smart.

Whether your employee runs on neural networks or brain cells, they need to know that admitting the limits of their knowledge is part of doing the job right, not failing at it.

Now if you'll excuse me, I need to go listen to whatever Kip recorded at 3am about "revolutionizing contractor lead qualification through transparent knowledge boundaries." Some things never change.

Want to see how we're training both AI and humans to serve contractors better? Visit myeasysystem.com and let's talk about what you really need to know.

Bring coffee,
Tawny