Coached by Code: What Training My Coachbot Taught Me About Being Human

When I agreed to co-create an AI-powered coaching platform, I thought the most significant challenge would be training the technology. After all, the goal was ambitious: Give employees a way to ask questions and get immediate coaching—anytime, anywhere—based on the expertise of multiple thought leaders, including me. My focus was simplification, innovation, and leading through change.

But as the project unfolded, I realized something unexpected. The real story wasn’t about building smarter technology. It was about what that technology reflected to us about being human.

Why People Open Up to Machines

AI coaching works for a surprisingly human reason: People feel safer talking to it.

Research from MIT, USC, and the CISPA Helmholtz Center confirms what we saw firsthand—people often disclose more to an AI coach than to a human one. With no risk of judgment, embarrassment, or “silly” questions, they lean into candor. AI creates a space where honesty flows more freely, which makes it an unusually powerful tool for growth.

And the benefits don’t stop there. A recent Korn Ferry study showed that 76 percent of global workers are more likely to stay when they have strong development opportunities. AI makes those opportunities widely accessible, extending coaching beyond executives and into every level of an organization.

Even better, the technology can adapt in real time:

• Tailoring guidance to roles, goals, or schedules

• Rehearsing tough conversations with employees or clients

• Giving live feedback in meetings or presentations

• Offering round-the-clock support on challenges like imposter syndrome or giving feedback

It’s not hard to imagine the impact. Someone asks: “How do I disagree with my team without hurting morale?” or “Can you suggest three ways to simplify this strategy presentation?” The AI responds with concrete, contextual advice, echoing the perspective of actual leaders.

That was the vision—and the reason I lent my voice to it.

The Identity Check I Didn’t Expect

However, training an AI coach who sounded like me uncovered more than technical hurdles. It raised deeper questions about identity and integrity.

How could I be sure the advice was truly mine, not a distortion? Was the tone aligned with how I coach in real life—measured, empathetic, focused on lasting change? Could the system inadvertently deliver advice that was biased or cold, even harmful?

One test case still stands out. A leader asked: “My team is resisting a new innovation initiative. What should I do?”

An AI response might be: “Reassign the resisters. Work only with fast adopters.”

On paper, that’s efficient. But it’s not leadership. True innovation is about inclusion, not exclusion—bringing people along, addressing fears, and creating cultural buy-in. Efficiency at the expense of empathy undermines the very change you’re trying to lead.

That’s when it became clear that my coachbot had to reflect not just my knowledge but also my values.

We addressed this by feeding it updated materials, refining language, testing with difficult prompts, and auditing for “hallucinations”—those moments when AI offers confident misinformation. But even more, we worked to layer in empathy, context, and nuance, so the advice wasn’t just smart, it was human.

The Art of Humanizing AI

The lesson? AI doesn’t replace human connection. It needs to be guided by it.

We encouraged users to engage with the coach in ways that sparked richer dialogue. Prompts like:

• “Lisa, what would you say if I feel overwhelmed but don’t want to admit weakness?”

• “Role-play firing someone—with empathy.”

• “Show me how to push back on a senior exec’s bad idea—without burning bridges.”

• “Based on your innovation framework, what experiments should I try with my team this week?”

• “What’s one thing I could eliminate from my workflow to simplify?”

These aren’t transactional questions. They invite simulation, perspective, and emotional intelligence—the ingredients that make coaching effective.

What the Coachbot Taught Me

In the beginning, I thought I’d be teaching the machine. In reality, the machine was teaching me—about leadership, learning, and the nature of coaching itself.

AI coaching isn’t about algorithms. It’s about creating safe spaces for growth. It’s about giving more people access to development, not just a privileged few. It’s about ensuring that what emerges from the code still carries authenticity, humanity, and care.

That paradox is what makes this moment so exciting. The more we use AI to scale learning, the more we’re reminded of the qualities that can’t be automated: empathy, trust, and human connection.

So yes, the future of coaching may be powered by technology. But at its best, it reminds us that the real work—helping people grow into their potential—has always been, and will always be, profoundly human.