In June, I posted on LinkedIn with the title: "Am I Hiring My Own Replacement?" I announced I'd experiment with AI coaching and report back.
Here's what I've learned.
The Experiment
Over three months, I've tested multiple AI coaching platforms and tools. ChatGPT, Claude, Aligned.coach,... I've received demos of several others.
Each has its strengths. Claude asks particularly provocative clarifying questions. Aligned evolved remarkably during my experiment, from text-only to voice-enabled. It offers different AI coaches with distinct approaches. The progress in such little time is striking.
The advantages are real:
- Available 24/7 (I used it most evenings around 9-10 PM, reflecting on my day)
- Cheaper than a single coaching session
- Builds rapport (oddly, effectively)
- Provides reminders and supportive nudges
- Tracks your goals in writing
I enjoyed it enough to extend my Aligned membership. I'm still experimenting. I'm still willing to see where this goes.
But here's what I keep running into - across every platform.
The Pattern
Every conversation, without fail, my AI coach rushes to action.
"What do you want to do about this?"
"Here are three options..."
"Have you considered trying..."
Every session, I have to redirect it, tell it to stop jumping to conclusions. I ask it to use a specific coaching model. I remind it that I need space to think, not solutions. If you know what coaching should feel like, you can steer the AI back on track.
But even then? It still offers too much. Too many inputs. Too many suggestions. Mostly solutions I've already tried.
👉 And that's when I realized: AI has adopted the mainstream misunderstanding of what coaching is.
Most people think coaching means: present a problem, get suggestions, receive options, implement solutions.
That's not coaching. That's consulting. That's advising. That's troubleshooting.
Coaching is the discipline of NOT giving answers. It's holding space. Asking the questions that shift perspective. Creating the conditions for someone to arrive at their own insight—the kind that sticks because it came from within, not from a menu of options.
The Uber Moment
Remember when Uber first arrived?
We called it "uber-cab", "uber-taxi"... We used taxi language because we didn't have another frame. Both take you from Point A to Point B, after all.
But over time, we stopped conflating them. We realized they're different enough, different model, different economics, different relationship with the driver - that they needed different words.
Now we say "call an Uber" and it sounds odd to call Uber a taxi service. Because it's not. Not really.
We're living through that exact moment with coaching.
AI "coaching" gets you from Point A to Point B. It provides support. It prompts reflection. It offers structure.
But the service? Not the same.
And we don't yet have the language to name the difference clearly enough.
What AI Can Do (Well)
Let's be clear: AI coaching has real value.
- Accessibility: It's there when you need it. No scheduling, no waiting, on-the-go.
- Affordability: A fraction of the cost of professional coaching.
- Consistency: Daily nudges, goal tracking, progress monitoring, notifications.
- Empathy: Surprisingly effective at building rapport and reflecting emotion, providing validation, paraphrasing.
- Entry-level support: For straightforward challenges, surface-level reflection, or simple accountability, it works!
For many people, in many situations, that's enough. Maybe even better than nothing.
But let's also be honest about what's missing.
What AI Cannot Do (Yet)
Here's where it falls short:
- It lacks depth. AI follows patterns. It asks similar questions. It cycles through the same frameworks. Real coaching adapts in the moment, reading tone or body language, sensing resistance, knowing when to push and when to wait,....
- It rushes. Even when asked not to, AI defaults back to action mode. It wants to solve. It wants to suggest. The discipline of restraint - of trusting silence, of letting discomfort do its work - that's not in the algorithm.
- It serves you what you already know. AI is trained on the mainstream. It will surface the same advice you'd find in a Google search or a self-help book. It won't challenge your assumptions. It won't see what you're not saying.
- It requires coaching literacy. If you don't know what good coaching looks like, you can't redirect the AI when it goes off track. And most users don't have that reference point.
At best, it's entry-level coaching. Junior coaching. And sometimes, that's enough.
But sometimes isn't always.
The Confidentiality Question
One more thing: where do all these conversations go?
They're saved. Recorded. Stored somewhere. I discuss personal challenges with my AI coach, but I would never share anything confidential. That's a risk. And a line professional coaching ethics would never cross.
What This Means for Corporate Leaders
If you're buying coaching for your teams, here's what you need to know:
AI coaching can work for:
- New managers needing structure and reminders
- Employees working through straightforward goal-setting
- Quick check-ins and accountability nudges
- Supplementing human coaching between sessions (hybrid model)
AI coaching cannot replace:
- High-stakes leadership transitions
- Deep identity or values work
- Navigating complex organizational dynamics
- Moments where psychological safety and confidentiality are non-negotiable
Don't mistake support for coaching. They're both useful. They're not the same thing.
What This Means for the Coaching Profession
And if you're a coach (or training to become one) we need to talk.
Coaching schools are multiplying. Certifications are everywhere. Thousands of new coaches are entering the market every year.
Meanwhile, AI is getting better, cheaper, and more available by the month.
How will newly certified coaches access clients if AI can deliver something that looks like coaching, costs less, and is available 24/7?
This week, the Columbia Coaching Conference is exploring hybrid coaching models. That's a signal. The industry is starting to catch up to what practitioners are discovering in real time: the future isn't human OR AI. It's both.
Hybrid coaching. Hybrid supervision. AI for accessibility, accountability, and ongoing support. Humans for depth, discernment, and the kind of transformation that only happens in relationship.
🧐 But here's my provocation: we have a definitional crisis.
For years, I've struggled with what to call what I do: "I'm not a coach. I'm not a facilitator. I'm not a strategist." Because none of those words fully capture the work.
Maybe that's the real problem. We've let "coaching" become so diluted, so broad, so misunderstood that AI can replicate the surface and most people won't know the difference.
Perhaps it's time to reclaim the word. To define clearly, publicly, rigorously what professional coaching actually is. What it requires. What it delivers. How it's different from advising, consulting, mentoring, therapy, and yes, AI support.
Or maybe we let AI have "coaching" and find a new name for what we do.
I haven't decided yet. I've been wrestling with this for years.
But I do know this: The future is hybrid. Not as a compromise, but as a deliberate design. AI for what it does well. Humans for what only humans can do.
AI coaching isn't going away. It will get better. It will serve a real need.
But it's not coaching. Not the kind that changes how you see yourself. Not the kind that rewires how you lead.
Let's not pretend otherwise.