A lesson with a good golf pro costs $100 to $200 for 45 minutes. Most serious amateurs take one a week if they're committed. But that same golfer also hits the range three or four more times on their own — an hour each, no guidance, no feedback. Roughly 80% of their actual practice time happens without any expert present. That ratio is the problem.
The Coaching Gap
Here's what actually happens during those solo sessions. You show up, buy a bucket, and start hitting. Maybe you remember a thing or two your pro said last Tuesday. By the thirtieth ball, you've drifted. You're working on something, but it's probably not what your coach prescribed. There's no feedback loop. No correction. No one to say "you're rolling your wrists again." You hit 100 balls, feel productive, and leave. Some of those reps made you better. Some made you worse. You have no way to know which.
This isn't a golf-specific problem. It's the same pattern everywhere in skill development. The expert is expensive, scarce, and available for a fraction of the time you actually need guidance. Physical therapy patients do their rehab exercises at home with a printed sheet of paper. Music students practice between lessons with a vague memory of their teacher's instructions. Language learners drill alone with no one to correct their pronunciation. The most critical practice hours — the ones where habits actually form — happen without any expert present.
What OnPlane Built
I wrote about the experience of building OnPlane in a previous post, so I'll keep this brief. OnPlane is a voice-first AI golf coach that lives in your AirPods. You talk to it while you practice — describe what just happened, ask what to work on — and it responds with real coaching advice in natural speech. It watches your swing through your phone camera. It remembers your tendencies across sessions. It's there every time you practice, not just once a week.
What interests me here is not the technology. It is the economics.
AI Replaces Absence, Not Expertise
The most common framing for AI in professional services is replacement. AI will replace lawyers, doctors, coaches, teachers. I think this framing is almost entirely wrong, and building OnPlane made that obvious to me.
A good teaching pro does things no AI can do right now and probably won't for a long time. They read body language. They understand where your head's at that day. They make judgment calls about when to push and when to back off. They physically reposition your hands on the club. The value of a great coach isn't information delivery — it's the full-spectrum human interaction of teaching.
But that coach isn't on the range with you on Wednesday afternoon. Nobody is. You're alone with a bucket of balls and your best intentions. That's what AI replaces. Not the coach. The absence of the coach.
The real competition for AI coaching is not the $200/hour lesson. It is the $0/hour of practicing alone with no feedback at all.
This reframe changes everything. You're not asking golfers to fire their instructor. You're offering them something useful during the 80% of practice time that's currently unguided. That's a completely different conversation.
The Pattern Beyond Golf
Once you see this pattern, you see it everywhere. Anywhere expert guidance is expensive, expert availability is scarce, and people spend significant time practicing or performing without feedback — there's the same gap.
- Physical therapy. A PT visit costs $150-300. Patients do most of their rehab exercises at home, unsupervised, often with poor form. An AI that watches via phone camera and corrects form in real time is not replacing the physical therapist -- it is replacing the nothing that happens between appointments.
- Music practice. Private music instruction runs $60-120/hour. Students practice daily, alone. An AI that listens and provides real-time feedback on timing, pitch, and dynamics fills the same gap.
- Language learning. Tutors charge $40-80/hour. The student needs thousands of hours of conversational practice. AI conversation partners are already proving this model at scale.
- Sales training. Onboarding a new sales rep costs companies thousands in coaching time. Reps make dozens of calls a day with no real-time guidance. An AI coach in-ear during calls is not replacing the sales manager -- it is being present when the manager cannot be.
In every case, the AI isn't competing with the expert. It's competing with nothing. And competing with nothing is a pretty good position to be in.
Growing the Pie
Here's the part that surprises people: I think AI coaching will make the human coaching market bigger, not smaller. Students who practice with AI guidance between lessons improve faster. Faster improvement means more engagement with the sport. More engagement means more willingness to invest in premium human instruction. The AI raises the floor, which expands the market for everything above it.
This is the opposite of the "AI replaces jobs" narrative. A golfer who gets better faster doesn't stop taking lessons — they take more lessons, because they're engaged enough to care about the next level. The AI doesn't cannibalize the pro's business. It creates a more invested, more serious student who values that human time even more.
The Bigger Idea
What I keep coming back to is how many domains have this same structure. Wherever there's an expensive expert who's only available a fraction of the time someone actually needs guidance, there's an opportunity for AI to fill the gap between sessions. Not to replace the expert. To be present when they can't be.
That's a very different starting point than "how do we automate this person's job." It leads to different products, different business models, and different relationships with the existing ecosystem. You're not fighting incumbents. You're serving a need that was previously unserved because the economics didn't work.
At BCK Systems, this is one of the lenses we bring to AI engagements — whether we're automating workflows, streamlining operations, or building something entirely new. The best AI projects start by asking where real value is being left on the table. Sometimes that's a process that's too slow. Sometimes it's expertise that's needed but unavailable. Either way, the answer usually points to something worth building.
If you're working on something in this space, I'd like to hear about it.