At two in the morning, someone lying awake opens an app instead of calling a friend.
They type out the thing they have been avoiding at work, and within seconds they have a structured reflection, a set of pointed questions, and a plan for the morning. It costs less than a coffee subscription, and it never judges, never cancels, never gets tired of hearing the same worry for the third week running.
A few miles away, a senior executive who is about to remove her closest co-founder from the business pays several hundred dollars an hour for ninety minutes with a human coach who has sat across from this exact kind of decision before, who will hold the discomfort with her rather than resolve it too quickly, and who she trusts enough to say the hard thing out loud before she has to say it to anyone else.
Both of these are casually described as coaching. The popular assumption is that they are the same service delivered through two different channels, and that the cheaper, faster, always-available one will eventually absorb the other, the way digital retail absorbed the corner shop. Look closely at what is actually happening across the coaching industry, and a different picture appears. Rather than artificial intelligence hollowing out human coaching, the two are pulling apart into distinct products serving distinct needs, and the human side of that split is becoming more valuable, not less.
The coaching profession, in other words, is not being replaced. It is being sorted, and the sorting is moving faster than most people, including many coaches, appear to have registered.
The Two Kinds of Advice Nobody Used to Separate
For most of its short professional history, coaching has been sold as a single category: a trained outsider helps a person get clearer on a goal and holds them accountable for pursuing it. In practice, that category has always concealed two very different jobs. One is largely mechanical: structuring a goal, tracking progress against it, prompting reflection on a schedule, offering encouragement when motivation dips. The other is fundamentally about judgment: helping someone weigh a genuinely uncertain decision, read a political situation they cannot see clearly from inside it, or say a difficult truth that a person’s friends and colleagues have been too polite to say.
These two jobs were bundled together for decades largely because there was only one way to deliver either of them, a human being, on a call or in a room. Once that constraint disappears for one half of the job, the bundle stops making sense.
What Software Is Actually Good At
The mechanical half of coaching turns out to be close to what current AI systems already do well. Structured goal tracking, habit prompts, journaling analysis, and round-the-clock encouragement do not require lived judgment so much as consistency, patience, and the ability to notice patterns across thousands of small data points, which is precisely where software has an advantage over a human who can only hold one client’s context in mind at a time. It is genuinely useful, which is why it has spread so quickly. Within a short span of years, tools offering some version of AI-guided coaching have gone from a novelty to something a majority of practising coaches themselves now build into their own services, not as a threat to route around but as infrastructure to work alongside.
This has had a real effect on the lower end of the market. Generic, undifferentiated encouragement, the kind that used to justify a modest hourly fee, is now abundant and often free. Coaches whose entire offer was warmth and accountability, with no specific domain of judgment behind it, are finding that offer harder to sell than it was even recently.
Why Scarcity Is Growing at the Top, Not Shrinking
The more interesting effect is showing up at the other end. As the mechanical layer becomes commoditised, three rough tiers are emerging where there used to be one blended market. At the bottom sits credential-light, experience-light coaching competing purely on price, the segment most exposed to being absorbed by AI tools directly. In the middle are competent generalists who use AI to extend their reach while still relying on a broad, non-specialist skill set. At the top sits a smaller group of specialists with deep, specific experience, an industry background, a particular kind of transition, a particular kind of crisis, who increasingly price themselves less like a service and more like an advisor, sometimes even sharing financial risk in the outcomes they are hired to influence.
That top tier has not become less valuable as AI has spread through the rest of the market. If anything, demand for it has grown, partly because the noise created by cheap, AI-generated coaching content has made undifferentiated advice harder to trust in general, and partly because the population of people facing genuinely complex, high-stakes decisions, career transitions, leadership pressure, identity shifts in later life, has not shrunk at all. What has changed is how that top tier proves its worth. Certifications alone are increasingly treated as a baseline rather than a differentiator, and documented outcomes in a specific domain are becoming the real currency, which is a meaningfully higher bar than the industry has historically asked of itself.
The Businesses Being Built on Either Side of the Split
This stratification is already shaping how money moves through the industry. Considerable investment is going into AI-driven coaching infrastructure built for enormous scale, embedded in workplace benefits platforms, wellness apps, and career tools that can deliver the mechanical half of coaching cheaply and consistently to very large numbers of people. A separate, and in some ways larger, current of investment is going into positioning human coaches explicitly as a premium layer above that infrastructure: credentialing bodies raising the bar on rigor and ethics, corporate programmes reserving human coaching for senior leaders and pivotal transitions, and specialist practices marketing themselves not as generalists but as the specific person to call for one high-stakes kind of decision.
Both currents are growing at once, which is unusual and worth paying attention to. It suggests this is not a market being disrupted in the conventional sense, shrinking on one end as it grows on the other, but a market being restructured, expanding at both extremes while the vague, undifferentiated middle empties out.
Choosing Which Layer to Build For
The practical implication for anyone with a stake in this industry, coaches themselves, the bodies that certify them, and the employers who fund coaching as a benefit, is that the strategies which win in one layer are close to the opposite of the strategies that win in the other. Winning the mechanical layer means going broad, cheap, and consistent, competing on accessibility and scale. Winning the judgment layer means going narrow, expensive, and provably effective, competing on trust and specificity, not on the number of letters after a coach’s name. Trying to do both at once, offering generic encouragement while also claiming elite judgment, increasingly satisfies neither audience.