■ SAFE RISK ■ Management & Business
The bottom of the coaching market is being eaten by chatbots that never charge by the hour. The top survives because senior leaders pay for confidential judgement from someone with skin in the game and a reputation to lose.
“AI gives data-driven advice. But who else can charge $1,000/hour to ask 'How does that make you feel?'”
Our AI replacement risk score — how we score jobs
Executive coaching is a mix of structured diagnostics and unstructured conversation. A typical engagement starts with 360-degree feedback interviews, psychometric assessment, stakeholder alignment on goals, then a series of sessions over six to twelve months. Between sessions the coach reads the politics of the client's organisation, holds sponsor conversations with HR and the client's boss, and quietly manages the gap between what the client says they want and what the business needs from them.
AI does more of this than coaches like to admit. Conversational agents provide reflective questioning, goal tracking and accountability nudges at effectively zero marginal cost, and evidence from adjacent wellbeing tools suggests plenty of people open up more readily to a machine than a person. Synthesising 360 feedback into themes, generating development plans, summarising session notes, drafting stakeholder reports and matching coaches to clients are all now software tasks. Large employers are deploying AI coaching to whole management layers that were never going to get a human coach anyway — which grows the category while commoditising its lower half.
The defensible ground is narrow but real. A chief executive deciding whether to fire their co-founder is not going to consult a chatbot with an unclear data policy, and the value of the conversation lies in the coach's read of that specific board, that specific investor, that specific history. Coaches also carry accountability and reputation, which is what makes hard feedback land. Our risk score of 21 assumes the profession bifurcates: AI absorbs the volume tier, human coaches concentrate at senior levels and in team and succession work, and the untrained mid-market coach with a weekend certification has the roughest decade.
Automatability: our editorial assessment of current and near-term AI capability
Already biting at the entry level. Enterprise AI coaching platforms are being rolled out to middle managers now, and by 2030 expect the sub-senior coaching market to be predominantly machine-delivered with human oversight. Senior and board-level coaching stays human well past 2040, but the pool of coaches shrinks toward those with genuine operating credibility or clinical training.
For structured reflection, goal tracking and accountability, they are surprisingly effective and available at 3am. Where they fall short is organisational context — knowing your board, your co-founder history, your industry's politics — and carrying accountability for the advice. Companies are using them precisely where they were never going to fund human coaching anyway.
Generalists with a short certification and no operating background, serving middle managers at modest rates. That is exactly the tier enterprise AI platforms are targeting, at a price no human can match. Coaches with senior credibility, clinical training or a hard specialism are seeing demand hold or grow.
Some already do for low-stakes reflection. But conversations about firing a co-founder, board conflict or a personal crisis carry legal and reputational exposure that makes confidentiality architecture the whole issue. Until a leader trusts where that transcript lives, the sensitive material stays with a human bound by a professional relationship.
Audit your client list for who a chatbot could serve adequately, and assume that revenue is leaving. Push toward more senior clients and team-level work, deepen a specialism, and use AI yourself for the prep and admin so your sessions become denser. Also: get explicit about your confidentiality practices, because clients are asking now.