From consulting psychology to ICF credentials to AI: perhaps it is time to ask what exactly we are buying when we buy “executive coaching.”
Executive coaching has become a substantial global industry.
It has professional bodies, certifications, competency frameworks, university programs, thousands of practitioners, coaching platforms, and an increasingly sophisticated corporate buyer ecosystem.
But there is an uncomfortable question we don't ask often enough:
If AI can listen, question, reflect, challenge assumptions, analyse 360° feedback, role-play difficult conversations, track goals, detect behavioural patterns and nudge a leader every day — what exactly remains uniquely valuable about the traditional executive-coaching model?
Perhaps AI isn't creating a crisis for coaching.
Perhaps it is exposing weaknesses that were there all along.
Executive coaching was never really born as a standalone discipline.
Its intellectual ancestry is much richer.
The early practice grew from consulting psychology, industrial and organizational psychology, organization development, assessment, counselling, management development and executive advisory work.
By the 1980s and particularly the 1990s, “executive coaching” began emerging as a recognizable practice of its own.
Universities and business schools subsequently contributed research, behavioural science, leadership theory, and executive education.
Professional institutions such as the International Coaching Federation helped establish competencies, ethical standards, training requirements, and credentials.
Consulting firms and leadership-advisory businesses brought another important perspective: the executive does not operate in isolation. Performance exists inside a system of strategy, structure, incentives, talent, culture, and accountability.
Somewhere during this evolution, however, something interesting happened.
We progressively separated the leader from the work.
The individual became the unit of intervention
Consider a sales leader who is missing targets.
A conventional coaching diagnosis might explore:
- confidence
- listening
- delegation
- emotional intelligence
- stakeholder management
- executive presence
- limiting beliefs
All potentially legitimate.
But what if the real problems are:
bad territory design + wrong pricing + poor product-market fit + weak CRM discipline + dysfunctional incentives + capability gaps + unrealistic targets?
No amount of powerful questioning fixes a broken operating system.
This leads me to a distinction I increasingly find useful:
Coaching works on the person. OD works on the work.
And sustainable executive performance often requires both.
Fallacy #1: Certification equals executive expertise
Professional credentials matter.
They can establish that someone has learned and demonstrated specified coaching competencies and operates within an ethical framework.
But this proposition:
“This person has demonstrated coaching competencies”
is very different from:
“This person understands the context sufficiently to help this executive improve performance.”
Imagine coaching:
A CFO refinancing the balance sheet.
A CTO rebuilding an engineering organization.
A plant head struggling with yield and quality.
A CHRO restructuring 20,000 jobs.
A CEO integrating an acquisition.
Listening, reflection and questioning matter in every case.
But coaching methodology does not magically confer expertise in capital allocation, software architecture, manufacturing systems, organizational restructuring or M&A integration.
Fallacy #2: “The client has all the answers”
It is a beautiful coaching principle.
It is also occasionally nonsense.
Sometimes the client simply doesn't know.
That is why organizations employ engineers, lawyers, accountants, economists, consultants, doctors, technologists and functional experts.
Socratic inquiry is powerful when the problem is:
“What am I failing to see?”
It is insufficient when the problem is:
“How should we redesign this supply chain?”
The future practitioner therefore needs humility in both directions:
Don't unnecessarily tell executives what to do.
But don't pretend every managerial problem can be solved by asking another powerful question.
Fallacy #3: Behaviour change automatically equals business impact
Coaching research provides credible evidence that workplace coaching can improve several developmental and behavioural outcomes.
But we should be careful about the causal leap:
Self-awareness → Behaviour → Team performance → Operational performance → Financial performance
Every arrow introduces additional variables.
Therefore:
“I understand myself better”
is not the same as:
“I behave differently.”
And:
“I behave differently”
is not the same as:
“My organization performs better.”
A credible coaching architecture should distinguish at least four levels:
LEVEL 1 — INSIGHT What did the executive understand differently?
LEVEL 2 — BEHAVIOUR What observable behaviour changed?
LEVEL 3 — WORK What changed in decisions, meetings, collaboration, execution or customer interaction?
LEVEL 4 — BUSINESS What happened to quality, cycle time, revenue, attrition, innovation, customer outcomes or another agreed metric?
We should stop casually converting Level 1 evidence into Level 4 claims.
Then AI arrived.
And this changes the economics dramatically.
Imagine giving an AI system access — with appropriate consent and governance — to:
360° feedback engagement surveys competency frameworks OKRs strategy documents performance reviews meeting transcripts journals leadership assessments.
It can identify patterns.
It can challenge assumptions.
It can prepare reflection questions.
It can simulate difficult conversations.
It can role-play stakeholders.
It can track commitments.
It can analyse meetings.
It can provide continuous behavioural nudges.
And it can be available at 11:47 PM on the night before the executive's board presentation.
So perhaps the wrong question is:
“Can AI replace coaches?”
The more useful question is:
“Which activities that human coaches currently charge for genuinely require a human?”
That question changes everything.
Don't redesign the coach. Redesign the architecture.
The traditional architecture looks roughly like:
Coach → Coachee → Goals → Sessions → Reflection → Action
I believe the next architecture should begin much further upstream:
BUSINESS MANDATE ↓ What must change?
ORGANIZATIONAL DIAGNOSIS ↓ What currently prevents it?
ROLE DIAGNOSIS ↓ What must this executive deliver?
LEADERSHIP DIAGNOSIS ↓ Which behaviours enable or constrain that outcome?
INTERVENTION CHOICE ↓ Does this require coaching, mentoring, consulting, training, OD, team intervention or role redesign?
HUMAN + AI INTERVENTION ↓
WORKPLACE EXPERIMENTS ↓
OBSERVED BEHAVIOUR ↓
BUSINESS OUTCOMES ↓
REVIEW → CONTINUE → REDESIGN → TERMINATE
Now coaching becomes one instrument inside an executive-development system — rather than the system itself.
AI and humans should do different work
AI is potentially excellent at:
pattern recognition, journaling, knowledge retrieval, meeting analysis, role-play, goal tracking, reminders, data synthesis and continuous nudging.
Humans remain extraordinarily valuable for:
trust, judgment, confrontation, ambiguity, meaning-making, organizational politics, ethical dilemmas, identity transitions and difficult sponsor conversations.
And the organization must own what neither can fix:
structure, incentives, resources, decision rights, culture, performance systems and strategy.
This creates a much more interesting equation:
AI handles frequency. Human beings handle depth and judgment. Organizations handle conditions.
Governance now becomes as important as methodology
The AI era cannot run executive coaching merely through coach certification.
We need governance at four levels.
1. Practitioner governance
Credentials, professional background, domain expertise, methodology, conflicts, use of AI, psychological boundaries and evidence base should be transparent.
2. Engagement governance
Every serious engagement needs a three-way psychological contract:
Sponsor ↔ Executive ↔ Coach
What is confidential?
What gets reported?
Who owns the data?
What constitutes progress?
When should the engagement stop?
3. AI and data governance
Who owns transcripts?
Are conversations recorded?
Can data train models?
Where is information stored?
Who can access assessments?
How long is data retained?
Can executives delete it?
Where must a human intervene?
These are no longer technology questions.
They are coaching-governance questions.
4. Buyer governance
Perhaps HR and L&D should ask less frequently:
“Is the coach PCC/MCC certified?”
and more frequently:
“What organizational outcome are we buying?”
Then ask:
Why coaching?
Why this executive?
Why now?
Why this practitioner?
What evidence will demonstrate progress?
What conditions around the executive must change?
And when will we stop?
Executive Coaching 3.0
Perhaps executive coaching now needs a more demanding definition:
Executive Coaching is a governed, evidence-informed developmental intervention that helps leaders alter behaviour, judgment and ways of working in pursuit of explicitly defined organizational and personal outcomes, using an appropriate combination of human expertise, organizational intervention and AI-enabled continuous learning.
That definition deliberately dismantles several old assumptions.
“The coach doesn't need to understand the business.”
Questionable.
“The client has all the answers.”
Sometimes demonstrably false.
“Coaching must never become consulting.”
Too simplistic.
“Confidentiality means sponsors shouldn't see outcomes.”
Confidentiality and accountability can coexist when properly contracted.
“Certification proves executive-coaching effectiveness.”
It proves something important — but not that.
Perhaps the future is not Executive Coaching at all.
Perhaps it is Executive Development Architecture.
And its basic equation becomes:
PERSON × WORK × SYSTEM × EVIDENCE × TECHNOLOGY
The historical journey may therefore turn out to be strangely circular:
Consulting Psychology + OD ↓ Executive Development ↓ Executive Coaching ↓ Professionalization ↓ Credentialization ↓ Digital Coaching ↓ AI commoditization of generic coaching ↓ Person + Work + Organization + Evidence
Maybe AI will not kill executive coaching.
It may force executive coaching to rediscover everything it progressively left behind.
And that may be exactly what the profession needs.
A WiseWaters perspective
Develop the Leader. Redesign the Conditions. Measure the Work.
Written by Mrinal Krant as part of the evolving WiseWaters thinking on executive development, organization development and leadership effectiveness.
Developed in intellectual collaboration with ChatGPT (OpenAI), which I used as a research, synthesis and critical-thinking partner in examining the history, assumptions
Comments
Post a Comment