The issue of coach versus AI became one of the hot topics in coaching for 2017 and looks set to continue to generate interest, debate and – hopefully – some research. As yet, there is little evidence that existing coaches have been given the sack and replaced by a bit; nor of coaches working closely with bots in an integrated practice. This is likely to change over the course of 2018 and 2019.

Our take on the topic, from the evidence that is available, is that, while simplistic coaching is under threat, truly developmental coaching can be greatly enhanced by partnerships between coaches and AI. Indeed, we can go further and say that it is in the coaching profession’s best interests actively to encourage such integration.

The tables below provide what we hope are practical descriptions of the strengths of human and AI coaching, both separately and combined. The first table looks at the coach/AI issue from the perspective of the four modes of learning: information, knowledge, skills and wisdom. The second explores the same issue from the perspective of what coaches do and how they do it.


Modes of learning:

Mode Coach-Mentor AI
Information Limited depth, high breadth High depth, limited breadth
Knowledge Limited depth, high breadth Moderate depth and breadth, depending on what databases linked to
Skills Observation and feedback, plus motivation Potentially higher levels of observation (faster and more comprehensive, including micro-movements and tonal analysis)
Wisdom Mentor draw on narratives and values important – the parable as source of learning Currently beyond the scope of AI

Tasks, skills and qualities of a coach

The first part of the table below is based on coaching at its simplest – an expanded view of the GROW model. The more complex the relationship and the nature of the change intended, the less effective an AI will be. However, the AI can continue to add value in partnership with a coach at each level of complexity.

The rest of the table looks at the skills and qualities coaches and AIs bring to the coaching relationship and conversation.


Tasks Coach AI Coach & AI together
Establish purpose and goals Effective coaches work with context and values before agreeing goals


Focus on the goal and routes to achieving it

Unable to work easily with evolving goals

Deeper exploration of context and purpose

Able to look beyond initial goals


Building client self-awareness Uses diagnostics alongside intuition to guide the client towards self-insight.

Builds on insights to shape new horizons

Uses standard tools and questions to help the client become more self-aware. Stops at the point of potential insight. Unable to check how deep the insight has been.


Identify new avenues to explore. When bot brings the client to an insight, it creates the platform for the client to explore it more deeply with the coach
Decision-making & critical thinking More creative, but more susceptible to failures of reasoning and decision-making traps Follows logic and decision-making processes more closely. Unable to include tacit knowledge or “unknown knowns” Better at finding solutions that are both/and rather than either/or
Generating options Intuitive understanding of client of possibilities in light of cultural variables and values; and of what does and doesn’t work Offers both “liner thinking” (obvious) options and “way out” options Coach can moderate and add to AI suggestions to create a wider palette of options. Capacity to be genuinely innovative
Motivating The Pygmalion Effect – motivating power of one person’s belief in what another can achieve Client has to generate own motivation Combining intrinsic and extrinsic motivation
Follow-up Coach acts as a conscience to the client. Difficult to keep reminding the client without appearing to nag and taking responsibility from the client to the coach More rigorous at reminding Easier to monitor progress and give continued support without seeming intrusive
Skills Coach AI Coach & AI together
Listening Has wider store of mental associations to aid sense-making

May filter out important data

Has large, but narrow store of associated algorithms and data to draw upon.

May pay too much attention to irrelevant data

Shifts focus more towards how the client makes sense of their issue
Questioning Intuitive recognition of the “right” question

Intuitive understanding of when not to ask a question

Able to draw upon a large database of questions from previous conversations. Difficulty in deviating from the “script”. Coach spends less time worrying about the next question, knowing that, if they don’t have one, they can fall back on the AI
Rapport building Building deep trust enables the client to delve further into issues and face their fears AI can seem less judgemental, but can only build “transactional trust” A big unknown! However, rapport with a coach may be undermined, if the client suspects “collusion” between coach and AI.  Transparency is vital
Giving feedback Coach gives feedback both on aspects previously agreed and on other things they notice AI gives feedback only on what it is programmed to do. Can make comparisons with other people in its database, to provide a sense of proportion. (E.g. 83% of people fall into this category…) Automating multirater feedback and analysis can put ownership of the process firmly in the client’s hands and suggest topics to discuss in coaching
Use of self In Gestalt mode, the coach is able to use their own feelings and associations to generate new avenues of enquiry AI lacks a sense of self and can only draw upon observation or comparison with other similar conversations Coach can use AI’s observations to check their intuitions. (E.g. when the coach a sense of discomfort, does the AI observe relevant changes in the client’s tone or micro-expressions?)
Being a role model More an aspect of being a mentor than a coach N/A N/A
Qualities Coach AI Coach and AI together
Credibility Combination of who the coach is and the experience they bring — leads the client to place more weight on their guidance The Wikipedia effect – generally helpful but not to be trusted! Likely to increase client confidence – but needs research to verify
Compassion Feeling for the client and understanding their perspective Rudimentary understanding of the emotions people generally in this situation may feel May help the coach avoid over-sympathising and losing their objectivity
Curiosity The instinctive desire to learn more and to follow a conversational path wherever it may lead Algorithms require AI to follow the mist logical path May make explorations more thorough. High potential to ensure that the conversation comes back to “parked” issues that might otherwise be forgotten
Courage The instinct to do or say what feels right N/A AI could potentially act as the coach’s own conscience, prompting them to reflect on their own motivations both during coaching sessions or in reflection afterwards



© David Clutterbuck 2018


This entry was posted in Blogs, Featured Blogs and tagged , . Bookmark the permalink.