Are human coaches headed for the same disruption experienced by human travel agents?  Should you start asking for recommendations on coding boot camps or career coaches? Before making any of these leaps, let’s start with how digital coaching may develop and reconfigure today’s coaching market.    

What is a Digital Coach? 

You’ve probably already shouted at - or thanked, on your better days - the precursors to digital coaches: Siri, Cortana, Alexa, Echo and Google Home. These digital assistants, as you find out on your tougher days, can handle only a relatively narrow range of questions and tasks. That’s “narrow” artificial intelligence (AI). 

Today’s digital agents will progress to successfully handle more complex questions. Agents will ask questions to get information about you, enabling them to offer a response that you consider useful and worth your time investment. Digital agents need our input to grow their knowledge base through machine learning. We will continue to make a familiar trade-off that we already make many times each day: we provide information about ourselves in return for a response or service that is useful and personalized. But the data we will share with digital coaches in the future is more than our preferred flight departure times. Instead, we’ll share qualitative, emotion-laden, messy data about our preferred lives and work.

In designing an AI-driven coaching agent, the design goal is to create the conditions that enable a coaching conversation, not to replicate a specific relationship.  The perceived quality of the conversation will rise as digital agents gain greater capabilities in processing natural human language and deriving the meaning of our input. Through humanizing features such as appearance, gender, voice and facial expressions, we can expect conversations in which we feel comfortable disclosing more about ourselves. As emotion recognition capabilities become available on consumer-friendly devices, and agents demonstrate that they “get” us, we can also expect to feel heard and understood.     

Our sense of being heard and understood will also depend on the quality of the agents’ input to these coaching conversations. The agent will use a rules engine to make selections from its knowledge base and experience. For example, the software will be programmed to ask open-ended questions, paraphrase user responses and offer the observations and follow on questions a human coach typically would.  

This is the speed bump in the development of digital coaching. Technology such as AI is just not good enough yet at what we associate with coaching – making moment to moment nuanced judgments about what is best for a client at a specific point in a coaching conversation. (Gratch, 2017). Machines learn and apply what they learn to the scope for which they are programmed. But, AI will evolve to take on wider knowledge sets and more complex interactions. And, digital assistance will become even more embedded in our lives. Regardless of how we or product providers choose names for digital coaches, we can expect to experience them as intelligent relational agents:

• Intelligent because the digital coach learns about us, adapts, asks smart questions and works with our responses;  
• Relational because of a partnership in which the agent personalizes our interactions, demonstrates awareness of our context, appears non-judgemental, and shows that it is working on behalf of our goals; and
• Agent, because it chooses and acts in an interactive conversation.

Although we can’t yet experience this form of digital coaching, we can recognize now, dispassionately, some of the potential advantages of digital coaches over human coaches such as:    

• Bandwidth to deliver on constantly rising consumer expectations of 24/7 access;  
• Potential to generate a prompt just at the moment a leader has a choice between a default thought and behaviour pattern, and an opportunity to shift both of these. Agents may serve up tips like the development recommendations in assessments such as Hogan that we wish clients would use real-time. They will remember the full leadership mnemonic, not just the first three letters;
• Focused on leader accountability for follow up; 
• Giver of advice only under a limited set of if/when conditions. Perhaps a digital coach would honor more seconds of silence and give leaders more time to think than a human coach?  
• No risk of burnout, of making different decisions based on whether it had a good night’s sleep and no need to practise mindfulness to stay present in coaching sessions; 
• Higher accuracy and speed of recall from previous coaching conversations, potentially enabling the digital coach to test and confirm new points of view expressed by its client and build a deeper connection with its client; and  
• Elicit greater self-disclosure and less impression management than human coaches (Lucas, et al., 2014).

The Market for Digital Coaching

Scale - the potential to offer a resource more broadly at low cost - makes the business case for digital coaching to change the shape of today’s coaching market.  Digital coaching may appear in the market as a direct-to-consumer (D2C) offer or via organizations investing in developing their leaders.  For the former market segment, D2C, niches such as tech entrepreneurs, self-employed millennials, and clients of today’s life coaches may provide the initial customer base. For the latter, digital coaching enables organizations to scale coaching to employees in several ways: 

• Today’s “have-nots”, the high numbers of employees traditionally ineligible for external coaching, largely because of cost, will have a new option for their development;  
• Hybrids of digital and human coaching may be offered to today’s “haves”, i.e. to leaders who typically, in today’s market, are eligible for the investment in an external coach; 
• Organizations investing in a coaching culture can reach all employees and customise delivery for large, important groups such as front line managers; and
• Organizations investing in internal coaching by managers may use digital solutions to develop coaching skills, sustain learning transfer, track coaching utilization and evaluate the impact of internal coaching.  

In both market segments – direct-to-consumer and organizational - there is a potential early mover advantage for a few players such as a high-profile coach who brings his or her well-established brand and experience to a provider of digital solutions. Today’s vendors of digital content and learning solutions could extend their product lines by offering digital coaching to their large user bases. These vendors already have access to a rich set of customer analytics to use in product tests and launches.  

The promise of scale and gains in AI – plus the agile methods used to develop new technology products – suggest it is time to prepare for digital coaching. And time to shape it for the better.  

Our Call to Action

It’s a “call-to-thinking” – to temporarily suspend judgment on outcomes for the human coach and to explore the possibilities free from any  protectionism as  the current market incumbents. If we prematurely dismiss digital coaching as superficial “robo-coaching”, we are likely to miss elements of the bigger picture and lose our voice at the design table. And we belong at the design table working alongside the Data Scientists, AI experts, UX experts, and Software Engineers to influence discussions on issues such as: 

• The knowledge base and rules engine at the heart of the AI. For example, we should advocate for capabilities such as processing user input to infer the user’s readiness for a coaching session and energy for post-session follow up;
• The purpose and value of a coach giving feedback on the dynamics of coaching conversations and relationships. Design teams already understand that their creations need to acknowledge the emotional state of their users in order to sustain the interaction. However, they may not yet appreciate that this acknowledgement via feedback on the experience of working with the user is part of the user’s development journey;  
• Adaptation of coaching codes of ethics for the digital design process, e.g., recognizing the user’s agency; 
• Pre-launch tests that include parallel oversight of digital coaching by a human coach; and
• The nuances of known AI issues - data protection, privacy and biases in programming – in the context of coaching.

Excited and uneasy? Seeing complementary possibilities in digital coaching, or only competition? To build on this introduction to how digital coaching may develop, the next post will explore familiar coaching topics such as accreditation and research in the unfamiliar territory of digital coaching. 

Missed Parts 1 and 2 of this blog mini-series? Read Tech in Coaching Ready for an Upgrade? Part 1 and Tech in Coaching: Time for an Upgrade? Part 2

References

Gratch, J.  (3 February 2017). Personal communication.  http://people.ict.usc.edu/~gratch/

Lucas, G.M.; Gratch, J.; King, A.; Morency, L-P. (2014).  It’s only a computer: Virtual humans increase willingness to disclose. Computers in Human Behavior, 37, 94–100.