How can we empower humans in healthcare? Could AI be the green shoots of a revolution in the profession? Sam Isaacson reports
At first glance it was a routine meeting, one of the countless video calls that seem to fill my time as a work-from-home coach / consultant / general dogsbody. But as I clicked the ‘join’ button that crisp morning in late February, I wasn’t fully aware just how exciting it would
prove to be.
Gathered on the screen before me were the faces of a small number of critically important people. There was Darren Leech, director and head of coaching at NHS Elect, a network dedicated to providing a range of services, including coaching to people across the NHS. There was Dr David Tee, coach, author, researcher, chair of the Wales Coaching Centre at the University of South Wales and editor of the British Psychological Society publication, The Coaching Psychologist. And there was Boas Loeb, co-founder with me of AIcoach.chat, a tool powered by artificial intelligence (AI), designed to provide the experience of coaching conversations to the 99% of the population who’ve never had access to coaching before.
Their expressions were a mixture of excitement and trepidation as Dr Tee prepared to unveil the results of our pioneering AI coaching pilot.
Months earlier we’d embarked on an ambitious collaboration, one that pushed the boundaries of what AI was capable of achieving. Could an AI truly understand the nuances of the human experience? Could it forge authentic connections and catalyse personal growth in the same way as flesh-and-blood coaches have been able to?
Ethical responsibility
A lot has been said over the past few years about the relationship between ethics and AI in coaching. There are concerns about a loss of confidentiality, about bias existing in datasets, about AI ‘hallucinating’, and more. Ethics is a critically important issue. In my book, How to Thrive as a Coach in a Digital World (McGraw Hill, 2021), I explained that while some ethical boundaries should never be crossed, others are more subjective and we should be gracious with others when we disagree with them.
In this case, I don’t think we’re talking about crossing an ethical boundary. We have a different sort of responsibility. If we can test AI and demonstrate that it enables identical outcomes to human-delivered coaching but at a lower price, knowingly withholding the technology would simply be wrong. Here’s how I put it in the book: “Where technology offers opportunities to better achieve positive outcomes from coaching, with an acceptable and transparently understood level of risk, and at a price that justifies it, it would be ethically wrong to not embrace it fully.”
With that in mind, perhaps the NHS is the ultimate testbed for this emergent technology. Those working within the NHS network are the individuals tasked with protecting the wellbeing of a nation. Doctors, nurses, administrators and many more, all united by a common purpose, committed to caring for others, and yet far too often driven to de-prioritise the things that coaching is so good at elevating: Strategic thinking, systemic awareness, personal wellbeing, and so on. If AI-powered coaching could genuinely help to solve that problem, within a budget that’s appropriate for the NHS, surely it ought to enthuse us that wider adoption of AI coaching is an exciting next step for the profession.
Might we therefore allow ourselves to dream for a moment, of a future where the transformative power of coaching – the catalyst for self-awareness, growth and systemic change that we know it can be – could be truly democratised and made accessible to all, regardless of circumstance.
A bold experiment
Our first conversation about piloting AIcoach.chat within NHS Elect was like setting out into unfamiliar territory in deep fog. Here was a chance to put our version of AI coaching to the ultimate test, to see if it could truly live up to our vision of it empowering individuals on a mass scale. Other so-called AI coaches have started to appear around the world – some have already disappeared – and too many clearly don’t have the coaching profession at their heart, so while we believed that what we were doing was ethically right, there was undeniable concern that we might have got it wrong.
For one month each, 53 volunteers from two NHS Trusts would become our pioneers, moving into the fog one step at a time. In some ways they were a challenging group, comprised of existing coaches and senior leaders; the tool was going to have to work well if it were going to prove its worth. I was at arm’s length from what was happening in those conversations. While a survey had been sent out prior to the participants being granted access, to gauge their level of goal attainment and self-efficacy I hadn’t seen the results of this, didn’t know who was taking part, and wasn’t able to witness AIcoach.chat sparking epiphanies in the AI coaching client. The only anecdotal feedback that trickled in was in the form of instinctive reactions to using the tool, my favourite of which simply started with, “Blimey that was very good!”
And so we found ourselves on that call, in which Dr Tee had crunched the numbers, to allow the facts finally to speak for themselves. No more speculation, no more hope, no more fear of what might be, just the data, as objective as one can realistically get in the world of coaching. It’s fair to say the outcomes shone brightly over the dull glow of my laptop screen:
A 10% average increase in goal attainment for most users, after just a single AIcoach.chat session. An average 5% uplift in self-efficacy, further empowering a group with a high baseline with the self-belief to conquer challenges; such a valuable attribute to have in any organisation. These results didn’t change based on the client’s gender – and improved proportionately to the number of sessions they took.
As I stared curiously at the outcomes, I couldn’t help asking myself what difference I would have made as a human coach had I delivered coaching sessions to 53 people within one month. Would I realistically predict that my clients would see an average 10% increase in goal attainment and
5% increase in self-efficacy? And yet, with no shadow of a doubt, these had been catalysed by a machine – one built on deep learning algorithms prompted by the best in coaching ethics – but a machine nonetheless.
In that moment, I knew we’d uncovered something of value. I’ve heard Peter Hawkins quote a young man who described coaching as “expensive personal development for the already highly privileged”. The tool we were looking at tears the curtain restricting that access, making the expensive affordable and even free, turning personal development into systemic positive change, and opening the profession of coaching up to those who would not consider themselves highly privileged. The scarcity of accredited human coaches is now no excuse to limit the reach and impact of this invaluable practice.
The road ahead
I’m not sure I need to say this: I’m excited about what this could mean. If AI coaching, in tools such as AIcoach.chat, becomes an indispensable part of our day-to-day experience in the same way that social media has, what could that mean for the future, for individuals, organisations, the wider systems?
Perhaps leaders across industries, emboldened on demand to shatter self-imposed limitations, will elevate their organisations to new heights of innovation and impact. Maybe individuals struggling with the challenges of the day-to-day will find their spirits buoyed by AI-powered coaching support and challenge, approaching every step they take with confidence and a mind like water.
At the very least, it will mean good news for the individuals who make use of it and the people they interact with. I’d hesitate before guaranteeing self-actualisation but it certainly feels like a step in the right direction.
It won’t be easy for AI coaching, I’m sure. The pioneering pilot with NHS Elect was merely the first step on this journey, a proof of concept that’s proven its worth as much as it can.
It might not be easy for the coaching profession as a whole. I have some coaching clients where the organisation’s budgets are tight, and can see the attraction a cost-conscious buyer will feel. Moving to an AI coach will give comfort that people are receiving coaching, even if humans aren’t involved. If there’s a significant shift in this direction, there’s plenty of thought the professional bodies will have to put into rapidly reworking documentation that has been evolving slowly over the past few decades.
But to those who feel the same tug that I do, that we have an ethical responsibility to embrace technology if it delivers outcomes that benefit our clients, we must embrace the future, alive as it is with possibility, without letting go of the foundations we’ve always worked from.
Maybe the dawn of AI-powered coaching is upon us. If it is, plenty will try to use it to make a quick buck, others will be well-meaning and trip themselves and others up in the process. I’m sure there will be plenty written about the Wild West of AI Coaching. But if we work together I have no doubt the coaching profession as a whole will respond in the same agile way we did when in-person coaching became remote coaching in 2020. This is simply a different-sounding call to a different-looking next chapter in our shared adventure.