Recently while I was mentoring another clinician I came across an interesting conundrum.

This clinician is highly experienced, with a sharp mind for clinical reasoning and a deep understanding of physical rehabilitation. Yet, as our session unfolded, he shared a frustration that many in healthcare will recognise; despite clear, logical treatment plans, some clients just weren’t following the professional advice given to them.

The Roadblocks

The clinician described how clients would agree to their rehab programme but then fail to make progress. It wasn’t their bodies that were holding them back, but something less visible – emotional barriers. Who would have guessed that clients’ thoughts, attitudes and beliefs could actually interfere with them getting better?

“If only I could work with someone who had no emotions,” he sighed, “like Spock from Star Trek. Just pure logic, no drama!”

The AI Parallel

A few weeks later, in another 1 on 1 professional development session the conversation took an unexpected turn. The clinician began using AI to support his clinical work and soon voiced a new frustration: “AI just doesn’t get it. It never gives me the answers I’m looking for.” There was a pause, as we both realised the irony – AI, with its logical, emotionless responses, was exactly the “Spock-like” client he had wished for. Yet, the interaction was still unsatisfying.

The Turning Point

This realisation was a lightbulb moment. The common thread in both scenarios was the clinician, not the emotionally blocked clients or the unhelpful AI. The ensuing conversation was a rich vein of self-appraisal and curiosity about what to do different and how to do that.

The magic ingredient in this case was curiosity. Curiosity around how to interact differently with an emotionally blocked client and about how we can better prompt AI to achieve what we really need from it.

Being a highly skilled clinician it can be tough to step back from tried-and-true strategies to do something different. Curiosity is a great elixir when we feel challenged.

Actions speak louder than words – especially when your action uses good words

One of the key actions was to channel Michael Bungay Stanier’s Advice Monster book. Instead of instantly purging advice immediately onto the client we utilised powerful questions that invited the client into the process. Three simple questions that really worked for this experienced clinician were:

“What’s on your mind about doing these exercises?”
“What else is going on that might slow your progress?”
“What’s the real challenge here for you?”

There was more utilised than just these three questions, but the point is that the important parts of the clinical interaction were more about context than content!

And what did we do with AI? Here’s a really simple hack to use this AI tool better than most. Utilise the acronym “CRISP” when prompt engineering. The letters stand for Context, Role, Intent, Style and Precision.

C – Context
Set the background: patient details, scenario, or challenge.
(e.g., “Male, 25, post-lumbar surgery, persistent lower back pain, footballer”)
R – Role of AI
Tell the AI who to be, or what hat to wear.
(e.g., “Act as a senior musculoskeletal physiotherapist”)
I – Intent/Instruction
What do you want? Research? A plan? Critical feedback?
(e.g., “Generate a staged rehab protocol with psychological return-to-play markers”)
S – Style/Tone
How should the answer sound? Clinical, empathetic, plain English?
(e.g., “Write in clear, jargon-free Australian English”)
P – Precision
Be specific about output: format, length, focus, or exclusions.
(e.g., “Bullet points only; focus on both physical and mental readiness; max 150 words”)

A CRISP prompt example could be this:
“C: My client is a 25-year-old male footballer, 8 weeks post-lumbar surgery, reporting fear of return to sport and ongoing low back pain. R: Act as a senior Australian sports physiotherapist. I: Create an evidence-based staged rehab plan, including both physical and psychological progression. S: Use plain, supportive language suitable for sharing with the client. P: Keep it in five bullet points, max 150 words.”

The Professional Learning Point

The results were transformative. Clients became more engaged and invested in their own progress, voicing their own solutions and taking greater ownership of their rehabilitation. The AI, in turn, started providing more useful and nuanced responses.

As neuroscientist Fabrizio Benedetti aptly put it, “The interaction creates the intervention!” The heart of communication, whether with patients or AI, lies in the questions we ask and the space we create for meaningful dialogue. In the end, it’s not about eliminating emotion or relying on logic alone, but about mastering the art of the right question at the right time.