AI Therapy Role-Play: How to Practise With a Simulated Client

Traditional role-play asks a trainee to act as the therapist while a peer or actor portrays a client. It remains a valuable training method, but organising practice can take time and often depends on having someone available to provide feedback. AI therapy role-play offers an additional format: a fictional client who can take part in short, repeatable training conversations.

What is a simulated client?

It is a constructed character with a defined educational profile: a concern, a communication style, likely emotional responses, and a learning objective. It is not a real patient, nor a dependable model of every clinical reaction.

A 2024 proof-of-concept study by Fung and Laing showed that a language model could maintain the role of a fictional client in a text-based CBT training exercise. That demonstrates a possible use for training; it does not validate the system as a replacement for supervised experience.

Example: a reluctant first-time client

Scenario: A person attends an initial session after encouragement from their partner. They are unsure why talking should help.

Simulated client: “I don't really know what to say. I'm only here because other people insisted.”

Possible response A: “You'll see that therapy works if you give it a chance.”

Possible response B: “It sounds as if coming here wasn't really your choice. What's it like for you to be in this conversation?”

Response B acknowledges reluctance and opens exploration instead of offering reassurance too early. But there is no universally perfect line: the best response always depends on context, the relationship, and what happens next.

A simple practice structure

Before the role-play: choose one skill, such as avoiding premature reassurance. Read the fictional client's profile and decide what you want to practise.

During the conversation: keep the exercise focused. Notice whether your response explores the client's experience or rushes toward a solution. You do not need to demonstrate every technique in one encounter.

Afterwards: revisit two or three moments. What did you assume? Which feeling did you recognise? How might you phrase one response differently? Compare alternatives before starting another attempt.

Why scenario design matters

A client who agrees with everything can make practice unrealistically easy. A character displaying many challenges at once may overload a beginner. A useful training scenario balances one identifiable challenge, an appropriate level of difficulty, and a clear target skill.

A 2025 qualitative study of counsellor trainees found that detailed prompts helped improve perceived realism. Learners also noticed that AI clients could be too agreeable and limited in emotional or cultural nuance.

Where role-play fits

AI role-play may make focused rehearsal easier to access. It cannot establish that someone is ready for complex clinical work or replace supervision, live interaction, and professional judgement.

TheraMentor offers a setting for simulated skills practice and subsequent reflection. For a useful next exercise, try focusing on reflection of emotion rather than on asking more questions.