SimTherapy
A clinical-thinking trainer for psychologists
A student runs a session with an AI client, builds a live graph of the case, and receives feedback from an AI supervisor. The project is already used in work with students.
We check whether a student's thinking can be made visible and measurable while live supervision stays in place.
- Status
- Pilot, used in work with students
- Direction
- Augment
- For
- Psychology and psychotherapy programs, instructors, supervisors
- What is measured
- Conversation style, method techniques, structure of the case map
- Access
- By request
Status
Pilot. The project runs in agreed scenarios with educational partners and is used in work with students.
For Whom
Students of psychology and psychotherapy programs who need controlled practice.
Instructors and supervisors who want to see the path of students' thinking before they read a transcript.
Programs and clinics that need more practice without a matching growth in hours of live supervision.
How it can work
Situation. Clinical thinking needs practice. Training clients are scarce, mistakes with a real client are costly, and supervision arrives late. Thinking stays invisible and cannot be compared between sessions or cohorts.
System. An AI client conducts an adaptive dialogue. The student builds a case graph at the same time: people, events, environment, request, goal, links. An AI supervisor reads the dialogue and the graph together, and after the session the system codes it into comparable indicators.
Human role. The student makes the decisions and answers for them. The instructor sets the learning goal, sees the group's trajectory, and works through the gaps.
What we study
Whether learning transfers beyond the trainer. We look at whether students justify decisions better, notice gaps in their own model of a case, and handle a new situation, and whether the trainer leaves them with confidence the work does not support.
What is measured
Conversation style (MITI). Quality of contact regardless of topic: the ratio of reflections to questions and the share of open questions, with a breakdown of specific utterances.
Method competency checklist. Presence of the techniques of a chosen approach. For CBT these are session structure, agenda, Socratic questions, and cognitive restructuring. Trauma protocol (CPT), Jungian, and integrative approaches have their own lists.
Graph IoU (structure of thinking). The student's case graph is compared with the scenario's reference graph by nodes and links. A pleasant conversation and an accurate clinical model score differently.
After a session the system produces a report within seconds. A language model does the coding, with an automatic fallback to rule-based analysis if it fails. Reports export to Markdown or PDF. The student sees a progress line over sessions and the instructor sees a cohort table. In Discovery mode the student starts from a hidden reference graph and builds the model from a live conversation.
Limitations
SimTherapy works next to live supervision and professional education. Simulation is not support for a real client, and AI feedback is a training reference for the student and the instructor.
It is not a diagnostic tool for real patients and does not assess a student's fitness for the profession.
The indicators describe one session. They are compared across sessions and cohorts only within the same scenario.
Access
Access is by request. Terms of use are agreed before you start working with the project.
HandS Lab research tools and prototypes are not a catalog of commercial products. If a specific applied solution involves commercial deployment, it is carried out by an external operator or partner.