Useful feedback starts with your words.

Why pair practice with feedback?

A 2019 study of ClientBot randomly assigned 151 nontherapists to practice with immediate feedback or receive initial instruction and encouragement. The feedback group used more reflections, including during a short test after feedback was removed. The benefit for open questions did not persist in that test. It’s a reason to explore this approach, with care about what the results can tell us.

Tanana et al. (2019), Development and Evaluation of ClientBot.

A randomized study of CARE, published at CHI in 2026, points the same way from the other side: novice counselors who practiced with simulated clients and got feedback improved their reflections and questions, and the group that practiced without feedback didn’t, and showed less empathy afterward. That’s why every SeatTime rep ends with feedback, and why our simulated clients are written to respond differently when you get it right than when you don’t. The study evaluated CARE; it does not establish SeatTime’s effectiveness.

Louie et al. (2026), CARE randomized study, CHI ’26, Article 15, pp. 1–31 (published April 13, 2026).

How practice is structured.

Deliberate-practice training in psychotherapy drills a small set of skills with short, repeatable exercises, graded from easier to harder, with the trainee adjusting difficulty as they go. SeatTime follows that shape: each clip trains one objective with one fictional client, every clip has beginner, intermediate, and advanced versions of the client, and your own “too easy / too hard” read steers what comes next. Exploration skills come before insight skills, and the library adds tracks for alliance ruptures, broaching culture and identity, couples and families, and risk.

APA Essentials of Deliberate Practice series: sample chapter · Muran, Eubanks & Wallner Samstag (2021), alliance rupture and repair.

What our feedback looks for.

Three layers. First, behavior counts: each line you type is coded (reflection, open or closed question, advice, and so on), in categories adapted from helping-skills research and MI behavior coding. Second, a checklist of what that clip asks for, with the turn that shows each one. Third, a read of how you came across (empathy, warmth, emotional engagement, collaboration, and repair when things strained) on behavioral anchors adapted from the Facilitative Interpersonal Skills task, which was designed to be rated from short written responses. Then three moments from your transcript and one line to try instead.

SeatTime’s automated feedback is not a formal MITI, CCS-R, FIS, or Counseling Skills Scale assessment, and it does not produce a clinical competency score. The CCS-R in particular is a supervisor’s rating of a whole practicum; we map skills to it so you know what your supervisor will look for, not to score a rep.

Moyers, Manuel & Ernst, MITI Coding Manual 4.2.1 · Zech et al. (2022), a text-based Facilitative Interpersonal Skills task.

What we have yet to establish.

We haven’t published a validation study of SeatTime or agreement rates with counselor-educator raters. We’re building that check: expert-coded transcripts compared with the coach, code by code, and any count that doesn’t meet a substantial-agreement bar is marked provisional in the app. Research on other tools informs our approach; it doesn’t establish SeatTime’s effectiveness with real clients.

AI feedback can miss context and get things wrong. Use it as a starting point for another try, and bring questions to your instructor or supervisor. You can flag feedback that feels off from the review screen.

Practice alongside your training.

Every SeatTime client is fictional. Please keep real client information out of the tool. Practice here doesn’t replace therapy, supervision, coursework, or required clinical hours.