Do AI-Simulated Sessions Count Toward CACREP Direct Hours?
AI client simulations have arrived in counselor education faster than almost anyone expected. Platforms now let students conduct full mock sessions with lifelike AI clients — voice, video avatars, multi-session continuity, instant feedback — and universities are signing pilot agreements to embed them across the curriculum. Which raises the question every program director, clinical coordinator, and student is now asking:
If a student spends an hour counseling an AI-simulated client, can that hour be logged as direct service under CACREP?
The short answer is no — not under the 2024 CACREP Standards as written. The longer answer is more interesting, because AI simulation still has a legitimate (and growing) place in a CACREP program. It just isn't in the direct-hours column. Here's the full picture, with the standards language, the emerging research, and what your program should do about it.
What the 2024 CACREP Standards actually require
Section 4 of the 2024 CACREP Standards (Professional Practice) governs practicum and internship. Two standards do the heavy lifting on hours:
- Standard 4.R: Practicum students complete at least 40 hours of direct service with actual clients that contributes to the development of counseling skills, within a practicum totaling a minimum of 100 hours (4.Q).
- Standard 4.V: Internship students complete a minimum of 240 hours of direct service with actual clients, within a 600-hour internship completed "in roles and settings with actual clients" (4.U).
The operative phrase appears four times across these standards: actual clients. That's not incidental wording. CACREP has long defined direct service as the supervised use of counseling, consultation, or related professional skills with actual clients — individuals, couples, families, or groups — for the purpose of fostering social, cognitive, behavioral, and/or affective change. Activities like observing others provide counseling, record keeping, and administrative tasks are explicitly excluded from direct service.
An AI simulation, no matter how lifelike, is not an actual client. There is no real person experiencing real change, no informed consent, no therapeutic risk, no genuine clinical relationship. By the plain language of the standards, an hour with an AI client cannot be logged as a direct hour — in practicum or internship.
You'll find the relevant standards on CACREP's Section 4: Professional Practice page, and the principle holds across both the 2016 and 2024 standards: the "actual clients" language predates AI and was originally intended to keep classroom role-plays with peers out of the direct column. AI simulation inherits the same status — it's a role-play with a very sophisticated partner.
So where can simulation hours live?
This is where programs need to be careful, because there are really three buckets, and they're often conflated:
1. Direct hours. Off the table for AI sessions, as above. Only interaction with actual clients counts toward the 40-hour practicum and 240-hour internship direct minimums.
2. Indirect / total field hours. This is grayer. The non-direct portion of practicum and internship (the other 60 of 100, and 360 of 600) covers activities that support clinical work — documentation, supervision, case review, site training. Whether an AI simulation completed as part of the fieldwork experience could be logged as indirect time is not explicitly addressed in the standards, and interpretations vary. Some programs treat structured simulation as a defensible professional-development activity within indirect hours; others keep it out of field logs entirely. If your program intends to count simulation time anywhere in the practicum or internship total, get it in writing from your CACREP liaison first, and document the rationale in your fieldwork handbook. Do not let students discover the answer at audit time.
3. Pre-practicum and curricular use. This is where AI simulation unambiguously belongs today — skills courses, techniques labs, pre-practicum readiness checks, and remediation. No accreditation question arises because no field hours are being claimed.
The research: simulation as preparation, not substitution
The peer-reviewed literature is consistent on this framing. In a pilot study published in the Journal of Counselor Preparation and Supervision, Beeson, Zhai, Fulmer, Burck, and Maurya (2025) ran master's students through 53-minute mock intake sessions using ChatGPT-4o's voice mode and rated the AI's performance across eight domains, including authenticity, consistency, emotional expression, and empathy. The AI client showed good-to-excellent fidelity compared to a typical client — but the authors also flagged formulaic, idealized responses that raised authenticity concerns, and positioned the technology as a "scalable, cost-effective tool for pre-practicum training."
A related qualitative study by Akkurt, Maurya, and Brown (2025) found that trainees experienced AI-simulated sessions as a psychologically safe space that reduced performance anxiety and built confidence before real client work — while noting the AI's tendency toward overly agreeable responses and limited emotional nuance. Trainees themselves described AI as a supplement to live practice, not a replacement.
The Journal of Counselor Preparation and Supervision has devoted substantial recent attention to AI in counselor education — including work on teaching, supervision, and clinical training strategies — and the through-line is the same: simulation accelerates skill development; it does not constitute clinical experience.
That distinction matters because counseling is choosing a different path than some neighboring professions. Nursing education, for example, has long permitted simulation to substitute for a portion of clinical hours in many states, backed by large-scale simulation research. Counselor education, at least under the current standards, holds the line: skills can be rehearsed in simulation, but direct service happens only with real people.
Real programs, real companies: who's building this
This isn't hypothetical. In January 2026, Marshall University's Department of Counseling announced a formal pilot with SimCare AI (Soma Lab Inc.) to co-design AI-driven clinical simulations and field experience tooling. Department chair Dr. Eric Beeson — lead author of the fidelity study above — framed the value precisely: the technology solves the problem of getting "high fidelity training repetitions for students prior to field experience," supporting rather than substituting for the relational work of counseling.
A few of the platforms counseling programs are evaluating:
- SimCare AI — Y Combinator-backed platform offering real-time voice conversations with AI client avatars, including multi-session continuity, group/couples/family scenarios, and automated skill feedback exported to instructors.
- Skillsetter (formerly Theravue) — a deliberate-practice system co-founded by psychotherapy researcher Dr. Bruce Wampold, built around stimulus-response skill drills, rubrics, and instructor evaluation.
- Mursion — avatar-based simulations blending AI with trained human "interactors," widely used in education and human-services training.
- Simucase — simulation-based learning with a virtual patient/client video library for observation, assessment, and intervention practice across health and behavioral disciplines.
Each is a legitimate training tool. None of them — and notably, none of their serious academic partners — claims that simulated sessions satisfy CACREP direct-hour requirements.
What this means for your program: five practical moves
1. Put it in your fieldwork handbook now. Students are already using AI tools informally. A one-paragraph policy — "AI-simulated sessions do not count toward direct or indirect practicum/internship hours unless explicitly approved in writing" — prevents the painful conversation where a student discovers 30 logged hours don't count. Standard 4.G requires a fieldwork handbook detailing requirements and expectations for all delivery types; this belongs in it.
2. Keep simulation hours categorically separate in your logs. The fastest way to create an accreditation problem is a logging system where simulated and actual client hours blur together. Your hour-tracking workflow should make the direct/indirect/simulation distinction structurally impossible to get wrong — which is exactly the problem purpose-built practicum and internship tracking solves, separating direct from indirect service the way accreditors define it.
3. Use simulation where it shines: readiness and remediation. AI clients are a genuinely excellent answer to "is this student ready for practicum?" Programs can require a benchmark simulated session before field placement, use simulation data in competency and disposition evaluations, and prescribe simulation reps as part of a remediation plan — all without touching the hours question.
4. Bring simulation artifacts into supervision, not into the log. A recorded AI session can be reviewed in individual or group supervision exactly like a recorded real session — often with less confidentiality friction. The supervision hour counts (it's supervision); the simulated session itself doesn't. Clean supervision tracking keeps that boundary visible to students and supervisors alike.
5. Document your policy as accreditation evidence. When the site visit comes, a written AI-simulation policy, clean hour categories, and evaluation rubrics that show how simulation supports skill development are assets, not liabilities. They demonstrate exactly the kind of intentional program design the 2024 standards reward. If your evidence currently lives across spreadsheets and email threads, see how CACREP-aligned compliance tracking and automated CACREP reporting pull it into one defensible record.
The bottom line
AI-simulated sessions do not count toward CACREP direct hours — the 2024 standards require direct service with actual clients, full stop. Treating simulation time as indirect field hours is a gray area you should resolve with your CACREP liaison in writing before counting a single minute.
But "doesn't count toward direct hours" is not the same as "doesn't count." The research and the early institutional pilots point to the same conclusion: AI simulation is becoming the bridge between the classroom and the first real client — a way to give every student dozens of safe repetitions before the repetitions start to matter. Programs that adopt it with clear policies, clean hour categories, and honest logging will get the pedagogical upside without the accreditation risk.
And if keeping direct, indirect, supervision, and simulation hours cleanly separated across an entire cohort sounds like exactly the kind of thing that shouldn't live in a spreadsheet — that's what Pracadium was built for.
Sources & further reading
- 2024 CACREP Standards (combined version, PDF)
- CACREP Section 4: Professional Practice
- Marshall University news: Marshall to pilot artificial intelligence software in counseling training (Jan 2026)
- Beeson, Zhai, Fulmer, Burck, & Maurya (2025), A pilot study evaluating the fidelity of ChatGPT in client simulations, JCPS 19(3)
- Akkurt, Maurya, & Brown (2025), Learning Through Simulation: Counselor Trainees' Interactions with ChatGPT as a Client
- Journal of Counselor Preparation and Supervision — AI in counselor education
- SimCare AI · Skillsetter · Mursion · Simucase