Mursion
A virtual human simulation training platform for high-stakes workplace conversations
Tool details
Mursion is a B2B immersive conversation training platform that grew out of the UCF TeachLivE lab: learners practice performance feedback, layoff conversations, clinician-patient communication, parent conferences and classroom management against virtual humans in the browser. There is no self-serve sign-up; it is sold to enterprise L&D, universities and hospitals on annual contracts.
Core capabilities
- Two conversation tracks: from 2015 to 2024 the dialogue was voiced live by human Specialists, and in 2025 a GenAI on-demand tier was added; AI now carries most sessions while humans cover the high end
- Industry scenario library: five lines — leadership, customer service, education, healthcare and manufacturing — with classroom management and parent conferences as progressive modules
- Skill scoring: ratings from Developing to Strong, tied to observable behaviors and located by timestamp
- AI coaching: AI-driven orientation before practice and reflection afterwards
- Enterprise integration: SCORM 1.2/2004, SAML/OIDC, JIT provisioning and deep linking into the LMS
Where it fits Executive conversation rehearsal for L&D at large enterprises, teacher-preparation courses at universities, and communication and de-escalation training in hospitals. It does not fit individuals or SMBs who want high-frequency per-seat practice for thousands of people, or who want to try it self-serve before buying.
What sets it apart It is willing to draw a line against 'crowd-pleasing AI role-play': its virtual humans do get defensive and frustrated. But the company never discloses the mix of AI versus human sessions, and the homepage still carries the old 'with a live human behind the avatar' copy; the stack visible in job postings is DeepGram + LiveKit + ElevenLabs + OpenAI.
Editorial assessment The assessment depth is solid, and so are the hard edges: every human session consumes a specialist and needs 8 weeks' lead time, while pivoting to AI drops it into the same crowded field as Virti/Zenarate/Attensi; G2 shows only 2 reviews, one of them from an affiliated party; there is no public price and no self-serve trial; there is no public SOC 2, FERPA or HIPAA-BAA, yet it sells to K-12 districts, universities and hospitals at the same time. Evidence of effectiveness has to be layered: the independent NSF-funded research comes from the human-driven TeachLivE era and cannot be extrapolated to the 2026 LLM version.
Pricing
Custom pricingCustom (quote-based)Default term of 1 year, auto-renewing for another 1 year unless written notice is given 30 days in advance; subscription fees appear in the SOW, and the contract specifies neither per-seat nor per-session pricing
A human Simulation Specialist role-plays in real time; first delivery needs 8 weeks' notice; rescheduling within 5% is free, while overages or cancellations within a week are billed per session
LLM-driven on-demand practice, which the site says suits 'most projects'; the ratio to human sessions and the unit price are agreed in the SOW
Custom virtual humans, environments and software are billed on a time-and-materials or deliverable basis, with premium support purchased separately
Additional notes
The official /pricing, /plans and /price pages all return 404, and none of the 32 pages on the site is a pricing page, so procurement has to go demo → SOW → annual contract. Third-party estimates are about $49–$164 per person per session (not officially confirmed); the widely circulated $38–$84/month tiers directly contradict the official statement that there are no public tiers, so they are not adopted. Payment is due 30 days from the invoice date, amounts already paid are non-refundable, the SLA is 99%, and the liability cap is $2M or 10 times fees paid.
FAQ
User reviews
User Reviews
No reviews yet
No one has reviewed this tool yet
