The most useful AI OSCE feedback identifies something you can change in the next encounter. A fluent virtual patient and a polished score are not enough. For UKMLA CPSA preparation, compare the link between the conversation, the feedback and a specific improvement in performance.
iatroX publishes this article and includes its simulations alongside Geeky Medics and Quesmed. Product descriptions were checked on 19 September 2026. No standardised head-to-head station results or pass predictions are claimed.
Compare the whole encounter
A clinical-skills platform may provide station instructions, a virtual patient, examiner-style feedback, teaching resources and progress records. Inspect each separately. A large station library is not automatically the same thing as a large collection of interactive AI cases.
Geeky Medics combines clinical-skills material with virtual-patient and feedback features. Its AI use can be governed by credits, so compare the allowance and how it is consumed. Do not translate credits into a number of complete stations unless the provider documents that conversion.
Quesmed's AI OSCE page advertised unlimited text and voice practice for £39.99 annually when checked. That specialist offer costs less than iatroX annual all-access. Scope, feedback design and the wider learning functions still differ.
A useful debrief needs evidence
A comment such as "explore concerns more fully" can be reasonable but difficult to act on. Better feedback identifies the point at which a concern was mentioned, what the learner did and what could have been explored next.
The same standard should apply to positive feedback. "Good communication" is less informative than an observation that the learner checked understanding after a complex explanation. Specific evidence helps distinguish a justified judgement from a generic response.
According to its September 2026 launch information, iatroX provides transcript-linked feedback by domain and Tutor-led follow-up. These are design features to inspect. They are not proof that every judgement is correct or equivalent to an examining body's assessment.
An original sample-review method
Use a fictional consultation scenario and choose one intended skill, such as exploring the patient's main concern. After the encounter, locate the relevant section of the transcript and compare it with the feedback.
Ask whether the comment matches what happened, whether an important omission was missed and whether the proposed improvement is proportionate. Then repeat or practise an appropriate related case with one change in mind.
This is a proposed method, not a test already conducted across the named providers. Publishing comparative results would require the actual encounters, consistent criteria and suitably qualified review. We do not invent a league table from product descriptions.
Unlimited practice is useful only when it includes review
A generous allowance can reduce the pressure to ration attempts. It does not automatically create better learning. Rapidly completing many encounters without reviewing the feedback may reinforce the same habits.
Credit-based access can still be good value when the learner needs only selected practice, especially if the accompanying teaching resources are useful. Conversely, a candidate planning frequent interactive sessions should understand the allowance before buying.
Check whether the product supports the examination track you need. General consultation rehearsal and dedicated CPSA preparation are related but not identical claims.
Keep practical limits visible
Conversational simulation can support history-taking, explanation and clinical reasoning. It does not replace examining real patients, practising procedures or receiving appropriate supervised feedback. A virtual description of a physical sign is not the experience of eliciting it.
Use simulation alongside teaching and practice that address those tasks. The strongest learning plan assigns each resource a role rather than expecting one subscription to replace all clinical education.
The verdict
Choose Quesmed when its lower-priced specialist AI OSCE offer meets the need. Choose Geeky Medics when its clinical-skills resources and AI allowance fit your preparation. Consider iatroX when transcript-based review, Tutor follow-up and related written learning are useful together. Judge the package by the next improvement it helps you make, not by which patient sounds most convincing.
A practical way to evaluate iatroX
The September 2026 simulation release includes UKMLA CPSA, with voice and text interaction, coached and uninterrupted modes, and transcript-linked feedback. Tutor-led follow-up is part of the published design, not an independently validated prediction of examination performance.
The introductory offer is one complete simulation across the platform, not a separate free attempt for every examination. Inspect the debrief as well as the conversation; practical skills still require suitable supervised work.
At the UK prices published on 19 September 2026, annual access costs £99 upfront, equivalent to £8.25 per month billed annually; monthly access costs £29. The paid package includes questions, Tutor, planning, simulations and CPD tools. Value depends on using several relevant learning methods, not on collecting access to unrelated exams.
Frequently asked questions
Does unlimited AI practice guarantee better preparation?
No. Its value depends on relevant cases, useful feedback and time spent applying the feedback.
Can an AI feedback score predict my CPSA result?
No pass prediction is established here. Treat commercial feedback as a learning aid rather than an official assessment outcome.
Is the free iatroX simulation available once for every track?
No. The introductory offer is one complete simulation across the platform, not one per examination.