This workflow is for GP ST3s using SimPrep AI's AI patient consultations who want a disciplined record–review–repeat loop rather than open-ended practice. It addresses one job: turning AI consultation practice into calibrated, transferable performance across the RCGP domains. Its principal limitation is twofold — an AI patient and AI-aligned feedback approximate but do not replace a human examiner, and several of SimPrep AI's specifics could not be confirmed at the last check, so verify them before you buy.
What SimPrep AI offers for the MRCGP SCA right now
Be cautious here. At the last check, SimPrep AI's public pages confirmed the format but not the numbers, so much of this needs verifying on the product page rather than being taken as fact.
| Item | What SimPrep AI reports (last checked 19 July 2026) |
|---|---|
| Format | AI-powered medical consultation practice with AI patients, covering both OSCE and MRCGP SCA preparation |
| Feedback | "Instant RCGP-aligned feedback", described as informed by senior GPs (vendor-reported); domain-by-domain scoring detail not confirmed — verify on the product page |
| Case volume | Not publicly confirmed at the last check — verify the current case or scenario count on the product page |
| Price / access | Not publicly confirmed at the last check; a free consultation to start was noted. Verify current pricing and access period on the product page before purchase |
| SCA components addressed | AI voice consultation practice with RCGP-aligned feedback across, in principle, all three domains; confirm exactly which are scored on the product page |
The honest position is that SimPrep AI's modality is clear — AI voice patients for SCA and OSCE with aligned feedback — but its scale, pricing and scoring granularity were not verifiable from its public pages. Do not infer a case count or price from this article; check the source. This piece pairs with the companion SimPrep AI SCA simulator audit rather than repeating it, and the workflow below holds regardless of the exact figures.
The exam you are actually preparing for
The SCA is twelve simulated remote consultations of twelve minutes each — 144 minutes total — sat in ST3, with nine diets a year and a fee of about £1,207. Each is examiner-judged across three domains: Data Gathering and Diagnosis; Clinical Management and Medical Complexity; and Relating to Others.
Cases are blueprinted across twelve Clinical Experience Groups, and the RCGP states plainly that not every group appears in every diet, that one case can span several groups, and that no single case covers all three domains — they balance across the whole exam. The cases are unseen and the score is a human judgement. Because SimPrep AI also serves OSCE-style exams, be deliberate that your SCA practice uses the twelve-minute remote GP consultation format and GP management, not a generic station script; the crossover is useful for structure but the content and pacing differ.
Step 1 — Build a balanced case set across the Clinical Experience Groups
Whatever SimPrep AI's case count turns out to be, drive your selection with a grid rather than picking whatever appears. List the twelve groups down the side and vary four things across the top — acuity, patient age band, complexity or comorbidity, and communication challenge.
| Clinical Experience Group | Vary acuity | Vary age | Vary complexity | Vary communication |
|---|---|---|---|---|
| Under 19 / reproductive & sexual health | Routine to urgent | Child to adult | Safeguarding, confidentiality | Third party present |
| Long-term condition / older adults | Stable to acute | Working-age to frail | Multimorbidity, polypharmacy | Adherence, sensory loss |
| Mental health / urgent care | Low risk to crisis | Any | Physical–mental overlap | Risk, time pressure |
| Health disadvantage / diversity | Any | Any | Language, literacy | Interpreter, beliefs |
| Undifferentiated / prescribing | Any | Any | Uncertainty, interactions | Managing not-knowing |
| Investigation-results / professional dilemma | Any | Any | Incidental findings, ethics | Uncertain news, disclosure |
If SimPrep AI's library is small or OSCE-weighted, this grid will quickly expose which SCA groups it cannot cover — a finding worth acting on, since coverage matters more than case count. The same logic drives the blueprint-coverage matrix framework.
Step 2 — Record the first attempt cold, and preserve unseen cases
Run each AI consultation for the full twelve minutes without pausing, restarting or reading any mark scheme first. That cold run is your baseline. Whatever the library size, ring-fence a preserved block you do not open until the final fortnight — ideally one to two full twelve-case mocks' worth — so you keep genuinely unseen material for a final readiness check. If the library is too small to preserve that much, that is itself a signal to add a second source rather than re-sitting seen cases.
Step 3 — Review by scoring twice, calibrated by observability
Score every consultation twice: first take SimPrep AI's RCGP-aligned feedback, then score the same recording yourself against the three domains, ideally with a peer or trainer doing so independently. Record the disagreements.
Sort each AI comment by how observable the underlying thing is:
- Observable — clearly present or absent in the transcript: red-flag questions, a named-timeframe safety net, an explicit plan, checking understanding. Most reliable; act on these once the clinical content is confirmed current.
- Inferred — rapport, patient-centredness, shared decision-making, deduced from proxies. This is Relating to Others and is where AI feedback is weakest; confirm with a human before acting.
- Generated — model plans, suggested phrasing, any score. Check against current guidance and a human; treat a score as a hypothesis, not a verdict.
With SimPrep AI specifically, because the scoring granularity is unconfirmed, do not assume its feedback maps cleanly onto the three RCGP domains — check that mapping yourself during calibration. Voice AI also adds transcription error and a tendency to reward fluent talking; both need a human check. The method is set out in how to calibrate automated feedback before you trust the score.
Step 4 — Convert feedback into two observable behaviours
Pick two observable behaviours per case and ignore the rest until they hold. Replace "improve management" with "give a follow-up interval and one worsening trigger before closing". Replace "be more patient-centred" with "elicit and reflect the patient's specific concern in the first two minutes". Two tracked behaviours change consultations; a long list does not.
Step 5 — Repeat with deliberate variation
Do not re-run the same case. Take the same principle into a different group with a different agenda, comorbidity or time pressure — shared decision-making from a statin case into contraception, then into deprescribing for a frail patient. The surface changes; the principle stays fixed. That is what makes a rehearsed skill survive an unseen exam case, and it is more valuable than sitting more new cases you never review.
Keeping the clinical management current — where iatroX fits, and where it does not
Any AI simulator can generate a confident but subtly dated management plan. Closing that currency gap is the one job iatroX does in this stack. The boundary is firm: iatroX is a question-bank and clinical-knowledge platform, not a consultation simulator. It does not voice a patient, grade rapport or replace SimPrep AI's consultations. It measures whether the knowledge under your consultations is current — unseen, SCA-style clinical MCQs and citation-first clinical answers grounded in NICE, CKS, SIGN, the SmPC/eMC and NHS content. Consult in SimPrep AI; verify the medicine in iatroX. And read any score as a measurement, not a prediction: your Q-bank percentage is not your exam score.
A seven-day plan for a working ST3
SimPrep AI does one job — voice and review the consultation; iatroX does another — measure unseen knowledge. No proprietary predictive-algorithm claim is made for either.
| Day | SimPrep AI (record + review) | iatroX (knowledge job) |
|---|---|---|
| Mon | Pick two under-covered groups from the grid; run one AI consultation | 15 unseen MCQs in those topics |
| Tue | Cold-run a second consultation; do not read feedback first | Check any management you doubted |
| Wed | Score both twice; log observable / inferred / generated disagreements; check the domain mapping | — |
| Thu | Re-run one shaky principle in a new group | 15 mixed unseen MCQs |
| Fri | Read feedback; pick two observable behaviours | Recheck one wrong guideline |
| Sat | Sit one preserved, unseen case end-to-end | 20-item timed block |
| Sun | Review the AI-vs-human log; refill grid holes | Log recurring knowledge gaps |
Three mistakes this workflow is designed to stop
Assuming OSCE crossover is SCA practice. SimPrep AI serves both; a generic station is not a twelve-minute GP consultation, so keep the format honest.
Trusting unconfirmed scoring. If the domain mapping is unverified, an aligned "score" may not mean what you think; calibrate it against a human before you rely on it.
Practising without preserving unseen cases. If the library is small and you burn it all, your final mock is a memory test, not a readiness check.
Exit standard and the continue / supplement / switch / stop decision
Ease off when three signals hold across unseen cases: consistent rather than one-off performance; agreement between SimPrep AI's feedback, your own scoring and a human reviewer, especially on the inferred tier; and no recurrent safety-critical omission.
Decide on gaps, not novelty. Continue if coverage is filling and unseen performance is rising. Supplement — and this is more likely here than with better-documented tools — if the library is too thin to preserve unseen cases, or if Relating to Others stays weak and you need live human role-play. Switch the primary tool if, after an honest audit, it does not cover the SCA groups you keep failing, or if you cannot verify its scoring. Stop adding platforms once you have one consultation source, one human reviewer and one knowledge check. The comparison hub sets the options side by side if you are still choosing.
Frequently asked questions
Is SimPrep AI enough for MRCGP SCA on its own? Probably not on its own, and partly because its scale and scoring were not fully verifiable at the last check. The AI voice-patient practice is useful for building consultation structure, but you still need human calibration, a confirmed way to cover all twelve groups, and a separate current-knowledge check. Verify its specifics before you rely on it as a primary tool.
Which MRCGP SCA component does SimPrep AI not reproduce well? Relating to Others is hardest for any AI to grade, and with SimPrep AI the risk is compounded by unconfirmed domain-level scoring — an aligned rating may not map cleanly onto the interpersonal domain. Calibrate that domain with a human rather than trusting the AI feedback alone.
How many unseen SimPrep AI cases or stations should I preserve for final MRCGP SCA calibration? Aim to ring-fence one to two full twelve-case mocks — around twelve to twenty-four cases — that you never open until the final fortnight. If SimPrep AI's SCA library is too small to spare that many, treat that as a prompt to add a second source rather than re-using seen cases.
When should I stop using SimPrep AI and move to mixed mocks? Once single-case behaviours are stable and, importantly, once you have satisfied yourself that its feedback broadly agrees with a human reviewer, move to full, mixed, timed mocks in the final two to three weeks to test stamina and switching between unrelated cases.
How should I combine SimPrep AI with iatroX without duplicating practice? Keep the jobs distinct: SimPrep AI for the AI consultation, iatroX for unseen clinical-knowledge measurement and current-guidance checks, never re-answering seen items. This follows the two-Q-bank rule: a second resource should add unseen material, not duplicate it.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; SimPrep AI's format was confirmed but its case count, pricing and domain-level scoring were not publicly verifiable at that date — treat all such details as vendor-reported and confirm them on the product page before purchase. Disclosure: iatroX operates a competing question bank and clinical-knowledge platform; its role here is confined to unseen MCQ measurement and current-guidance checks, jobs SimPrep AI's simulator does not claim to do. Corrections are welcome via the feedback route on iatrox.com.
References: RCGP, Simulated Consultation Assessment — overview, preparing, and case content (rcgp.org.uk/mrcgp-exams/simulated-consultation-assessment); SimPrep AI (simprep.ai); iatroX MRCGP SCA bank (iatrox.com/mrcgp-sca); "Your Q-Bank Percentage Is Not Your Exam Score" (iatrox.com/blog/qbank-percentage-not-your-exam-score).