The Royal Pharmaceutical Society's Prescribing Competency Framework, adopted by the NMC as its prescribing standard, needs no eleventh competency for the AI era, and that is the insight worth building on: AI changes how prescribers demonstrate and maintain the existing ten, which means every prescriber already holds the map, and the timing is pointed, with the NMC planning to consult between September and December 2026 on a revised Code and revalidation including stronger expectations around safe and effective use of AI and digital technologies. This article maps AI to each competency with the same three-part discipline throughout: the appropriate use, the unsafe shortcut, and the portfolio evidence that demonstrates the competency rather than asserts it.
The consultation competencies
Assess the patient: appropriately, AI structures information and surfaces the questions not yet asked; the shortcut is letting it fabricate coherence from gaps, and the evidence is a documented case where structured prompting changed what you examined or asked. Identify evidence-based treatment options: appropriately, retrieval and comparison of validated sources with links opened, the routing matrix in action; the shortcut is synthesis-as-terminal-source; the evidence is a worked question with sources compared and the conflict, where one existed, named. Present options and reach a shared decision: appropriately, AI-drafted plain-language explanations reviewed against the PIL and personalised, preserving uncertainty; the shortcut is fluent false reassurance handed over unread; the evidence is a before-and-after of an explanation you corrected. Prescribe: appropriately, verification of product, formulation, licensing status and legal requirements against the exact SmPC; the shortcut is prescribing from a generic summary; the evidence is a documented product-identity check on a decision where formulation mattered. Provide information: appropriately, accessible, accurate patient information generated, reviewed and personalised; the shortcut is unreviewed handouts; the evidence is the reviewed artefact with your amendments visible.
The governance competencies
Monitor and review: appropriately, seven-section monitoring plans assembled with sources and verified, per this cluster's framework; the shortcut is plans that stop at drug and dose; the evidence is one full plan with its SmPC and guideline links. Prescribe safely: appropriately, AI as a structured second look for interactions and contraindications, always verified against established resources, never as the sole screen, because omissions are invisible; the shortcut is treating silence as clearance; the evidence is a documented near-miss or check that changed a decision. Prescribe professionally: appropriately, scope, accountability and confidentiality carried into every tool interaction, nothing patient-identifiable in consumer systems; the shortcut is authority borrowed from an answer, acting outside scope because the AI made it look routine; the evidence is a reflective note on a case where you declined to act on an answer and escalated instead. Improve prescribing practice: appropriately, auditing your own AI-assisted decisions, logging where answers were wrong or incomplete, feeding corrections back; the shortcut is unexamined habit; the evidence is the audit itself, the most underused portfolio artefact in this list. Prescribe as part of a team: appropriately, sharing sources rather than conclusions, seeking specialist and pharmacist input where the framework's escalation triggers fire; the shortcut is the private oracle replacing the medicines-information call; the evidence is a documented escalation with what the specialist added that the synthesis lacked.
From mapping to revalidation
The mapping's practical payoff is that AI-era prescribing generates portfolio evidence continuously if the workflow captures it: a genuine clinical question, the sources consulted and how they differed, what changed in knowledge or action, and the reflection on the tool's limits, is a complete CPD entry in the exact shape revalidation rewards, and the reflective questions write themselves from the three-part discipline, where did I use AI appropriately, where was I offered the shortcut, what evidence did this leave? Building that capture into the daily loop, question to sources to reflection to record, is precisely the workflow this platform's CPD layer exists to hold, and the direction of regulatory travel suggests prescribers who install the habit now will find the coming expectations describing what they already do.
Frequently asked questions
Does the framework require me to use AI at all?
No: it requires reliable sources critically evaluated, and the mapping shows how AI use can serve that requirement; a prescriber meeting the competencies without AI is fully compliant, and one using AI carelessly is not.
How should supplementary prescribers apply the mapping?
Identically, with the clinical management plan as an additional governing document in every competency's verification step: answers must fit inside it, whatever the wider evidence would permit.
What single habit best future-proofs against the NMC's expected changes?
The documented verification: sources opened, checks recorded, limits reflected on; every signalled direction of the consultation points at exactly that behaviour, already required in substance by competency 2.7.