AI in Pharmacy: Knowing What not to Automate.

Written by Judit Mora, CEO and Co-founder, Nuumad 
Published on 27th July 2026 in Scottish Pharmacy Magazine

The use of artificial intelligence in healthcare is evolving rapidly. But knowing when and how to use AI is essential, especially in the community pharmacy sector.

The Pharmacists’ Defence Association (PDA) recently published new guidance which highlights a principle that should sit at the heart of every discussion about AI in pharmacy: technology may influence professional decisions, but it cannot take responsibility for them.

The PDA is clear that pharmacists must retain active oversight, professional judgement and accountability when AI is involved in their work. It also highlights the need for particular caution when AI provides recommendations, interacts directly with patients, reduces the visibility of decision-making or changes clinical processes.

A recent statement from the GPhC on AI also made it clear that pharmacy professionals remain personally accountable for their decisions and actions and must always use their professional judgement when AI tools are involved.

At Nuumad, we believe that clinical decisions belong to clinicians. For us this is a technology design principle as well as a regulatory a governance issue.

Removing the administrative burden not clinical judgement

We have built a platform that is intelligent, sophisticated and highly automated without asking artificial intelligence to become the clinician. Our objective is to remove unnecessary administrative burden, structure consultations and support pharmacists with transparent, evidence-based workflows. It is not to to replace clinical judgement with probabilistic AI.

A well-designed clinical workflow does not need AI generating recommendations in real time. What matters is whether the system produces a consistent, transparent and validated pathway that a pharmacist can understand, scrutinise and apply within their professional judgement.

It is why Nuumad uses tested, rule-based clinical algorithms rather than probabilistic AI for clinical decision-making. Where probabilistic AI system may generate an answer based on patterns learned from its training data, a rule-based clinical pathway operates according to defined logic and predetermined clinical criteria. For clinical decision support that distinction matters. A 2026 systematic review of predictive AI clinical decision-support systems found that performance varied significantly by setting and highlighted the gap between technical success and the real-world.

Confidence through consistency

While AI does undoubtedly have a role to play in healthcare, context and human oversight remain essential. A pharmacist needs to know what information has been considered, what pathway has been followed and what clinical criteria underpin the outcome. They also need to be able to apply their own knowledge and experience to the individual patient in front of them.

Technology should reduce the risk created by unnecessary complexity, inconsistency and administrative pressure, it should not reduce the pharmacist's responsibility. The PDA's position reinforces this point.

The Nuumad platform was designed to support the pharmacist by structuring the consultation, automating administration, guiding the workflow and providing evidence-based decision support. The clinical judgement however, remains with the qualified pharmacist. It demonstrates that technology can support a clinical decision without making the clinical decision.

Simplicity matters in a regulated profession

Introducing AI into a clinical environment brings additional considerations around governance, regulation, data quality, bias, system limitations, monitoring and professional accountability. The PDA's guidance highlights many of these issues for pharmacists to consider. This does not mean pharmacies should avoid AI altogether, but they should question whether they actually need AI for a specific task before introducing it.

Often automation can be achieved through deterministic, tested algorithms and structured workflows, that can offer a much clearer route to consistency and oversight. That is the approach we have taken with Nuumad. Our platform automates the parts of the consultation process that do not require a pharmacist's clinical expertise, while keeping the clinical decision firmly with the pharmacist. The opportunity for technology in community pharmacy is not to make pharmacists less important, it is to give them more time with patients using the expertise that patients rely on them for.

In contrast, administrative processes, data entry and repetitive workflow tasks are necessary, but they do not require the full extent of a pharmacist's clinical expertise. Technology can take more of that burden away.

An important conversation to have

The PDA's announcement is an important contribution to a conversation that pharmacy needs to have right now, rather than when AI has become embedded everywhere. We need to think carefully about where AI can generally add value and where it adds clinical risk and complexity.

Innovation in healthcare should be measured by how effectively technology helps healthcare professionals do their jobs. The smartest pharmacy technology is not the technology that tries to think like a pharmacist, it is the technology that gives pharmacists more time, more consistency, greater transparency and better workflows so they can do what only they can do - apply their clinical judgement to the patients in front of them.