The Pharmacists Patient Care Process (PPCP) in the Age of AI: A Pharmacist’s Playbook
TLDR: AI won’t simply “fit into” the PPCP; it will re-weight it—automating portions of Collect, Implement, Follow-up, and Communication now, while moving into bounded areas of Assess/Plan later under explicit guardrails. The 2025 PPCP already leans into person-centered, team-based, tech-enabled care; our job is to decide where AI belongs—and where it doesn’t.
The Pharmacist Patient Care Process (PPCP) was updated in August 2025 after its initial inception back in 2014 by the Joint Commission of Pharmacy Practitioners (JCPP). The PPCP has always been an interesting item for me, as it can be argued to be the extension of the traditional pharmacist roles focused on a product-based model (prescribing, transcribing, dispensing, administration, and monitoring) towards a more clinically focused model that the PPCP represents, entailing:
Collect – “The pharmacist ensures the collection of necessary subjective and objective information about the patient to understand the relevant medication and medical history, overall health status, and other pertinent factors. Information may be gathered and verified from multiple sources (e.g., the patient, caregiver, observations, existing patient records, other health care professionals).”
Assess – “The pharmacist assesses the collected information to identify and prioritize patient needs to inform the establishment of a care plan. In the context of new and existing goals of therapy and other patient goals.”
Plan – “The pharmacist develops a person-centered, evidence-based, cost-conscious care plan in partnership with the patient and/or caregiver and in coordination with other care team members.”
Implement – “In providing person-centered care, the pharmacist implements a prioritized care plan in partnership with the patient and/or caregiver and in coordination with other care team members.”
Follow-up: Monitor and Evaluate – “In providing person-centered care, the pharmacist implements a prioritized care plan in partnership with the patient and/or caregiver and in coordination with other care team members.”
*In addition to these five items, pharmacists must also be able to collaborate, communicate, and document to accomplish the PPCP.
Why this matters to me (and to us)
I trained before the PPCP existed and then watched it roll out as an educator. It’s been a background frame for my teaching—SOAP for documentation, therapeutics for assessment, etc. At the same time, I’m a technologist and semi-futurist who knows the workforce our students enter will be turbulent. So this piece is about where AI practically helps the PPCP now, and where we should pace ourselves—without losing the “why” behind clinical judgment.
Step-by-Step: Where AI will Impact the PPCP (now and later)
Collect
The one area I think that AI will make perhaps one of the biggest impacts on is how we take medical data and collate and feed it out to those who need to use it. EHRs can still be a jumbled mess. Patient data is sporadic. Communication between different software systems often takes middleware or an external portal. It’s a mess.
Take medication reconciliation, for instance, an item that in theory should be as simple as asking a patient what medication they take. And yet, we expend tremendous resources to determine multiple items that is both subjective and objective in nature. What prescription were they prescribed? Did they get it? Do they take it? Do they use it right? What other agents are they also taking? When? Why? Did they work? Etc.
Then where is this data? The pharmacy? Which one? Which EHR? Where’s the vaccination records? Where’s the list of previous medications and agents? Over the course of the patient's life, this list will ebb and flow. The cognitive labor associated with keeping this straight alone can drive a caregiver into burnout.
So yes, a problem with an opportunity for disruption that I think anyone would welcome. At this point, you should say, “But companies and vendors are out there solving it!” And I would agree, to a point. They have made the process better, and saved time and efficiency, but it’s still a mess. Errors still exist; ergo, we can still do better.
And this is just one instance that I can point to that most health professionals can relate to, it doesn’t get into the issues in separate specialties and therapies that also have their own baggage.
The rise of digital health technologies (DHTs) is only going to add on to this process, whether it's real-time patient data (e.g., vitals), ePROs, genomic data, etc. Data is great, but access and utility are hampered.
AI I think has a great opportunity to help us do better with the mass patient data that will inundate us by the end of the decade, and close health silos amongst health systems will be unable then to keep up with changing market demands and payment schemes (especially if you want value based care) as the work force burns out drowning in this data stack.
Having AI help better manage and organize this data will be key. Creating mass summaries using NLP from clinic notes, real-time scribing, and post-care notes will be drastic to the healthcare space for the collection of data.
Assess/Plan
This is where my consternation lies. We’ve long struggled with “cookbook” medicine; adding GenAI hallucinations doesn’t help. The core risk is students (and clinicians) knowing the what/when but skipping the why. My stance: AI can rank options; it cannot replace accountability. Near-term, keep AI advisory in bounded domains (e.g., protocolized insulin titration), require source-linked outputs, and document the pharmacist’s rationale (“why this, why not that”). Over time, some therapeutic areas will qualify for AI-assisted assessment with validation, KPIs, and pharmacist sign-off. None of this removes the pharmacist’s duty to reason.
I say this because we’ve been having issues with the use of medical information for years. I do not want to go back to analog data formats of paging through a book to find information, but the rise of certain platforms that can almost generate cookbook medical practice can also lead to issues on treating the individual patient. Add on the integration of GenAI and the lingering issue of hallucinations and we see this getting more problematic. Every week I hear from colleagues about providers entering orders for therapies that do not exist, don’t work, or make no sense, and increasingly when the provider is challenged, they repose “the [AI tool] says it works!”
It comes back to knowing Why. Bypassing this ability to assess, whether due to lack of knowledge, skills, experience, or laziness causes this problem. We want to save time with technology and AI may help address the work associated with find an answer, but we cannot bypass the quality assurance of that answer.
So then, that’s why the application of AI in assessment will prove troublesome. It’s not that it isn’t feasible. It’s that the users (e.g., health workforce) may not do it well. So, we have a lot of training still before I can say AI is ready for this area, and will need much more time. Keeping the human/pharmacist in the loop is key, and if we start having CDSS and related software integrated into the EHR that help us assess patients that impact care, we will need further refinement of these AI tools to get there.
Saying all that, I will acknowledge though that AI can be a good tool currently with several facets of helping to assess and plan care for the patient, even in its current form. I see continuous research ebbing and flowing on the ability of models to diagnose and analyze data, each with its own caveats. That’s enough for me to think that we likely will eventually get there, even if we must spin our wheels for a little bit. I suspect that what will happen is that some AI models will qualify to help with certain therapeutic categories enough to warrant us depending on them to deliver care. An example will likely be insulin management with the use of AI/ML models that will likely automate the whole process in the coming decade.
Using AI tools to make sure the planning component works out probably makes the most sense. HTN management, initiated, which therapy, and how to see it through, can be a rote format that AI can help support for health teams. I see this as helping to ensure the safety of much of what we do with our medications once it’s decided what to utilize.
Implement
Great plans die on logistics. AI can orchestrate tasks, benefits checks, scheduling, referrals, and personalized education, closing gaps between a plan and the first dose. This is ripe now. AI tools could readily be created to support helping with care coordination and data coming in regarding personalized patient education and lifestyle management. Helping to further schedule further care plans to support the assessment and planning of a pharmacist, I think AI really can help carve through the logistics of care. After all, every great plan falls apart at the start, and being able to adapt is key to seeing the process through. However, that’s always been a huge logistical burden that AI can help with in terms of actual implementation.
Follow-up: Monitor and Evaluate
Similarly, F/U is another good area for the use of AI tools to help with data coming back from patients and interaction asynchronously. AI can watch signals asynchronously (RPM, ePRO, refill/adherence patterns), surface safety/efficacy concerns, and prompt timely human escalation, closing the loop we too often lose in daily churn. How are those medication changes working out, for instance? What actionable insights do we have to help intervene to ensure efficacy and safety of treatment? Closing the loop can often be lost in the glut of everyday work, and having the ability of AI to help us monitor what is going on will be to fully implementing the PPCP loop.
Communicate/Collaborate/Document
Finally, the soft skills to make the PPCP work. These skills are essential to work with patients and the rest of the health care ecosystem. While I would argue much of this is taught and reinforced in the pharmacy curriculum, IPPE/APPEs are the point where we can push students to really practice them.
That being the case, I think this is probably one of the most low-hanging fruits for AI utility that will hit healthcare now. We see this with AI documentation and scribing popping up, and EHR integration with many tools to help facilitate communication across multiple healthcare teams and departments. We must prepare our students to enter this workforce with these skills.
AI × PPCP at a glance
What’s the Timeline to Prepare for?
I would say the immediate practical concerns of AI and the PPCP are as follows:
Now (0-12 months) – Immediately work with preceptors to have them get students involved with any AI tools that are being used, especially related to Communicate/Collaborate/Document, which are probably being rolled out within the workforce (e.g., EHR, PMS, related)
1 – 5 Years – Prepare students to understand how to integrate AI tools with Collect, Implement, and Follow-up in the PPCP. It is likely a lot of this will be operations/administrative in nature, and heavily vendor-supported across multiple pharmacy environments. I expect it’ll be sporadic in nature and not necessarily as quick as we see with the communication aspect, as vendors are still creating and building, but it's likely to get wider market traction in the next five years as they work it out with pilots and testing.
3 – 7 Years – This is debatable. Some may say this will go faster than I think but I do think AI in Assess & Plan will take longer. Several reasons include the issue of regulatory and scientific accuracy and application. Specialties will range in speed for the application of AI in their sectors. Educationally, we have an uphill battle. And then there are the cultural issues of acceptance by the health professions and patients themselves. I expect this to take a long time to hammer out, and that in some areas of pharmacy, this could go faster than others, depending on where you practice. So, from an educational perspective, be heavily adaptable in whatever sector you are practicing.
Pharmacy Cross-Sector Outlook
While it is hard to predict where AI impacts the PPCP, and what that may entail for each type of practicing pharmacists, these are some examples of where I think we will see AI rollout and impact very soon, and possible areas to highlight for current students and for practitioners to look into:
Community: AI pre-reconciles meds/immunizations; pharmacist resolves exceptions and counsels.
Ambulatory: HTN/diabetes pathways surface titrations; pharmacist documents reasoning and SDOH barriers.
·Inpatient: Ambient scribing and med-rec triage shrink pre-rounds time; pharmacist focuses on DTPs/deprescribing.
Specialty: AI flags access/toxicity risks and coordinates REMS tasks; pharmacist monitors labs/adherence.
Managed care: Population signals prioritize outreach; pharmacists design high-value interventions.
Industry/Reg/MedAffairs: Safety signal detection from real-world data; pharmacists interpret and educate.
Hard Questions We Need to Answer Now
There remain several areas that I still believe need to be addressed in the short term, as they could pose as barriers or rate-limiting steps for AI integration into the PPCP. Most of these questions span multiple pharmacy stakeholders, and I do think a lot of professional work will need to be put in as we think of policy and governance in the future, such as:
Who owns the assessment? - If AI proposes a plan, does the note explicitly show independent pharmacist reasoning?
Disclosure - When and how do we tell patients an AI tool contributed to the assessment/plan?
Liability - If software suggests a non-existent dose and we sign it, where does accountability land?
Equity - Will AI-triaged follow-up miss those with low digital access, which contradicts PPCP equity aims?
Formulary pressure - What happens when payer-preferred algorithmic pathways conflict with clinical nuance?
Documentation - What minimum elements prove critical thinking occurred (e.g., “Model suggested X; pharmacist accepted/rejected because Y”)?
Closing Thoughts
If you’ve heard me speak on Digital Health and AI in the past, you’ll often hear me say that I think adoption will be ‘slow’ itineration’s within certain practice/therapeutic areas depending on current use of data as a marker for care. That’s the reason why diabetes is moving so fast, but other areas are still struggling. I always thought that Digital Health will eventually become just Health, but to be honest I still see mass transition not happening in the 2030s. I see this decade primarily focused on the administrative side of pharmacy, which honestly makes sense. We cannot go further clinically with the PPCP until the labor is lessened, which includes much of the scut labor we have to manage. Using AI in these areas will allow the clinical side to open further. Trying to do it all at once will only burn us out, and to alleviate burnout, having AI tools supplement and freeing up labor pools will help us out the most currently.
ChatGPT Ver 5.0 and Grammarly were utilized to help edit this article's content and improve formatting and readability.