At McKesson ideaShare 2026 last week, a quiet moment in a breakout session may have been the most practically significant thing that happened all four days. Michael Deninger, chief technology officer of Towncrest Pharmacies, described a workflow that sounds futuristic until you realize it’s already running in his Iowa stores right now.
The Moment That Stopped the Room
After completing a patient consultation, sometimes lasting 45 minutes, Deninger would input a de-identified brain dump of the interaction into an AI platform, which then organized it into a complete clinical note with relevant diagnostic and Current Procedural Terminology codes. What previously required 45 minutes of documentation time now takes approximately 5 minutes.
Sit with that for a second. A 45-minute patient consultation producing a fully documented, CPT-coded clinical note in 5 minutes. That is not a marginal efficiency gain. That is a complete restructuring of what is possible in a single pharmacist’s day.
If documentation burden drops by 40 minutes per patient encounter, and a pharmacist currently sees three to four billable clinical patients per day, the math on what becomes newly possible is genuinely significant. In clinical volume. In billing throughput. In patient impact. The hours saved on documentation are hours available for the next clinical encounter. And every encounter documented with AI assistance is one that can actually be billed at its true CPT value, with the note quality to support the claim.
The key, Deninger said, is identifying the most painful inefficiencies first. His immediate action step for pharmacists who haven’t started was even simpler: “The first thing they need to do is to step in and ask a chatbot a question. It doesn’t even have to be pharmacy related.”
The Journey That Got Him There: From Skeptic to Implementer in Six Months
Deninger’s path is worth understanding specifically because it mirrors where most independent pharmacists currently stand.
Six months before ideaShare, he described himself as largely skeptical of AI tools, uncertain of their reliability and wary of overpromising technology. As the platforms matured, he began incorporating AI into his daily operations with measurable results. One example stood out: reconciling insurance claims data across multiple pharmacy locations. What had previously taken colleagues tens of hours to sort manually, mapping insurance carriers across stores to distinguish Medicare, Medicaid, commercial plans, discount cards, and hospice coverage, Deninger accomplished in roughly 2 hours by feeding de-identified claims data into an AI agent. The result was a consolidated, usable reporting structure for all his locations.
Six months from skeptic to demonstrating a 9:1 documentation time compression on a national conference stage. The skepticism wasn’t the obstacle. The first use case was.
The Three AI Tools Towncrest Is Running Right Now
Deninger described a specific tool stack, not a theoretical wishlist, and each component addresses a different layer of the pharmacy workflow.
AI-assisted SOAP note generation: The core documentation workflow, a de-identified post-visit brain dump converted into a fully structured clinical note with diagnostic codes and CPT suggestions. This is the tool that turned 45 minutes into 5.
OpenEvidence for clinical literature access: He introduced staff to OpenEvidence, an AI application that draws answers from the primary medical literature rather than general web sources, allowing technicians and pharmacists to quickly access clinically reliable information on unfamiliar topics.
OpenEvidence is the most widely used medical AI and clinical decision-support platform among U.S. physicians, with over 757,000 verified physician registrations and a milestone of exceeding 1 million consultations in a single day in March 2026. It is trained exclusively on peer-reviewed journals and guidelines, cites every claim, and withholds answers when uncertain. A PMC-published study found it provided accurate, evidence-based recommendations aligned with physician treatment plans in primary care settings, with high ratings for clarity, relevance, and evidence-based support.
OpenEvidence is free for verified U.S. healthcare professionals and includes evidence-grounded search with citations from peer-reviewed literature, Visits for clinical documentation with ambient note generation, and as of March 2026, Coding Intelligence for automatic ICD-10, evaluation and management, and CPT suggestions.
OpenEvidence’s Coding Intelligence delivers automatic CPT code sequencing to maximize reimbursement, ICD-10 diagnosis suggestions that reflect actual complexity of the encounter, and E/M level recommendations with the full MDM rationale already written into the note. For every hour of patient care, physicians spend nearly two additional hours on documentation. OpenEvidence generates the MDM breakdown automatically from the clinical note.
AI data entry bot: At one store, Towncrest implemented an AI data entry bot that autonomously processes orders and refills, freeing technicians for higher-value tasks.**
Three distinct tools, three distinct problems: clinical documentation compression, evidence access at the point of care, and routine data entry automation. None of them require proprietary infrastructure. Two of them are free.
The Compliance Architecture Matters
Before implementing any of these tools, pharmacists need to understand the HIPAA compliance landscape around AI documentation. The Deninger workflow depends critically on de-identification: the brain dump that goes into the AI platform contains no protected health information, no patient names, dates, or other identifiers. The clinical intelligence is preserved; the PHI is removed before any AI platform touches it.
OpenAI launched ChatGPT for Clinicians on April 22, 2026, and it is free for any verified U.S. physician, nurse practitioner, physician assistant, or pharmacist. It includes optional HIPAA support through a Business Associate Agreement. Standard consumer ChatGPT remains outside this compliance framework. If you are a verified U.S. clinician, ChatGPT for Clinicians is the free on-ramp with HIPAA cover available.
The practical taxonomy for AI use in pharmacy documentation breaks down into three distinct categories. General-purpose AI platforms like ChatGPT handle tasks that are primarily linguistic: note drafting from de-identified summaries, patient education template creation, prior authorization language, staff training materials. These work best with de-identified inputs and optionally with a BAA. Clinical decision support platforms like OpenEvidence handle evidence lookups, drug information questions, and guideline searches: trained on peer-reviewed literature, not general web content, appropriate for any clinical question a pharmacist would otherwise look up in a reference. Ambient documentation platforms handle real-time visit capture and structured note generation from audio: HIPAA-compliant by design, typically EHR-integrated.
The Deninger workflow sits at the intersection of the first and third categories. De-identified post-visit summary input into a general-purpose AI generates the structured note, which the pharmacist then reviews, corrects, and finalizes before placing it in the patient record. Human review is not optional. It is the compliance checkpoint that makes the entire workflow defensible.
Why Documentation Has Always Been the Hidden Constraint
The single biggest constraint on pharmacist-led clinical services has never been scope of practice, reimbursement philosophy, or patient demand. It has been time.
A pharmacist buried in dispensing verification, prior authorizations, and documentation cannot be in two places at once. The documentation burden is particularly acute in clinical pharmacy because the note quality directly determines billing success. A poorly documented MTM encounter may be accurate clinically but insufficient for reimbursement. A pharmacist who completes a comprehensive medication review and then spends 40 minutes documenting it has net positive impact on the patient and net negative impact on the clinical visit economics of their practice.
AI-assisted documentation that compresses that 40-minute note into 5 minutes doesn’t eliminate the pharmacist. It multiplies the pharmacist. Every hour saved on documentation is an hour available for a clinical encounter. Every encounter documented with AI assistance is one that can be billed at its true CPT value with the note complexity to support the reimbursement.
A clear message emerged from the session at ideaShare: the transformation is already underway, and pharmacies of every size can participate. Independent pharmacy owners who identify their most painful operational friction points, start with a single manageable use case, and apply appropriate skepticism to AI outputs are already seeing meaningful gains, in efficiency, in data quality, and in the time available for patient care. The tools are here. The question now is where each pharmacy wants to begin.
The Playbook: One Use Case, Thirty Days
Deninger’s journey into AI began with identifying the most painful inefficiencies first. Not the most impressive use cases. Not the most transformative. The most painful.
That is the Towncrest playbook, and it is specifically calibrated for the pharmacist who has been watching the AI conversation from the sidelines because the full transformation felt overwhelming.
This week, identify the single most time-consuming documentation task in your current clinical workflow. Prior authorization letters. MTM SOAP notes. Comprehensive medication review documentation. Patient education summaries for complex regimens. Refill counseling scripts for technicians.
Pick one. Then spend 20 minutes testing whether an AI platform, OpenEvidence for clinical questions, ChatGPT for Clinicians for note drafting, or any other platform you have access to, can meaningfully accelerate it. Not replace it. Accelerate it.
The first test doesn’t need to be perfect. Deninger’s first AI experiment was an insurance reconciliation spreadsheet, not a clinical documentation revolution. The clinical documentation revolution followed. It followed because he completed the first test, measured the result, and moved to the next use case.
One use case. Thirty days. Measure what changes. That is how the 45-minute consultation becomes a 5-minute note, and how a pharmacy with three billable clinical patients per day becomes one with five.
Sources: Pharmacy Times (McKesson ideaShare 2026: Why Automation and AI Are Reshaping the Dispensing Experience, June 2026), Pharmacy Times (McKesson ideaShare 2026: Recapping Key Insights from Denver, June 2026), OpenEvidence (Coding Intelligence Launch Press Release, March 26, 2026), OpenEvidence User Guide (Visits, Coding Intelligence, CPT Suggestions), Fierce Healthcare (OpenEvidence Launches Hands-Free Voice AI Feature, Expands Hospital Footprint, May 21, 2026), Commure (ChatGPT for Doctors: How to Use AI in Your Practice in 2026, June 2026), iatroX (Free Medical AI Tools Worth Using in 2026 by Country), Physician AI Handbook (AI Tools Every Physician Should Know, May 2026), Crescendo AI (2026 AI News, Innovations, Breakthroughs in Healthcare)