PDA Week 2026: Full Conference Notes from GKS

Full conference notes from PDA Week 2026 by George Kwiecinski, Denver, CO, March 22-27, 2026.

PDA Week 2026 marked a substantial shift in the maturity and competence of leaders, operators, and professionals in the understanding, acknowledgment, and rhetoric around AI tooling. AI stands to impact healthcare and life sciences in amazing ways, even in the postmarket. As it relates to manufacturing, drug development, quality, and related topics, we stand to see great benefit. In summary, the three largest takeaways from PDA Week are as follows:

1. AI maturity has increased significantly. Industry speakers no longer treat AI as theoretical. Companies like Takeda, AstraZeneca, Sanofi, and Genentech/Roche presented live deployments, not concepts. The conversation has moved past the pilot stage into questions of governance, validation, and organizational readiness.

2. The PDA Whiteboard is a significant new platform. PDA's new open-access article submission and peer rating platform lowers the barrier to sharing practical insights.

3. The industry needs more use cases and open dialogue, not more lectures. The most valuable sessions were the ones with real implementation details, honest failure stories, and open Q&A. Posters were also outstanding this year, with several presenting original research on FDA enforcement patterns, CRL trends, and AI applications in contamination control.

For a data-based summary of these topics, feel free to reference our originally posted analysis blog.

Day 1: Leadership, Audit Readiness, and Knowledge Transfer

Glenn Wright opened with PDA's 80th anniversary and 20 years of PDA Europe. The PDA Whiteboard was highlighted as the key new initiative for 2026. Members can submit articles with no peer review barrier, articles are posted within 4 business days and remain visible for 45 days, with interactive ratings that surface the most practical contributions. This is important. It is an acknowledgment that the industry is in need of better, more sophisticated tooling. We need better forums for industry feedback, and we need qualified ways to talk beyond a typical chatroom.

The opening plenary was delivered by Dr. Ed Hoffman, former NASA Academy Director and Chief Knowledge Officer (33 years at NASA), now CEO of Knowledge Strategies and Senior Faculty at Columbia University. He drew parallels between NASA and pharma as high consequence organizations where failures are far more visible than successes and knowledge management is mission-critical. He introduced the Smart Mission Canvas Framework across seven dimensions: Mission, Knowledge, Learning, Team, Culture, Story, and Global Collaboration. Dr. Hoffman has published these frameworks in his book The Smart Mission, worth a read for any team leader in the space.

Brian Trinda, GQA Director of Emerging Technology at AstraZeneca, presented the scale of AZN's global audit program: 31 sites across 15 countries, 50+ full-time auditors, and over 1,200 external suppliers in 60+ countries.

The core point of the talk was twofold. First, specialization and training are critical, but as things become more complex, ensuring more comprehensive but broad training is also important. AZN maintains a framework across 16+ specialization areas but is actively reducing granularity, instead providing technical resources to support auditors rather than multiplying credentials.

Second, there are clear steps being taken related to AI-enabled audit planning and execution: automated audit scheduling, risk-based focus area targeting, integration of regulatory data, and co-pilot tools.

When talking about the role AI plays, Brian gave a great example:

"The message we give our auditors is: you are the expert. You can't offload your responsibility to other experts in the loop."

— Brian Trinda, GQA Director of Emerging Technology, AstraZeneca

An excellent point. Using tools to prepare, rather than not preparing at all, and maintaining ownership throughout, can certainly be helpful.

Drew Bundschuh from Novavax rounded out the day with a session on human-centered strategies for knowledge transfer. The key takeaway: creating psychological safety for knowledge sharing is as important as the technical knowledge itself. He shared an example of working with an ESL team where one person kept apologizing for asking too many questions, and actively encouraging them to continue became a turning point.

The Q&A from both sessions included some notable exchanges. On creating safety in audit environments: "What it comes down to largely, in my opinion, is the auditor's acumen and their soft skills." On AI-automated audit scheduling: AZN has a dedicated person who coordinates scheduling using AI tools to match qualified auditors to sites, with the result being "an easier time and higher level of satisfaction with the team." AZN also uses technical guidances that go deeper than SOPs and is building an auditor capability webpage with week-long face-to-face training sessions.

Day 2: AI Deployments, Data Maturity, and Digital Transformation

The morning session opened with research data on the GenAI divide in pharma, reinforcing the theme that gamification and microcertification in training represents a significant shift from compliance-driven training to competency-based learning.

Brian Chivruk from Takeda presented one of the most detailed AI deployment case studies of the conference. Takeda's journey from the 2019 Shire acquisition through building AIDA, an AI investigation tool, was notable for its honesty about the sources of variation: language differences, GxP modalities, business area complexity, and the tension between global and local approaches.

The big takeaway was again similar to Day 1: an assistive tool that fills in gaps and context is still far more useful than either not doing the work, or lacking the time and expertise to accomplish a meaningful amount of the workflow. Takeda also detailed an AI maturity model for task completion: Advice Only, Propose and Approve, and Constrained Automatic. A useful framework for any team thinking about where AI sits in their operations today.

A Redica Systems session introduced the concept of Convergent Risk, where signals from quality, compliance, and procurement align to predict an enforcement action before it occurs.

Day 2 also included sessions on women's health innovation. Amy presented on the chronic underfunding of women's health research: less than 1% of global health research is dedicated to non-cancer women's health conditions. What is novel about Elvie and the collaborative landscape in this space is the work being done alongside Proov, Frame Health, Origin, Hera Biotech, Modern Fertility, MIDI Health, Kegg, Mosie Baby, and Elektra Health.

Jenny Garcia from Takeda followed with a session on clinical research participation diversity. 77% of clinical trial participants are white versus 58% of the US population, a gap that becomes more critical as US demographics shift through 2060.

PDA continues to be exceptional at granting outstanding poster spots. There were 20+ posters across a broad range of topics. It is best to refer to the full poster listings on the PDA Week 2026 page for the complete set.

Jack Prior from Sanofi Manufacturing and Supply Global MSAT closed out the day with one of the most practical frameworks of the conference. He addressed the problem that CMOs and manufacturing sites vary wildly in data readiness, and what he proposed is a ranking system across six dimensions: Fresh, Frictionless, Accessible, Authentic, Standard, and Structured.

He showcased dual heatmaps scoring sites on both what data they deliver and how well they deliver it. The best sites scored high on accessibility and freshness, the weakest on CQA visibility and data structure.

"The best way to capture long-term data for AI is better use by humans today."

— Jack Prior, Sanofi Manufacturing and Supply Global MSAT

In biologics, you rarely have big data. The greatest case is 200 batches, often just 2-3. Results from site assessments showed a 10-point average increase in maturity scores, though some sites slide back. "Large companies want smaller and smaller software footprints, which makes it hard."

Day 3: Contamination Control, Predictive AI, and Regulatory Expectations

Easily the highlight of Day 3 was the lecture-style talk by Peter Baker, a former FDA investigator. Full notes and a transcript were captured. Below is a summary.

FDA is now explicitly citing companies in warning letters for operating reactive quality systems. Peter noted this was the first time he had seen FDA use that specific language, calling proposed corrective actions "inadequate because your proposed corrective actions are based on a reactive quality system."

On the coming challenge of AI signal volume:

"We can barely manage the signals that we have now without AI. We can barely manage our quality systems at the moment without AI. Imagine the number of signals that are going to come our way with these new tools and models. They're going to point to potential process issues. We are struggling at the moment."

— Peter Baker, Former FDA Investigator

"We need to understand how to fulfill our obligations under 210 and 211 now. Because the time is coming when we will be overwhelmed with signals."

— Peter Baker, Former FDA Investigator

Using ASQ's root cause analysis framework, Baker argued that the root cause of most critical GMP failures is management strategy. Companies are so focused on getting product out the door that they neglect continuous improvement. This creates the conditions where QC lab staff run 35,000 unreported injections in wash folders to pre-test samples before official testing, hide corrections in Excel spreadsheets, and perform non-contemporaneous documentation. This example was also noted by Peter at the FDA UGA GMP Conference earlier this year. Read more in our recap here.

Peter also addressed risk assessment as a credibility crisis. Regulators no longer believe the integrity of industry risk assessments, and the idea of risk debt, similar to technical debt, captures this well: risk debt always gets paid. He cited the high rate of CRLs related to aseptic processing as evidence of inadequate risk evaluation, and challenged the standard 3-factor RPN scoring:

"You're using severity for occurrence and detectability with the exact same scoring. Probability of occurrence is just a guess. And by the way, why are you using those factors? Probably just because they were in the most popular template."

— Peter Baker, Former FDA Investigator

Peter concluded with three critical steps to transformation: critical thinking training for operators, quantitative risk assessment (a key reason ICH Q9 was revised), and following PIC/S guidance, avoiding unjustified assumptions, incomplete identification of risks, and inappropriate application of risk assessment tools.

Concluding Thoughts

Companies and innovators should continue to think about their ability to procure tools and resources internally, bridging the gap by using new technology and being educated on it. The reason implementation lags, or why implementation examples among leaders at conferences are becoming stronger, is that we need to acknowledge where the industry is as a whole. Maturity among leaders at conferences is correct and encouraging, but this represents the smallest, most forward-thinking, budget-able part of the industry.

A core reason these notes are made public and posted by GKS is alignment. Not all leaders and adopters of new technology, as well as the opposite, are in the room during these conversations. It is as much about who is in the room as who is not, in quality.

Full session recordings, transcripts, and slide notes are on file. For our data-based analysis of attended sessions, see the original PDA Week analysis post.

Explore KeyPedia, our AI-powered regulatory intelligence platform, or read our peer-reviewed research on FDA GMP inspection trends, published in the Journal of Pharmaceutical Innovation (Springer Nature, 2026).

Browse KeyPedia Lite for free access to our regulatory data, or request a demo to see the full platform.

Follow GKS on LinkedIn for ongoing updates.

Notes compiled by George Kwiecinski, Co-Founder & CEO of Global Key Solutions. Sessions attended at PDA Week 2026, Denver, CO.