April 10, 2026
On March 25, 2026, I attended MAX: Medical Affairs eXcellence Forum, Canada’s first forum dedicated entirely to medical affairs in biopharma. Organized by Agilis Health and held at the Sheraton Toronto Airport, it brought together 140+ professionals spanning Medical Affairs, Pharma, Clinical Research, Regulatory, and Medical Communications for a full day of candid, high-calibre conversation.
It was the kind of event where you leave with more questions than answers, energized by being in a room of professionals genuinely wrestling with complex, unresolved problems. Here are the five themes I have not stopped thinking about.
By the end of the morning sessions, the argument had already been made. Dr. Chris Pettengell (Pentavere/HEALWELL AI) shared results from live Canadian deployments of DARWEN™, a platform that converts unstructured clinical text into structured, actionable insights [2,3]:
Dr. Muhammad Mamdani’s (TCAIREM, University of Toronto) earlier talk further reinforced this. CHARTWatch achieved a 26% reduction in unanticipated deaths over five years at Unity Health Toronto [5,6]; AI-assisted physicians paradoxically underperformed unassisted peers by second-guessing correct AI outputs [7]; and patients rated AI responses as almost ten times more empathetic than human physician responses [8]. More highlights from his Temerty Medicine talk are available here.
“AI is structured to please you. If you’re not doing prompt engineering, you’re only seeing 10 to 20% of their capabilities.”
— Dr. Muhammad Mamdani, Director, TCAIREM, University of Toronto
The barrier to AI’s potential in medicine is not the technology itself, but rather the trust and human literacy surrounding it. Unfortunately, AI literacy in Canada remains critically low. The question is no longer whether AI belongs in healthcare — it’s whether we have the knowledge and governance structures in place to use it responsibly.
A panel featuring Michael Duong (Roche), Dung Pham (Mitsubishi Tanabe), and Nahal Nasseri (Sandoz), moderated by Amit Suri (Next Gen Pharma), landed on a consensus that stayed with me: healthcare is not short on data, it is short on infrastructure to make that data accessible and actionable.
The sharpest reframe of the session:
“The question isn’t just how AI brings value to medical affairs… It’s how medical affairs brings value to AI.”
— Michael Duong, Head of Innovation, Roche
Structured endpoints, well-annotated data, and reusable evidence packages are no longer just IT deliverables — they sit squarely within the purview of medical affairs. Every evidence plan, publication, and data asset generated by these teams becomes foundational input for building safe, accurate clinical AI. Nasseri’s experience makes this clear: her team’s hybrid chatbot was technically ready long before it could be deployed. The limiting factor wasn’t the algorithm, it was organizational trust, governance, and approval pathways, which remain the most common bottlenecks to real-world implementation.
Dr. Jacob Cookey (Virtual Hallway) brought a harder question into focus: how are clinical decisions actually being made? A Harvard study tracking pediatric cardiologists found that only 20% of clinical decisions were based on medical evidence, dropping to just 3% when limited to evidence specific to the patient [9]. A 2021 FDA study of more than 2,000 prescribers found the top two influences on prescribing behavior were colleagues and trusted peers [10]. Virtual Hallway turns that peer influence into a structured clinical network: with 11,000 clinicians and over 1 million peer-to-peer knowledge exchanges, 84% of consults kept patients off specialist waitlists, and 29% avoided an ED visit entirely [11,12].
Accuracy is necessary, but not sufficient. Medical content must reach the right person at the right time, from a trusted peer source — and be structured around a real clinical decision point.
A panel with Sophie Michaud (Galderma), Kai Carlsen (GSK), and Mark Surka (Sanofi), moderated by Stephanie Kwong (Methapharm), offered a clear and unambiguous perspective on the evolving role of the Medical Science Liaison.
AI already handles HCP profiling, document synthesis, pre-call preparation, and field report automation. What it cannot do is build a genuine relationship or show up with real-time empathy and judgment.
Sophie Michaud looks for resilient candidates and has stopped using the word “liaison” entirely. It implies passivity — information flowing in one direction — when the role is inherently bidirectional and grounded in strategic insight. Kai Carlsen added that the quality of scientific exchange now matters more than the frequency of contact. Mark Surka noted that the ideal candidate profile has shifted: individuals with computer science backgrounds paired with scientific training are increasingly valued, and hiring criteria written three years ago may already be outdated.
The 360 stakeholder engagement panel, featuring Taline Movsessian (Eli Lilly), Peyman Nakhaei (Daiichi-Sankyo), and Clinton McCracken (Leo Pharma), moderated by Daanyaal Bandukwala (INK PharmaSolutions), reinforced how this shift is playing out: moving from channel-centric to patient-centric is not a technology problem, it is a mindset problem. The future is not more touchpoints, but fewer, better-targeted, physician-specific interactions, with digital tools creating continuity between them.
MSLs must act as proactive territory owners who drive insights upward. This is not a new expectation — it is what the role was always meant to be: a strategic partner, not just a messenger.
The afternoon panel on bridging medical strategy into real-world impact featured Elizabeth Bongfen Silayuv (Alexion) and Tao Meng (J&J), moderated by Kai Carlsen (GSK).
Elizabeth Bongfen Silayuv was frank: a strategy that does not change year over year is not a strategy; it is a document. Without clearly defined unmet needs and concrete actions tied to it, it stays on paper. And it stays on paper precisely when MSLs are handed a strategy to execute rather than helping to build it from the ground up.
Tao Meng emphasized how the MSL’s role within the organization is being reframed: not as a support function, but as a strategic co-owner of the product. He was direct about the implication — stop trying to prove clinical value with commercial KPIs.
“If your field team feels empowered to make the right decision at the right moment… your medical strategy is working.”
— Tao Meng, Director of Medical Affairs, Johnson & Johnson
Strategy only reflects real-world demand when the people on the ground are the ones shaping it.
Getting strategy right externally starts with getting alignment right internally. Dr. Shurjeel Choudhri (Bayer) described arriving in Canada to find medical affairs operating in silos, disconnected from the functions it needed to influence. His response was deliberate: rebuild communication pathways first. His teams now save an average of two hours per week through AI-enabled tasks, and he pointed to synthetic advisory boards — AI personas modelled on real KOLs — as a near-term reality. The lesson is clear: AI adoption does not stall on the algorithm, but on the trust and governance structures that were never built in the first place.
This set the stage for the final leadership panel featuring Samuel Lee (Astellas), Kathy Foris (Advanz), and Glen Newell (Biogen), moderated by Brennan Smith (CTC Medical Communications), which tackled measurement. A live poll asked how confident we are in measuring Medical Affairs impact? The room answered: somewhat confident. As Samuel Lee noted, “If we answered that we were very confident, we could all go home.” Glen Newell highlighted the structural mismatch: MA outcomes take years to materialize, while commercial cycles run monthly. Kathy Foris offered a practical solution: a project tracker that captures interim indicators and builds a narrative of progress before final outcomes are realized.
Ultimately, measuring what MSLs truly do requires a shift from counting activity to tracking clinical impact. But that shift depends on alignment and trust — internally across functions and externally with partners — before meaningful metrics can follow.
What stood out most at MAX was the candour. These were not aspirational keynotes about what AI might do someday, but honest, grounded conversations from leaders actively shaping Medical Affairs at some of Canada’s largest pharmaceutical companies — honest about what is working, what is stuck, and what needs to change.
Three themes ran through every session:
At their core, many of the challenges discussed — from the knowledge-to-action gap, to content that fails to change clinical behaviour, to field insights that never reach decision-makers — are fundamentally communication problems. Issues of clarity, structure, timing, and the ability to translate complex evidence into something actionable by the right person at the right moment.
That is exactly where medical writing sits: between science and the audience, between evidence and decision.
At Craft Science Inc., this is the work we do every day. We are a women-owned team of PhD-level medical writers and scientific editors based in Toronto, focused on developing scientifically rigorous content that moves people, not just informs them. If any of these conversations resonate with your team, we would love to connect.
Thank you to Agilis Health for creating a space where these conversations can happen. Events like MAX don’t just inform — they move the field forward!
Disclaimer: The mention of specific companies, products, or organizations in this article is for informational purposes only and does not imply endorsement. The companies whose products were referenced were not consulted, involved in the preparation of this content, nor did they provide any funding or compensation.
References