My notes and reflections from Canada’s largest digital health gathering.
This week, I spent three days at e-Health26, Canada’s largest digital health gathering, hosted by Digital Health Canada, CIHI, and Canada Health Infoway, in Halifax. I returned home with one strong belief: the future of Canadian healthcare won’t be decided by who launches the most impressive AI. It will be determined by who makes care trustworthy, connected, patient-focused, and led by the people who provide it.
What I saw on the floor
AI was the main focus at nearly every booth. There were ambient scribes, AI “protocoling” for imaging, and AI agents that summarize charts. However, when I spoke with the teams behind these technologies, the most engaging aspect was rarely the model itself. It was the workflow surrounding it: how a referral moves, how a form becomes a task, and how every step is recorded.
The AI was the spark. The plumbing was the product.
That plumbing kept pointing to the same simple problem: forms, faxes, and scattered data. One platform’s whole pitch was “Axe the Fax. Fix Referrals.” Others changed paper requisitions to digital formats or combined patient reminders to reduce missed appointments. If you remove the branding, they all have one goal: capture structured information, move it through a process, and keep it trackable. This is the daily reality of Canadian healthcare, and very little of it needs advanced AI.
Data sovereignty is now table stakes
The strongest vendors didn’t focus on model size. They prioritized PHIPA compliance, Canadian data residency, SOC 2, and ISO 27001. Buyers are asking tougher questions now. They want to know not just “what can it do?” but also “where does our data live, who can see it, and can we prove it?”
The case for restraint
The most memorable session connected everything. In her keynote, Dr. Julielynn Wong of 3D4MD argued that we should not assume the advance of AI in healthcare is inevitable. She emphasized that knowing when not to use AI is just as important as knowing when to use it. Her test was refreshingly practical: is the AI significantly safer, simpler, and cheaper than the alternative, and can its creators show transparency and accountability? She advocated for independent, made-in-Canada systems that keep humans in control.
In a room full of AI, the loudest applause was for restraint.
What stood out to me was her focus on humanitarian issues. She encouraged us to consider the true cost of these tools, not just the license fee, but the total impact on communities. Every dollar spent on a hidden system is a dollar not invested in effective care, and these models have carbon and water footprints that are often overlooked in business discussions. Her point wasn’t to oppose technology. It was about being mindful: selecting tools with a human and humanitarian perspective rather than simply chasing the latest trends.
Why this felt familiar
I walked the floor with a focus. These are the problems we tackle every day at AOT Technologies. Our open-source platform, formsflow.ai, converts forms into managed, automated workflows. It is designed to be deterministic and built on open standards such as BPMN and DMN. It offers end-to-end auditing and uses AI only when it genuinely adds value, such as helping to build forms. The “axe the fax,” the digital requisition, and the patient engagement workflow are all examples of what formsflow.ai is designed to do. Open standards and interoperability, measured use of AI, deterministic workflows, and controlled data residency are not just extras. They define the platform. Because it is open source and operates on infrastructure controlled by the hospital, transparency and data residency are built in.
It’s the same principle Dr. Wong defended on stage.
Here is where I stand. Most of this isn’t a tough technical problem. You rarely need advanced AI to improve patients’ experience. AI should focus on genuinely difficult issues, like ambient scribing, image analysis, and early cancer detection. However, the majority of today’s patient experience can be significantly improved with simple solutions: clean forms, clear workflows, and systems that communicate with one another. Using the right tool for the right task is more effective than applying AI everywhere. In healthcare, a reliable, trackable workflow that avoids errors isn’t a drawback; it’s essential. For a country like Canada, which has a small population but a large public budget, directing that funding toward the simpler issues could greatly enhance patient experience. Being patient-centered isn’t just a slogan; it’s a measure: is the person on the other end truly better off?
Three takeaways for Canadian healthcare
1. Integration beats isolation: The floor was full of point solutions that don’t communicate with each other, from population-health platforms to integration specialists. Canada doesn’t need more islands. It needs the connections of open standards and interoperable workflows.
2. Open source is a strategy, not just a budget item: When the platform is open, hospitals can understand how decisions are made, keep their data at home, and avoid being tied to one vendor’s plans. In a moment of “buy Canadian,” that’s a purchasing strategy, not idealism.
3. AI belongs inside trusted workflows, not on top of them: The right spot for AI is as one step that a human supervises, drafting a note a clinician approves or flagging a message in a patient-engagement tool, not as an unaccountable decision-maker. Get the workflow and governance right, and AI becomes truly useful, not just impressive.
Where we’re going
I left Halifax feeling hopeful. The discussions have progressed. We are moving beyond the idea that “AI will fix everything” toward more important questions about trust, ownership, and whether a tool improves patients’ or clinicians’ lives. That’s the premise behind formsflow.ai: open standards instead of silos, predictable workflows that use AI where it makes sense, and data that the institution owns. The most advanced model doesn’t usually determine who gets better care. It’s the reliable, everyday systems that do. Get that right, and we can build a brighter future.
With thanks
This change comes from people, not platforms. The conversations were the real takeaway. I want to thank everyone who influenced my thinking in Halifax.
For patient access and virtual care, thank you to Dr. William Cherniak and Harry Cherniak (Rocket Doctor), Valda Duke (Fonemed), Scott MacIntosh (ClinicGlide), and Swami Sambasivan (Carelynk). For forms, workflow, and interoperability, a big thanks to Daniel Cheng (Process Fusion), Declan FitzGerald (DMF Systems), Komal Sardana (Ocean eReferral), Saravana Rajan (Pumkin AI), Damanpreet Kaur, and Jordan Schroecker (Venuiti). For patient engagement, I appreciate Luke Callanan and Tim Bonnell (TxtSquad). For population health and responsible AI, thank you to Hannah Seo (Verto Health) and Robyn Stewart (NTT DATA). For delivery and transformation, thanks go to Chris Bath and Mark Ferris (Barrington Consulting) and Ben Goldberg (Kainos). And for the operational backbone, I appreciate Peter Trajceski (StaffScheduleCare). I also want to thank Fahad Munir, Garrick Bikku George (Sunco), Jag Basrai, and Jubin Jose for discussions that went beyond the booths.
The technology will continue to improve. It’s the people like these who will hold it accountable.