The Intelligence Age: Why healthcare leaders must act now
- Sukhman Kalra
- May 10
- 11 min read
Artificial Intelligence (AI) is no longer advancing in isolated breakthroughs; it is compounding into a system-level transformation, becoming embedded into workflows that can reason, act and coordinate across entire ecosystems.
For health leaders in Canada, this is not a distant future. It is an immediate strategic inflection point.
The publicly funded health system is already under strain from workforce shortages, long wait times, tightening fiscal constraints, and rising demand for digitally accessible, equitable, high-quality care. In this environment, AI should not be viewed as a novelty or a collection of point solutions. It should be treated as a capacity multiplier (especially in documentation, triage, planning, and patient navigation) and, more importantly, as a design principle for how health systems need to operate differently.
The question is no longer whether AI will affect health care. It is whether Canadian health leaders will use it to redesign care delivery, or simply layer it onto outdated workflows.
What is changing in the AI landscape?
Signal 1: From tools to executable workflows
AI is shifting from passive models/tools (that respond to prompts) to autonomous and collaborative agents capable of executing multi-step workflows, coordinating across systems (and with other agents), reasoning, and adapting to dynamic environments.
The next stage will likely be more proactive: systems that anticipate needs, surface issues before they escalate, and take action within defined guardrails.
Implication: Organizations will move from deploying AI as a tool to embedding AI as a workforce layer.
For health systems, this means AI will not just:
Support clinical decision-making, but rather care coordination will be co-handled by AI agents (with a human-in-the-loop)
Act as documentation tools, but rather execute end-to-end administrative workflows with minimal manual intervention
The strategic question shifts from: Where can AI help? → Which workflows should be rebuilt with AI?
Signal 2: AI as core infrastructure
Like the internet before it, AI is becoming a foundational capability embedded across operations rather than a standalone innovation.
Implication: Competitive advantage will depend not on access to AI, but on how meaningfully and effectively AI is integrated into an organization's operating model.
For health systems, this means:
Differentiation will come from AI-enabled care models, not isolated pilots
Organizations that redesign care pathways and workforce models will outperform those that layer AI onto legacy systems
Signal 3: Compounding gains through learning systems
AI systems are increasingly capable of learning from their outputs, accelerating performance over time. This is possible due to the learning loops, replicability, scale/volume and workflow automations. Anyone who has been using Large Language Models (LLMs) over the last few years has witnessed this steep learning curve.
Implication: We are entering a phase of compounding / non-linear productivity gains.
For example, in an idea exchange with ChatGPT, I asked it to explain how it drew an inference. It generated two responses in a side-by-side comparison and asked me to select the response I prefer. This is similar to how we (as humans) generate options for a task and use the feedback from the intended audience to refine our approach to achieve the intended result.
For health systems, this means:
Clinical pathways can continuously improve using real-world data (especially as outcomes are observed)
Patient flow can be dynamically optimized
Triage models can become more accurate as they process more cases and outcomes
This creates a widening gap, as early adopters will compound gains and late adopters will face accelerating inefficiencies.
Signal 4: Productivity Gains → System Transformation
AI is moving beyond incremental efficiency toward end-to-end system redesign.
The advancement of AI depends on access to high-quality data, interoperable and computational infrastructure, and coordinated ecosystems. These will become key assets for any organization that wants to deploy and embed AI systems at scale and achieve increasing gains.
Implication: Organizations that treat AI as a piece of the puzzle (i.e, an IT investment) will see limited benefits. The organizations that treat it as a design principle will fundamentally reshape how they operate.
The shift: AI strategy → System strategy.
For health systems, this means:
Data governance becomes a core / strategic capability, not a compliance function
Interoperability becomes a prerequisite for value creation
Partnerships across public, private, and academic sectors become essential to unlock value and scale
In Canada, the challenge is not data scarcity; it is fragmentation across provinces and institutions.
Where can AI deliver measurable value in the Canadian health system?
As I stated before, the healthcare system in Canada is at an inflection point. An aging population, rising chronic disease burden, workforce shortages, and fiscal pressure are pushing the publicly funded system toward a structural constraint that traditional approaches alone will not solve.
According to CIHI, Canada’s total health spending is projected to have reached ~$399 billion in 2025, growing at ~4.2% annually. That is 12.7% of Canada’s GDP and an average of $9,626 per Canadian. Even with that level of investment, the system will struggle to keep pace with inflation, population growth, and rising demand.
That means the next wave of improvement cannot come from cost containment alone. It must come from care system redesign, supported by AI.
I see five key areas for impactful and meaningful use of AI for health system transformation:
1. Unleash Workforce Capacity (Immediate Impact)
AI’s near-term value in health care is strongest where work is repetitive, data-rich, and time-sensitive, i.e., documentation, scheduling, demand forecasting, referral management, and population risk stratification.
Potential Use Cases
Automated clinical documentation and transcription using ambient AI
Digital referral workflows via an AI-driven central intake models (e.g., this can be achieved through command centre capabilities in Ontario at the Local Delivery Group or Ontario Health Team-level)
AI-enabled scheduling, patient flow optimization and patient follow-ups (e.g., med adherence, symptom check-ins, etc.)
AI-assisted patient intake, triage, and communication with human oversight (i.e., human-in-the-loop)
Why It Matters
Workforce capacity remains the binding constraint amidst workforce shortages and access challenges across Canada. Administrative burden directly limits workforce capacity, especially available clinical time. Alleviating administrative burden through AI applications will expand system capacity without new hiring.
Note: This is not meant to suggest that we don’t need to address the workforce shortage crisis; rather, that AI can create efficiencies that provide clinicians more time for patient care.
Possible Constraints
Fragmentation across EHRs (Epic, Meditech, Oracle) and community EMRs prevents end-to-end visibility
Privacy and compliance requirements
Procurement timelines
Clinician adoption
Potential Impact
Increased effective clinical capacity without hiring
Reduced wait times, faster transitions and improved access
Reduced documentation burden and improved retention
Alignment with Funding (2023–2025 Federal + 2026 Ontario Budget)
The federal health funding agreements prioritize access, wait time reduction, and data sharing, while the Ontario 2026 Budget reflects rising spending pressures driven by demand. Automation of administrative functions / workflows is one of the few levers that directly convert funding into access gains.
2. AI-Driven Care Coordination (Near-Term Impact)
Potential Use Cases
Patient journey coordination across integrated care teams (e.g., OHTs)
Automated discharge planning between hospitals and home care providers (e.g., via Home and Community Care Support Services in Ontario)
Navigation support for seniors across post-acute, community, and long-term care
Chronic disease pathway management
Why It Matters
Canada’s health system is structurally fragmented. Patients, particularly seniors and those with complex care needs, experience disjointed transitions between hospital, home care, and long-term care. These gaps cause:
Avoidable readmissions
Extended hospital stays (ALC patients)
Poor patient and caregiver experience
Possible Constraints
Limited interoperability across provider settings
Data sharing gaps and a lack of standardization
Governance complexity
Potential Impact
Reduction in ALC days
Shorter acute care stays (i.e., less time in the highest cost of care setting)
Lower readmissions and ED revisits
Improved patient experience and continuity of care for seniors and complex care patients
Alignment with Funding (2023–2025 Federal + 2026 Ontario Budget)
Federal priorities emphasize coordinated care and data sharing, while Ontario’s 2026 Budget continues to invest in integrated care and community-based services to reduce pressure on hospitals.
3. Clinical Decision Augmentation (Scaling Phase)
The next opportunity is not AI replacing clinical judgment. It is AI improving the consistency, speed, and prioritization of clinical decisions.
Potential Use Cases
AI-supported diagnostic triage
Risk stratification using CIHI datasets
Clinical decision support embedded in CIS and workflows
Why It Matters
Canada faces persistent specialist and diagnostic backlogs and variability in care. AI can standardize prioritization and improve clinical consistency and how constrained specialist capacity is allocated.
Possible Constraints
Regulatory approval
Workflow integration
Clinician trust and adoption
Data quality and standardization
Potential Impact
Faster diagnosis and treatment
Reduced variability
Improved patient outcomes and safety
Efficient use of specialist capacity
Alignment with Funding (2023–2025 Federal + 2026 Ontario Budget) Both federal and provincial funding priorities emphasize reducing backlogs and improving system performance. AI-enabled triage ensures limited capacity is allocated effectively.
4. Reactive care model → Proactive care model (Transformational Shift)
Potential Use Cases
Remote patient monitoring programs
AI-driven risk identification (e.g., ability to identify high-risk patients)
Chronic disease management programs with automated escalation
Virtual agents to enable adherence, patient engagement and care plan follow-ups
Why It Matters
Demand is increasingly driven by chronic disease and an aging population. A reactive system is fiscally unsustainable. Proactive and preventative care shifts / reduces the risk of intervention, decreasing reliance on hospitals and emergency departments.
Possible Constraints
Uneven access to digital tools / infrastructure
Limited integration between virtual care platforms and core clinical systems
Funding models that still prioritize episodic, in-person care
Patient digital literacy and engagement barriers
Potential Impact
Reduction in avoidable hospitalizations and ED visits
Improved chronic disease outcomes
Lower cost growth
Increased patient autonomy and engagement
Alignment with Funding (2023–2025 Federal + 2026 Ontario Budget)
Federal investments in digital health and data infrastructure, combined with Ontario’s focus on community-based care expansion, support a shift away from hospital-centric models.
Manulife is one of the organizations in the Canadian system driving a move toward a proactive health model by providing its members access to a platform that connects them to preventative health, early diagnosis and intervention, and virtual care services. There is not only untapped potential in the data that can help drive population health analytics, but also an opportunity to establish public-private partnerships to better orchestrate care for Canadians.
5. System-Level Optimization (Long-Term Transformation)
Potential Use Cases
AI-driven bed management and patient flow optimization
Surgical scheduling optimization across regions
Workforce planning and staffing allocation
System-wide resource optimization (e.g., ICU beds, diagnostics) across regions
Why It Matters
Capacity constraints in Canada are not only due to a lack of infrastructure. They are also driven by inefficient utilization, poor coordination across institutions, and a lack of real-time system visibility. Even as governments invest heavily in infrastructure, without optimization, much of this capacity will go under-utilized.
Possible Constraints
Fragmented governance across provider settings and regions
Limited real-time operational data
Resistance to centralized decision-making
Potential Impact
Increased utilization of existing infrastructure and capacity
Reduced wait times and surgical backlogs
Improved patient flow and reduced overcrowding
Alignment with Funding (2023–2025 Federal + 2026 Ontario Budget)
Ontario’s 2026 Budget reflects significant increases in health spending driven by utilization pressures, while federal transfers continue to grow steadily. AI-driven optimization ensures funding translates into real capacity, not absorbed inefficiency.
Across all five areas of transformation, a pattern emerges - AI will not just improve efficiency, it will also unlock capacity in a system where capacity is a constraint.
For Canadian healthcare leaders, this is a critical shift:
From adding resources → maximizing existing ones
From fragmented delivery → coordinated systems
From reactive care → proactive management
The Strategic Reality for Canadian Health Leaders
For health leaders, the most credible AI opportunities are not the most futuristic ones. They are the tools that reduce clerical burden, improve flow, and help teams make faster, better decisions using existing data.
AI is a capacity multiplier: With funding growth constrained relative to demand, AI becomes the primary lever for expanding system capacity (e.g., patients seen per day, OR utilization, bed turnover, etc.).
The constraint is system design: Technology is not the barrier; fragmentation, governance, and workflows are.
Workforce roles will evolve, not shrink: AI will augment, not replace, clinicians, and redefine roles & responsibilities / scope of practice.
Data is the strategic asset: Canada’s advantage lies in establishing a longitudinal view of a patient’s health data, and the value depends on activation, not collection.
We need to be pragmatic and:
Prioritize high-ROI use cases (admin automation, triage, coordination)
Invest in data and integration infrastructure
Redesign clinical and operational workflows, not just adopt tools
Build governance, trust frameworks and strategic partnerships
Lead with a clear point of view on AI’s role in health system transformation
As an example, refer to the work being led by the Mental Health Commission of Canada and the Canadian Centre on Substance Use and Addiction to co-develop the world’s first National Guidance for AI in Mental Health and Substance Use Health Care. It is planned for launch in 2026/2027. The framework is focused on three pillars:
Trust and explainability → patients deserve to understand how AI affects their care
Human-centred care → AI supports practitioners, it doesn’t replace the human relationship
Equity and data governance → ensuring AI doesn’t replicate or amplify existing disparities in mental health access
Note: AI introduces real risks and structural trade-offs that leaders must manage.
Poor data quality = Clinical risk: For example, poor data quality can result in erroneous triage decisions, resulting in high-risk patients being deprioritized and bad outcomes.
Hallucinations, bias and optimization trade-offs: While a human-in-the-loop is essential to mitigate hallucinations and/or biases in AI decision-making, it is equally important to recognize that the learning loops could result in AI making trade-offs between optimization and patient needs. For example, an AI care coordination tool handling post-acute admissions could be focussed on optimization and result in prioritizing healthier patient for LTC beds over those who might need the beds the most.
Workflow misalignment = Adoption failure: If AI adds to the workload of clinicians, adoption fails.
Centralization vs Autonomy: Centralization improves efficiency but may reduce provider control.
Speed vs Safety: Human-in-the-loop is essential for AI agent deployment in healthcare to control potential errors / hallucinations.
Vendor lock-in: Closed systems limit scalability. Open APIs and interoperability are essential to scalable gains across the health system.
Cybersecurity: As AI agents gain permissions to access different data and enterprise systems to automate tasks, organizations cannot underestimate the importance of robust permission-based systems and privacy and security guardrails.
What should health system leaders do in the next 90-180 days?
Health leaders do not need a grand AI strategy document first. They need a disciplined starting point. In the next 90-180 days, leaders should:
Establish AI governance with clinical, operational, digital, privacy, legal, and patient representation. AI governance should address key areas, such as - policy framework, accountability, risk classification, human oversight, data governance, model transparency, bias monitoring, security protocols, audibility, explainability.
Identify one or two high-volume workflows where AI can create measurable value. Ensure the selected AI workflows generate value that directly aligns with the organizational strategic priorities.
Define success metrics in advance, including time saved, wait reduction, quality, and staff experience. Ensure an evaluation is conducted after the pilot to inform the scaling of the AI workflows.
Map data gaps, workflow blockers, and safety risks. The goal is to have strong data governance and open standards and interoperable systems for the selected workflows.
Develop a change management plan for the people who will actually use the workflow.
Pilot AI with human oversight and redesign the workflow, not just the tool. Continuous monitor for performance, safety, equity, and workflow impact.
Bold Predictions for Canada’s Health System (Next 5-7 Years)
Integrated Care Teams → AI-Orchestrated Care Networks: Integrated care teams (e.g., OHTs) will evolve into AI-enabled coordination platforms, where agents will manage patient journeys across care settings in real time, turning structural integration into operational reality.
Funding Models Shift to Enable Value-Based Care: As fiscal pressure intensifies, governments will increasingly tie funding to access, outcomes, and efficiency, enabled by AI-driven measurement and transparency.
A Federated National AI and Data Framework Emerges: Canada will develop a coordinated but federated approach to health data and AI governance, enabling innovation while maintaining provincial control and privacy standards.
The Intelligence Age in healthcare will not be defined by smarter tools, but by systems that can think, act, and improve at scale.
Canada’s health system does not have the luxury of gradual change. The fiscal math for it no longer works.
The question for health leaders is not whether AI will transform the system. It already is. The question is whether we as leaders will shape this transformation deliberately or be constrained by it.
In my next article, I plan to explore how healthcare organizations can start this journey and capture the momentum – i.e., What does this AI momentum mean for healthcare leaders on Monday morning?
Share your thoughts (what does this AI momentum mean for your organization)…ideas (which workflows would you prioritize)…concerns (what are your biggest showstoppers)...and let’s explore together.
Comments