AI in Action: Transforming Clinical Care Across Canada Working Group

The scan catalogues initiatives across provinces, territories, care settings, technology types, and stages of deployment. It is intended to help leaders see where AI adoption is visible, and where opportunities and gaps may exist.

This environmental scan was conducted by a group of volunteers and relied exclusively on publicly available sources, including media reports, health system announcements, vendor publications, and outputs from national funding programs. Each province and territory was reviewed individually using a standardized framework to support comparability. The following fields were recorded for each initiative: Province/Territory, Initiative Name, Organization, Type of Clinical Setting, AI Use Case Category, Technology Type, Stage of Deployment, Description of Application, Date of Launch/Announcement, Funding Sources, Partnerships, Outcomes/Impact, References/Links. All entries were verified for correctness prior to inclusion in this environmental scan and pivot tables were utilized to draw key insights and generate counts across standardized categories. While this scan provides a structured view of visible AI activity in Canadian healthcare, several important limitations must be acknowledged.

• Public-source dependency: The scan is based on information available through health system announcements, funder databases, and media. Undisclosed pilots or internal deployments are not captured, potentially underrepresenting the full scope of AI activity.
• Sectoral imbalance: Hospitals and health system–level deployments dominate public reporting. Primary care, long-term care (LTC), community health, and Indigenous/remote settings are much less visible.
• Vendor vs. clinical confirmation: Some rows are based on vendor or funder announcements without clear evidence of sustained clinical use.
• Dynamic landscape: With 89 projects in pilot stage, some may already be scaled or discontinued, and new initiatives are likely since the scan’s close.

These statistics are updated as new initiatives are added to the scan. Last updated October 27, 2025.

Scale of activity

175 initiatives were documented across Canada. Larger provinces (Ontario, Quebec, British Columbia) account for the majority, while smaller provinces and territories often report only 1–3.

Clinical settings

AI initiatives were most commonly deployed in hospital settings (51 initiatives), followed by acute care (30), primary care (25), specialty clinics (15) health system (14) ambulatory care centres (11). Diagnostic imaging (9), Community health (6), long-term care (6), public health (5), and remote care (3) are comparatively rare, underscoring an imbalance across care environments.

Technology patterns

Machine Learning (ML) (80) is most common, particularly for prediction and triage. Computer Vision (34) is concentrated in diagnostic imaging and endoscopy. Natural Language Processing (NLP) (37) appears primarily in AI scribes and chatbots. Deep Learning (DL) (7), Large Language Models (LLMs) (6), Robotics (2), and are less frequently documented. Nine (9) initiatives remain categorized as AI (general) due to vague or incomplete descriptions.

Stage of deployment

The majority of initiatives (100) remain in the pilot stage, while 38 are reported as in active use, 19 as scaled deployments, and 9 as system-wide or still in planning.

AI Use Case Category

Diagnosis & Decision Support (56) and Communication & Coordination (56) were most common, followed by Triage & Risk Stratification (30), Treatment & Intervention (19), and Patient Education & Engagement (8), and Follow-up & Remote Monitoring (6).

Signals for leaders

• Workflow integration matters: Tools embedded into existing clinical processes (scribes, stroke imaging, deterioration models) are more often reported as “in use” or “scaled.”
• Equity gaps remain: The underrepresentation of primary care, LTC, and remote/Indigenous settings suggests AI benefits may be unevenly distributed unless equity is intentionally prioritized.
• Evidence is scarce: Few initiatives publicly report outcomes, making it difficult to assess effectiveness, return on investment (ROI), or scalability.
• Future watch: Early appearances of LLMs and robotics highlight areas of experimentation but not yet wide adoption.

This scan is a living resource. We invite healthcare providers and organizations to share details of their clinical AI initiatives, so the database remains accurate and current. Feedback on inaccuracies or omissions is welcome. Please contact chief@digitalhealthcanada.com. The AI in Action: Transforming Clinical Care Across Canada Working Group will continue building on this work—engaging leaders across Canada, surfacing lessons learned, and supporting responsible AI adoption. Our shared aim is to ensure that AI in healthcare advances in ways that are evidence-based, equitable, and patient-centred.

AI Clinical Function Use Case Categories

• Triage & Risk Stratification: AI tools for early detection of high-risk patients, symptom assessment, and predictive analytics.
• Diagnosis & Decision Support: AI-driven technologies supporting radiology, pathology, and clinical decision-making.
• Treatment & Intervention: AI-enhanced robotic surgery, personalized medicine algorithms, and real-time intervention tools.
• Follow-up & Remote Monitoring: AI-powered patient monitoring, wearable device analytics, and predictive analytics for post-discharge care.
• Patient Education & Engagement: AI-driven chatbots, virtual assistants, and personalized patient education platforms.
• Communication & Coordination: AI solutions for clinical workflow automation, provider communication, and administrative efficiency.

Clinical Venues

• Hospital: Inpatient, general or specialized hospital facilities (regional, community, academic/tertiary).
• Acute Care: Units/services for time-sensitive, high-intensity medical needs (ED, ICU, critical care, oncology wards, diagnostic imaging within hospitals).
• Ambulatory Care Centre: Outpatient hospital-affiliated clinics (e.g., wound, diabetes, oncology follow-up).
• Community Health: Non-hospital local health centres, community clinics, Indigenous health centres.
• Primary Care: Family practice, GP clinics, virtual-first care platforms.
• Long Term Care: Nursing homes, residential facilities, supportive housing for seniors.
• Public Health: Population-level health monitoring, screening, disease surveillance.
• Specialty Clinic: Stand-alone or hospital-affiliated specialty units (endoscopy, fertility, orthopedic).
• Health System: Multi-site or integrated health networks spanning hospitals, clinics, LTC, and/or public health.
• Diagnostic Imaging: Radiology and imaging services, including standalone centres or hospital-based units (e.g., MRI, CT, ultrasound, mammography, stroke/neuroimaging).

Technology Types

• AI (general): Used when details of the underlying model are not specified or span multiple techniques.
• Machine Learning (ML): Algorithmic models trained on structured or semi-structured data for predictions or classifications.
• Deep Learning (DL): Neural network–based AI, typically used for imaging, signal, or complex pattern recognition.
• Natural Language Processing (NLP): AI tools for analyzing and interpreting human language (transcription, summarization, entity extraction).
• Large Language Models (LLM): Transformer-based generative models (e.g., GPT-type), specialized for text generation and reasoning.
• Computer Vision: AI applied to image/video recognition, detection, and segmentation.
• Robotics: AI-enabled robotic systems in surgical, diagnostic, or rehabilitation contexts.

Stage of Deployment

• Planning: Announced or being designed; not yet piloted.
• Pilot: Limited-scale testing in a real-world setting.
• In Use: Deployed in routine practice but not yet scaled across multiple sites or full systems.
• Scaled: Broad deployment across multiple units/sites within an organization.
• System-wide: Adopted at the full health-system or multi-institutional level.