Access
People need practical ways to locate, understand, and navigate care across geography, systems, and differing local realities.
Harvard Medical School Executive Education · 2025
Transforming Advocacy and Access
A proposed, culturally grounded, AI-supported strategy for healthcare access and advocacy in rural and Indigenous communities.
A healthcare transformation capstone that became a broader argument about the structural, cultural, and relational changes organizations must make in response to AI.
- Context
- Director of Organizational Development, Physiotherapy Association of British Columbia
- Work
- Strategy, organizational development, AI transformation
- Status
- Capstone proposal with a limited early advocacy prototype
01Overview
A tool proposal became a study of organizational readiness.
This capstone was developed through my work with the Physiotherapy Association of British Columbia as part of Leading Digital Transformation in Health Care, an executive education program from Harvard Medical School.
It began with a practical question: could an AI-supported platform help rural and Indigenous communities navigate health care services while helping practitioners and communities turn local knowledge into more timely public advocacy?
The proposed platform was the key piece in the very beginning. But the further the project developed, the clearer it became that technology was only the visible layer of a much larger transformation.
02The challenge
Access and advocacy were being treated as separate problems.
The project began in physiotherapy, but the pattern extends across and beyond all Allied Health professions. Rural and Indigenous communities face overlapping barriers to health care access, continuity of care, digital participation, and influence over the policies that shape services. Practitioners may understand the local need but lack the time, evidence, or organizational support to turn that knowledge into sustained policy influence.
How could technology support access and public voice without reproducing the exclusions already present in the health care system?
A shared platform for access, advocacy, and participation
This was a strategic proposal—not a completed public product.
- 01
Culturally safe access
Resources for understanding, locating, and advocating for appropriate health care services.
- 02
Localized advocacy
AI-assisted letters, briefs, and policy messages grounded in local circumstances.
- 03
Policy-window awareness
Signals from public discussion and policy developments that could make advocacy more timely.
- 04
Indigenous data governance
Community authority, data sovereignty, cultural protocols, and defined limits built in from the start.
- 05
Secure Canadian infrastructure
Privacy-conscious systems suited to Canadian health care and nonprofit contexts.
03The early prototype
A small experiment, not a product launch
The technology was the easiest part of the problem.
A limited prototype used early Custom GPTs to explore automated advocacy drafting. The tools were primitive by today’s standards, but they made the underlying possibility tangible: AI could reduce some of the labour involved in turning local experience into policy communication.
They also exposed the more important constraints. A better model would not determine who had authority, create trust, redesign workflows, establish culturally grounded governance, or make an organization capable of listening and changing course.
A technically functional system could still fail if:
- communities were consulted but lacked decision-making power;
- staff did not understand or trust the tools;
- advocacy remained disconnected from clinical realities;
- data governance arrived after the system had been designed;
- success was reduced to adoption or efficiency;
- local differences were flattened into one provincial model.
Responsible AI adoption is not a software rollout. It changes how an organization listens, decides, learns, shares power, and acts.
04Transformation
Six conditions for meaningful transformation
These conditions work as one system. Weakness in any one can undermine the legitimacy and usefulness of the whole.
Advocacy
Clinicians and communities need support turning local knowledge into credible and timely policy intervention.
Governance
Those most affected must have real authority over a system’s design, data, uses, limits, and evaluation.
Organizational capacity
Staff and members need time, literacy, support, and permission to shape tools rather than simply receive them.
Clinical relevance
Technology must remain connected to frontline care, actual advocacy practices, and differing regional conditions.
Relational implementation
Trust is developed through long-term partnership and shared accountability, not extracted through consultation.
Three strategic shifts
Fragmented advocacy
ToShared voice
Isolated services
ToConnected access
Technology administration
ToOrganizational learning
05Governance and cultural safety
Governance before scale
The capstone envisioned a lean, cross-functional structure in which cultural safety, clinical value, and technical feasibility would be accountable to one another.
Indigenous Partnership Circle
Proposed to embed community authority, cultural governance, and Indigenous data sovereignty from design through evaluation.
Clinical Digital Leadership
Proposed to connect the platform to clinical practice, PABC member realities, patient needs, and frontline workflows.
Technology Guidance
Proposed to oversee infrastructure, privacy, feasibility, implementation, and alignment across the organization.
Shared accountability
for public benefit
for public benefit
A necessary correction
These groups were proposed, not established. Indigenous people did not participate in the capstone itself, and the project did not receive Indigenous endorsement. Any responsible future implementation would need to begin with Indigenous leadership and community-defined authority—not add consultation to a predetermined design.
Where First Nations are involved, OCAP® principles are not a compliance label or automatic guarantee. They point toward a larger obligation: governance must determine what is built, what data may be used, who benefits, what limits apply, and whether the work should proceed at all.
06Implementation
Start small enough to learn
The proposed operating rhythm rejected a single large rollout in favour of focused pilots, feedback, and earned expansion.
- 01Listen
- 02Co-design
- 03Pilot
- 04Evaluate
- 05Adapt
- 06Scale
Make learning operational
Four-to-six-week micro-sprints, listening sessions, modular deployment, manual alternatives, open feedback loops, peer mentorship, and lessons-learned documentation would allow the organization to adapt before expanding.
Let partnership be structural
Partners would help define purpose, governance, benefits, limits, and measures of success from the beginning. Growth would be earned through trust, evidence, and mutual benefit—not assumed from good intentions.
Build workforce judgment
Digital maturity would not be measured by the number of tools adopted, but by the organization’s ability to question, shape, responsibly use, and sometimes refuse them.
Measures of success
Adoption was never enough.
Access
Navigation to appropriate services, regional reach, reduced difficulty locating care, and continuity between information and treatment.
Participation
Clinician and community engagement, repeat participation, diversity of representation, and the quality of feedback.
Influence
Responses from public officials, evidence entering policy discussion, service commitments, and new partnerships.
Trust
Partner-defined cultural safety, transparent AI use, confidence, meaningful governance, and the ability to correct or opt out.
07Reflection
What the project changed in my practice
Organizations do not become ready for AI by purchasing better tools.
I began with the idea of an AI-enabled advocacy and access platform. I finished with a much larger understanding of transformation.
The technology was one part of the system. Meaningful implementation also required cultural change, Indigenous governance, relational partnership, clinical leadership, privacy, workforce learning, and the institutional ability to revise or stop what was not working.
Organizations become ready by strengthening how they make decisions, share responsibility, protect people, learn from experience, and remain aligned with their purpose.
The question now extends far beyond one association.
Healthcare associations—and, by inference, most nonprofit organizations—cannot respond to AI simply by adopting new tools. They must also examine the structures, attitudes, authority, skills, relationships, and institutional habits that determine how those tools will be used.
This orientation work became part of the foundation for Sympathetic Technology.
Questions I carry into the work now
- What problem are we actually trying to solve?
- Who is affected, and who has authority?
- What knowledge is missing?
- What must remain human?
- What would responsible refusal look like?
- What happens when the technology is wrong?
From capstone to organizational practice
AI changes more than the tools.
It changes the conditions under which organizations listen, decide, learn, and act.
Work with Sympathetic Technology