Enterprise Governance,
AI Risk & Digital Trust

Enterprise Governance &
Digital Transformation
Management Consultant

Recovering stalled AI revenue beyond legacy governance and standard frameworks.

Recovering stalled AI revenue beyond legacy governance and standard frameworks.

Strategic Public Sector AI GTM:
From Sovereign AI Infrastructure to Public Trust Outcomes


CUSMA Did Not Fail on July 1, 2026. Something Larger Just Began.

How the architect of institutional trust foresaw, nearly two centuries ago, the Sovereign AI question Canada now faces as CUSMA enters its first trilateral joint review. Reflections on the day the United States declined to extend CUSMA, and the day Canada's sovereign AI conversation moved from policy aspiration to institutional accountability.  


CUSMA Did Not Fail on July 1, 2026. Something Larger Just Began.

How the architect of institutional trust foresaw, nearly two centuries ago, the Sovereign AI question Canada now faces as CUSMA enters its first trilateral joint review. Reflections on the day the United States declined to extend CUSMA, and the day Canada's sovereign AI conversation moved from policy aspiration to institutional accountability.  

Six years ago CUSMA replaced NAFTA.

Six years later, the United States Trade Representative formally declined to extend the agreement. Canada and Mexico had both signaled willingness to extend to 2042. The agreement remains in force but shifts into annual reviews for the remainder of its current term. The next round of talks convenes in Mexico City on July 20.

The treaty layer of North American trust is now openly in contest. That does not diminish the sovereignty conversation. It clarifies it.

Because if the trilateral trust framework we have relied on for six years can no longer be assumed as continuity, the question of what sovereign AI produces inside each Canadian institution stops being an abstract policy debate. It becomes the practical answer to what Canada actually controls when the treaty layer is contested.

There is a Millian question underneath the sovereignty work that classical political economy answered long before AI existed. 

Six years ago CUSMA replaced NAFTA.

Six years later, the United States Trade Representative formally declined to extend the agreement. Canada and Mexico had both signaled willingness to extend to 2042. The agreement remains in force but shifts into annual reviews for the remainder of its current term. The next round of talks convenes in Mexico City on July 20.

The treaty layer of North American trust is now openly in contest. That does not diminish the sovereignty conversation. It clarifies it.

Because if the trilateral trust framework we have relied on for six years can no longer be assumed as continuity, the question of what sovereign AI produces inside each Canadian institution stops being an abstract policy debate. It becomes the practical answer to what Canada actually controls when the treaty layer is contested.

There is a Millian question underneath the sovereignty work that classical political economy answered long before AI existed. 

John Stuart Mill argued that institutions are judged not by the control they preserve, but by the tangible improvement they create in the lives of the people they serve. It is a nineteenth century idea. It has aged into something newly urgent.

Sovereign AI meets a harder test than infrastructure delivery. When state investment is measured in fiscal terms, the public is implicitly promised returns. In the fiscal architecture of a democracy, those returns are not measured in shareholder value. They are measured in citizen outcomes: faster services, more accurate decisions, lower administrative friction, greater institutional trustworthiness. Public trust is the yield curve of state investment. Nothing else legitimizes fiscal architecture over time.

CUSMA reviews trade flows, dispute mechanisms, treaty compliance. It does not and cannot review whether Canadian sovereign AI produces Canadian public trust, or whether the compute simply enables decisions whose economic value migrates elsewhere. That review belongs to Canadian institutions, and it does not depend on Washington's willingness to extend a treaty. 

Three questions the treaty review does not ask, but every Canadian institution should and today, more urgently:

John Stuart Mill argued that institutions are judged not by the control they preserve, but by the tangible improvement they create in the lives of the people they serve. It is a nineteenth century idea. It has aged into something newly urgent.

Sovereign AI meets a harder test than infrastructure delivery. When state investment is measured in fiscal terms, the public is implicitly promised returns. In the fiscal architecture of a democracy, those returns are not measured in shareholder value. They are measured in citizen outcomes: faster services, more accurate decisions, lower administrative friction, greater institutional trustworthiness. Public trust is the yield curve of state investment. Nothing else legitimizes fiscal architecture over time.

CUSMA reviews trade flows, dispute mechanisms, treaty compliance. It does not and cannot review whether Canadian sovereign AI produces Canadian public trust, or whether the compute simply enables decisions whose economic value migrates elsewhere. That review belongs to Canadian institutions, and it does not depend on Washington's willingness to extend a treaty. 

Three questions the treaty review does not ask, but every Canadian institution should and today, more urgently:

- When AI systems trained on Canadian data make decisions affecting Canadian citizens, does the improvement in citizen outcomes justify the fiscal architecture that funded it? Not in output metrics. In trust metrics.

- When AI reduces the cost of expertise inside institutions, does the value accrue to those institutions and their citizens, or migrate to the platforms that industrialized the expertise? (Ford recently rehired engineers it had replaced with AI. The lesson was not that AI failed. The lesson was that institutional context and decision accountability had left the building along with the people.)

- When a public sector AI capability delivers a citizen outcome, is the accountability for that outcome located within Canadian legal and institutional authority, or contractually distributed across jurisdictions that CUSMA cannot reach and today, may not protect?

A necessary clarification about what this argument is not.

It is not a call for digital isolationism. Canada does not need to build walls around its data or its models. Nor is it a lament for treaty continuity. Treaties are instruments, not foundations.

The stronger frame is a durable trilateral Norh American trust architecture between three sovereign economies, in which each partner is stronger because the others are stronger. Today's decision does not remove that possibility. It removes the assumption that treaty structure alone will produce it. Durable trilateral trust must now be actively built, not passively inherited.

CUSMA at its best is not a fence. It is the operating principle of trilateral trust between sovereign economies. Sovereign AI at its best is the same principle applied inside each of them.

Between and within, the architecture is the same. And when the outer structure is contested, the inner one is what determines whether the outer one gets rebuilt.

Six years into CUSMA, six years into the first review, and one year into the era of sovereign AI economics, the review that matters is not only the one that finished this morning.

It is the review each Canadian institution now owes its citizens.

- Infrastructure creates capacity.
- Governance creates yield.
- Accountability creates value.

That is the architectural work of the next twelve months. And today, it just became more urgent.

Open to selective advisory collaborations and industry speaking ->


Credentials:
Recognition and experience

2025 CIO Award recognized for leadership in digital transformation and enterprise analytics

Automotive Retail Leadership Extensive experience operating in high-pressure, performance-driven retail environments

Data & AI Strategy Specializing in real-time decision systems and governance for data-driven organizations

North America Cross-market experience across enterprise operations and executive alignment

CDO Magazine, IDC, Strategy Institute public speaker on governance, AI, and operating models for modern enterprises

2025 CIO Award recognized for leadership in digital transformation and enterprise analytics

Automotive Retail Leadership Extensive experience operating in high-pressure, performance-driven retail environments

Data & AI Strategy Specializing in real-time decision systems and governance for data-driven organizations

North America Cross-market experience across enterprise operations and executive alignment

CDO Magazine, IDC, Strategy Institute public speaker on governance, AI, and operating models for modern enterprises

Contact & Availability

  • Infrastructure creates capacity. Governance creates yield.

  • Capacity without accountability is not sovereignty. It is exposure

    If this resonates, start a conversation ->

Email: engage@andrerekhtine.com

Website: andrerekhtine.com | LinkedIn

Available for advisory and consulting engagements across North America.