Project summary

Canmora Collective is building the first operational protocol standard for AI decision accountability in consequential government decisions. This is not a framework or a white paper. The actual sequence - preregistered criteria, bifurcated AI/human scoring, configuration logging, divergence review, and integrated public accountability records - that a grants director follows on Tuesday morning when applications land, and that produces a decision any legislator, auditor, or rejected applicant can reconstruct. The same architecture applies to procurement, budget analysis, and legislative drafting: any consequential government decision where AI materially shapes the outcome.

The norms governing how AI enters government decision-making are forming now. Thirty-six governors' seats are on the ballot in 2026, with at least twenty-one incumbents leaving - incoming administrations will set AI decision defaults from scratch. Getting accountable process architecture in place before less accountable approaches become entrenched is among the most tractable near-term interventions available in AI governance. This is the window.

What are this project's goals? How will you achieve them?

Three deliverables in three months:

Deliverable 1: Organizational infrastructure completed: entity formation, USPTO trademark filing, and fiscal sponsorship agreement executed. This is the legal and operational foundation for everything that follows.

Deliverable 2: Protocol v1.0 finalized and published as a free open field standard any government agency can adopt independently. An early version is drafted now.

Deliverable 3: Confirmed recruitment of the first government demonstration cohort - four to six governments committing to implement the protocol across at least one full decision cycle. Relationships across a 36-state network and direct contacts from a decade of grantmaking reform work make this immediately actionable, not cold outreach.

How will this funding be used?

Three months of founder time - salary ($28,517) and benefits ($7,985): $36,502. USPTO trademark filing and legal costs: $1,500. Technology: $1,000. PPF fiscal sponsor overhead at 6%: $2,340. Total: $41,342.

Personnel is 87% of direct costs before PPF overhead. No contractors, no travel, no events. This is minimum viable infrastructure to produce three concrete deliverables in ninety days.

Minimum funding of $30,000 covers organizational infrastructure, technology, PPF overhead, and partial founder time sufficient to complete protocol publication. Full funding of $41,342 adds the remaining founder time for government cohort recruitment.

Who is on your team? What's your track record on similar projects?

Patrick Carter, founder. I have spent a decade building operational infrastructure that moves government - not policy recommendations, but the process architecture that changes how decisions get made and resources get allocated.

At Minnesota Management and Budget, I directed reform of $1.2 billion in annual state grantmaking - the precise decision context this protocol addresses. At Results for America, I lead national strategy across 36+ states: an implementation guide I authored directly preceded a doubling of state adoption of evidence-based budget principles from seven to fourteen states in under three years, and a scorecard redesign I led has driven concrete policy change in multiple states.

Published in Public Administration Review. Regular speaker and advisor to government leaders through CSG and NCSL. 80,000 Hours coaching participant, EA Global attendee.

What are the most likely causes and outcomes if this project fails?

If this project fails, the operational gap in AI decision accountability for government remains unfilled during the critical norm-formation window. Agencies racing to adopt AI without process architecture establish defaults that become self-reinforcing - through procurement requirements, vendor optimization, and staff training - that are orders of magnitude harder to displace once entrenched. A delay of even twelve months meaningfully reduces the tractability of this intervention. The risk of moving too slowly is not abstract: it is the quiet accumulation of unaccountable decision patterns across thousands of government processes that collectively shape the lives of hundreds of millions of people.

How much money have you raised in the last 12 months, and from where?

$0 - pre-funding, currently seeking anchor funders.