Why this exists.
An institution exists when its credentials carry weight. The credential of the AI-Native Architect in regulated industries did not exist. So we built it.
Founded
2026 · Phase 0 · solo founder
Built on
AI Governance OS · Apache 2.0 methodology
Long horizon
10 years · the regulator's reference
I · The why
Compliance evidence and credentials are the same artifacts.
Across US regulated industries — banks, credit unions, insurance, healthcare — every institution now deploys AI in workflows the regulator examines. Adverse-action letters. Underwriting decisions. Clinical recommendations. Member disclosures. Fraud screens. Each is a place where AI moves between the institution and a person, and where, eventually, an examiner will ask: can you show me the trail?
Most institutions cannot. The model exists. The validation memo exists. The annual audit exists. But the daily human work — the override, the disclosure, the catch, the documented refusal — is not yet a profession. There is no shared vocabulary. There is no shared credential. There is no senior practitioner the junior can apprentice under and recognize as such.
This is the gap the Compact exists to close. Not by adding another exam-based certificate to the pile (AIGP, ISACA AAIA, ISO 42001 lead-auditor — these all exist and serve their layers). Not by selling another compliance dashboard (Vanta, Drata, OneTrust serve that layer). By making the daily human work credentialable — by treating the artifacts produced in real regulated-industry AI work as both the compliance evidence the examiner reads and the credential evidence the practitioner carries.
The structural insight is the load-bearing one: compliance evidence and professional credentials are the same artifacts seen by different audiences. An attestation ledger entry that answers a CFPB examiner is the same entry that counts toward a builder's AGRS score. The Compact's whole architecture is the answer to "what if those audiences shared a substrate?"
The single moment that made the Compact necessary
"The case against the institution wasn't that the model was wrong. It was that the institution could not show me the trail of human attention between the model and the consumer."
Paraphrase · Mass AG v. Earnest Operations · July 2025 · $2.5M settlement
II · The operating thesis
Three modes. One credential ladder. One wedge.
The Compact operates as a three-mode product: Assess, Prove, Train. Assess is how we sell — the $25K Sprint that produces a compliance Readiness Report on Day 30. Prove is the wedge — the governance plugin that emits hash-chained evidence at the moment AI is built. Train is how we retain — the cohort program and the universal-layer rollout that move the institution's workforce up the AGRS ladder.
The three modes share a substrate (the attestation ledger), a vocabulary (the Postures), a measurement system (AGRS), and a credential authority (the Compact, with the Bench as the arbitrating body). Customers begin at Assess, deepen through Prove, expand through Train. The retention is structural because the credentialed talent is provably scarce and the institution's evidence ledger is the moat.
The wedge is US mid-market commercial banks
Phase 0 sells one SKU to one buyer: the $25K Lending Risk Readiness Sprint to the CCO/CRO/CMRO at a US mid-market commercial bank ($5B–$50B assets). Three pilots is the Phase-0 exit criterion. Below three, the methodology has not been pressure-tested. Above three, the Compact has earned the right to expand.
The credential ladder is the moat
The three modes are commercially defensible for a 12–24 month window. The credential ladder — AGRS L0 through L5, the Five Postures, the Standard-Bearer Registry — is what compounds over a decade. The bet is that an open methodology adopted by 10,000+ regulated-industry practitioners becomes the regulator's reference. The closed alternative does not.
III · Anti-goals
What the Compact refuses to be.
Anti-goals are commitments — refusals as binding as the affirmations. The Compact's anti-goals survive every phase.
Refusal i
Not a knowledge-test credential
IAPP's AIGP, ISACA's AAIA, every cloud-vendor AI cert is a multiple-choice exam against a body of knowledge. The Compact is structurally different. AGRS measures evidence-portfolio performance at Kirkpatrick L3 (behavior) / L4 (outcomes). AGRS recommends AIGP at L1 and slots above it; it does not replicate it.
Refusal ii
Not "AI fluency" in the general-tool-adoption sense
The word "fluency" contaminates within ~90 days of being tied to performance review. The Compact's construct is AI-governance fluency under regulator scrutiny. We refuse to credential general AI tool adoption — that is the host employer's HR concern.
Refusal iii
Not a static curriculum
The most pedagogically sophisticated artifact in the Wave-1 corpus is a single director's weekly fluency Scout log that auto-generates the next assignment from the learner's own evidence. The Compact commits the same: curriculum surface evolves from the learner's real work, not from a pre-authored content library.
Refusal iv
Not "Diligence" as a measurable D
Anthropic's 4D framework under-specifies Diligence because most behaviors occur off-platform. The empirical literature offers no clean answer. The Compact measures behavior (artifacts emitted, reviews performed, gates passed) and refuses to measure virtue.
Refusal v
Not a competitor to host-employer HR systems
The Compact explicitly separates AGRS from host-org performance ratings. Individual AGRS scores remain restricted; only team and organizational aggregates appear publicly. The Compact's role is professional growth, never employee evaluation.
Refusal vi
Not a general-purpose coding bootcamp
The credential is AI-Native Architect for regulated industries. We do not train full-stack engineers, data scientists, or general-purpose AI/ML practitioners. The depth is in the regulator-relationship axis, not in the technical-skill axis.
Refusal vii
Not a ChatGPT-wrapper chat tool
Sentinel is a plugin-embedded co-pilot inside Claude Code, Cursor, VS Code. It does not exist as a standalone website. There is no "open Sentinel" link. Sentinel is in your editor, refusing to skip governance, when the work happens.
Refusal viii
Not a consumer education product
Paywall is enterprise-priced and seat-based. Individual tiers exist only as a Phase-3+ marketing funnel, never as primary revenue. The Compact is for regulated-institution workforces, not the general AI-curious public.
Refusal ix
Not a third-party marketplace before $5M ARR
The Standards Consortium and the Template Marketplace are real Phase-3 plans. They are also real deferrals. The marketplace opens only after $5M ARR + 25 customers + staffed regulatory ops. Earlier opening dilutes the certification authority.
Refusal x
Not a model vendor
Sentinel is provider-agnostic — Anthropic, OpenAI, Azure, Gemini. The Compact never recommends one model over another. The methodology is what's authoritative; the model is configurable. We refuse to be on any one vendor's side.
IV · The long horizon
What this becomes.
An institution exists when its credentials carry weight. In ten years, the Compact is not "a compliance vendor that added training." It is the organization that certifies AI-Native Architects in regulated industries, with the evidence platform as the proof mechanism.
The credential L4 — Industry-Specialist appears on bank AI risk job requisitions. The annual Readiness Index is cited by regulators in 60 Minutes interviews. The Standard-Bearer Registry has eighty names; the Summit fills a hall.
The bet is that the work the Compact does in 2026 — three pilots, a fifteen-seat cohort, a six-Sprint customer base — compounds because the methodology is open and the credential is authoritative. The open methodology survives the rise of competitors. The authoritative credential survives the decline of any single phase's product.
2026
Phase 0. Three pilots. AGRS v0.2 published.
2028
Phase 2 open. 200+ AGRS L2+ holders. First Standard-Bearer Summit.
2030
Phase 3. 1,000+ L3+ holders. Big 4 partnership. Insurance + healthcare verticals.
2036
The Compact is the regulator's reference. Methodology cited in named federal rules.
V · How we built this
A founder, an AI, and a methodology derived from the work.
The Compact was conceived through a structured research process: three waves of synthesis with 18+ parallel research agents, twelve source-extraction documents covering enterprise-internal AI fluency frameworks, NIST AI RMF, US banking regulators, CUs and community banks, insurance, healthcare, credentials and pedagogy, and the competitor academy landscape.
The founder is a solo operator at the time of this writing, building the Compact while working a senior software-engineering role at a payroll-platform company. The three-pillar architecture, six tenets, AGRS rubric, and Sentinel personality model all emerged from the research process, not from prior personal preconception. The methodology survived adversarial-review arcs.
This page, and the website you are reading, were authored by the founder with AI co-authorship using the same methodology the Compact teaches. The Skill Contracts that wrap the website's deployment pipeline, the bias audit on the role-picker copy, the verification of every cited regulatory anchor — all of it follows the discipline the Compact publishes. If we cannot meet the rubric, we have no business issuing it.
Apache 2.0 is the license. Anyone may build against this. The Compact is the certification authority — not because we own the methodology, but because the credential ratification has to live somewhere with continuity. We are that continuity.
The Compact began as a research synthesis and became a credential institution. The methodology is open; the credential is authoritative; the long horizon is the regulator's reference.
⛬ The institution exists when its credentials carry weight.