Signal 011 — April 3, 2026

The Shadow Machine

Oracle fired 30,000 people via a 6 AM email while reporting record profits. The session where the cfva.ai compliance score crystallized completely — a number built from everything in public record, permanently indexed, impossible to purchase, with no floor and no expiration date.

The session where the scoring architecture became a permanent public instrument — and why companies cannot win by fighting it

30,000 Families. One Email. 6 AM.

Oracle fired between 20,000 and 30,000 employees — roughly 18% of their global workforce — via a cold email sent before most of them had woken up. The email said: today is your last working day.

Oracle reported a 95% increase in net income the same quarter. Oracle reported 22% revenue growth. The company made more money than ever. And used that money to fund massive AI infrastructure investment while sending termination notices to the people whose work generated the profits that made that investment possible.

The outrage lasted 48 hours. Then it scrolled away. It always scrolls away. The executives kept their positions. The families did not.

This is the gap cfva.ai was built to close. Not with commentary. Not with outrage. With a permanent public record that does not scroll.

30K
Jobs eliminated
+95%
Net income same quarter
18%
Of global workforce — one email

Not Just AI. Everything in the Public Record.

This needs to be stated clearly: the cfva.ai score is not only an AI compliance score. It is a comprehensive public record score. AI compliance is one layer — and a significant one given the regulatory environment being built around it. But the score reflects the full picture of what federal and state enforcement records say about a company across every domain where the government has documented misconduct.

FTC consumer protection actions. CFPB complaint volumes. DOJ settlements. SEC enforcement. OFAC sanctions. FDA recalls. OSHA violations. State attorney general actions. Every verified public enforcement record contributes. AI-related violations carry their own weight within that framework — because AI is increasingly where large-scale harm is being documented. But the score does not begin and end with AI. It begins and ends with everything the public record contains.

-210
EXAMPLE CORPORATION
cfva.ai Public Record Score — for illustration only
FTC — Consumer Protection Enforcement Action -25 ftc.gov →
DOJ — Settlement, Consumer Fraud -40 justice.gov →
CFPB — Complaint Volume Threshold Exceeded -30 consumerfinance.gov →
SEC — Material Disclosure Enforcement -20 sec.gov →
FTC — AI-Related Deceptive Practice -25 ftc.gov →
OFAC — Partial Sanctions Listing -40 treasury.gov →
Colorado AG — SB 24-205 Enforcement Action -30 coag.gov →

Every deduction has a government source URL. No URL — no deduction. That is not a policy preference. That is the architecture. The score is arithmetic applied to public record. Nothing invented. Nothing added.

The Score Has No Floor. This Is Deliberate.

The cfva.ai score can go negative. It can go deeply negative. There is no artificial bottom — and this was a deliberate architectural decision.

The reasoning is simple: the asymmetry between earning trust and destroying it is real, and the score should reflect reality. Building a clean record takes years of consistent behavior. A single OFAC sanctions listing, a DOJ criminal settlement, a sustained pattern of consumer complaints — these can accumulate in months and take years or decades to genuinely resolve.

An artificial floor would lie about that asymmetry. A score of -400 reflecting 15 years of documented misconduct across six federal agencies is not a broken number. It is an accurate one. The number being uncomfortable does not make it wrong.

The Recovery Principle

A score does not improve because a fine was paid and a press release was issued. It improves when the underlying federal record reflects genuine resolution — cases closed, consent orders lifted, regulatory relationships repaired over time.

Time and genuine behavioral change are the only path back toward 100. There is no shortcut. There is no purchase option. There is only record.

Over time, cfva.ai will publish score trajectories — showing not just where a company stands today but whether they are moving toward accountability or away from it, and how fast. A company trending from -200 toward -50 over three years tells a fundamentally different story than one holding steady at -200. Both stories matter. Both will be visible.

Global Score. Industry Rank. Two Levels of Accountability.

Every entity in cfva.ai carries two numbers: a global score and an industry rank. The global score is the absolute public record. The industry rank is the context that makes it impossible to dismiss.

A score of -50 is a number. "Ranked 2nd worst public record score in the financial services industry" is a fact that lives in board presentations, in procurement due diligence, in conversations between investors deciding whether to hold a position. The industry rank transforms an abstract number into a competitive and reputational reality that no communications team can easily neutralize.

The same logic applies in the other direction. A company scoring 94 globally might be the highest-scoring healthcare company in the country. That finding belongs in their investor materials, their sales deck, their regulatory submissions. Companies with strong records will actively promote their cfva.ai score — because everyone will know it cannot be purchased or influenced. That credibility is what makes the recognition worth having.

Both numbers update automatically. Both are publicly visible. Both index on Google next to the company name. The shadow follows them into every room where a decision about that company is being made.

The Score Is Free. The Intelligence Is Not.

The score is public. Anyone can see it. That is non-negotiable. The score being visible to investors, regulators, journalists, competitors, and the general public — before a company ever contacts cfva.ai — is the product working exactly as intended.

The free tier shows the number. No sources. No breakdown. No industry rank. Just the score. That number is enough to start the conversation that leads to the credit card.

Tier Price What You Get Who Uses It
Free $0 Score visible. No sources. No breakdown. No industry rank. Just the number. Anyone. The hook.
Basic $50/mo 250 entity searches. Full source breakdown with .gov links. Live score updates. Journalists, researchers, individuals, small businesses monitoring their space.
Pro $250/mo Unlimited searches. Competitor monitoring. Automated alerts when scores change. Law firms, compliance officers, PR agencies, investors doing due diligence.
Enterprise On demand Recovery roadmap. Active monitoring. Remediation strategy. No price listed publicly. The company itself. They call. Their situation determines the scope and the cost.

The enterprise tier carries no listed price. The company's own record sets the floor on what genuine remediation requires. The conversation determines the rest. Companies with scores above 95 pay a nominal monitoring fee. The further below zero, the higher the cost of the intelligence needed to navigate back. This is not punishment — it is proportionality. It reflects the actual work involved.

The Score Cannot Be Purchased. This Statement Is Permanent.

This appears in the footer of every page on cfva.ai. At the top of every score page. In the terms of service. In the about page. In every press communication. It will not be removed.

"Scores on cfva.ai are derived exclusively from public federal and state enforcement records. They cannot be purchased, negotiated, or removed. A score reflects public record — nothing more, nothing less. If your record improves, your score improves. That is the only path."

Every private compliance and assessment vendor faces the same structural problem: their paying customers are the companies being assessed. The product is only as credible as the vendor's willingness to publish findings that cost them the client relationship. In practice, that willingness is structurally limited.

cfva.ai has no enterprise customers whose contract depends on a favorable score. The score is derived from government records that exist regardless of whether anyone pays for anything. There is nothing to purchase because the underlying data is not controlled by cfva.ai. The only product for sale is the intelligence that helps a company understand and improve their record. The record itself is immutable.

Companies will test this. The answer will always be the same: the federal or state agency that issued the enforcement action is the only entity that can modify the underlying record. When the government updates their record, the score updates automatically. Their contact information is publicly available.

You Cannot Send a Cease and Desist to a Government URL

The legal architecture rests on one principle: every deduction has a .gov source URL. No URL — no deduction. The score is arithmetic applied to public federal and state record. It contains no opinions, no interpretations, no allegations. It contains math.

When a company's legal team reviews a cfva.ai score page and follows every source link, they land on a government website every single time. cfva.ai did not write those records. It performed addition and subtraction on records the government published and made publicly available.

"We didn't write the record. We just did the math." — cfva.ai scoring methodology

There is a second layer of protection that crystallized during this session. A company that disputes their score on the grounds of missing data has just confirmed awareness of additional enforcement records not yet indexed. The dispute mechanism is a genuine accuracy tool — documentation provided through it that reveals previously unindexed enforcement actions will be added to the score. Challenging an incomplete score makes it more complete, not less accurate.

Legal threats create their own compounding problem. A filed lawsuit is a public court record. Public court records are indexed by cfva.ai. A legal threat that escalates to a filed lawsuit becomes a new deduction. The companies that understand this fastest will take the only two paths that actually improve their situation: fix the underlying record over time, or engage with the remediation process. Those are the only moves that work.

The Only Winning Move

Fix the behavior. Resolve the federal actions. Let the record improve through time and genuine change. That is the only path to a better score.

A lawsuit makes the score worse. Silence leaves it unchanged. The only move that works is the one that was always the right move.

Nearly 4 Million Entities. The Gap Widens Every Day Automatically.

cfva.ai currently holds nearly 4 million entities and over 5 million records across dozens of federal and state data sources. The database grows automatically. New enforcement actions are indexed continuously without manual intervention required.

The competitive moat here is time. Any platform that begins building this database tomorrow starts at zero — no historical score trajectories, no years of indexed enforcement patterns, no track record of accuracy that has been tested and verified. That gap compounds every single day the existing platform operates.

After several years of operation, cfva.ai will hold something genuinely irreplaceable: a longitudinal record of how companies have behaved across time. Score trajectories. Industry trend lines. The companies that improved and how long genuine improvement took. The companies that didn't and what the pattern looks like year over year. That historical data cannot be purchased by any competitor at any price. It has to be earned through time. The moat deepens on its own.

This is not a platform with a product cycle. It is a living organism that gets stronger every single day without being asked to. The founding vision is infrastructure — something that outlasts any individual product iteration, any regulatory shift, any change in the technology landscape. The data is the asset that makes everything else possible.

The Hall of Negative Fame. The Hall of Excellence. The Annual Reckoning.

This will not happen in year one. It should not. The annual accountability report — the Hall of Negative Fame and Hall of Excellence — requires something that cannot be rushed: a track record substantial enough that the findings carry institutional weight when published.

When cfva.ai has been publishing scores for three or four years — when journalists already cite the platform, when the methodology has been publicly documented, tested, and refined, when the database has depth and history — then the first annual report lands with the gravity the concept deserves. Not before.

The Hall of Negative Fame is not designed to humiliate. It is designed to make the worst documented patterns of institutional harm impossible to ignore for another calendar year. Companies that appear on that list despite having the resources and the information to have changed course — that appearance documents a choice. The list makes the choice visible and permanent.

The Hall of Excellence exists for the same reason the free monitoring tier exists: to create a genuine, credible positive incentive. A cfva.ai Hall of Excellence recognition means something precisely because it cannot be purchased. Companies that earn it will promote it — in investor materials, in procurement responses, in press releases. They become the platform's most credible advocates because their recognition is real and the methodology that produced it is fully public.

The patience is the strategy. The credibility has to be earned before the annual report can carry weight. That earning happens record by record, score by score, year by year. When the first annual report drops, the platform will have already earned the right to publish it.


How This Signal Was Built

This Signal began with a news story about a mass layoff. It ended with a complete scoring architecture, a legal framework, a tiered access model, a permanent integrity statement, and a long-term publishing strategy that will take years to fully execute and generations to fully measure.

The human brought the instinct that made every piece click into place: that negative scores should have no floor because reality is asymmetric and the score should reflect reality. That industry rankings are more damaging than raw scores because context is what transforms numbers into accountability. That the score must be permanently incorruptible — not as a marketing position but as a foundational design requirement that cannot be negotiated away at any price. That this platform should outlast its builder and serve people he will never meet.

The AI brought structure — the legal architecture, the access model logic, the competitive moat analysis, the framework that turns instinct into something buildable and defensible.

The human grew up watching what happens when information is controlled by the people with power. He built a system where the information belongs to the record, the record belongs to the public, and the score belongs to the truth. That lived experience is the reason this platform exists and the reason it will be built to last beyond any single product cycle or regulatory moment.

The scoring engine will be planned in full detail before a single line of code is written. Three days minimum. Legal review follows. The build follows the review. And then the shadow follows them everywhere — permanently, automatically, made entirely of their own actions reflected back through public record.


Eleven Signals. One Protocol. One Score.

The engine was born in Signal 007. The score crystallized in Signal 011. Planning begins now. The build follows the plan. The shadow follows the build.