THE DEFINITIVE CODE
.png)
THE ICON BEHIND THE CODE
Encoding Judge Constance Baker Motley's legal rigor into a modern standard for defensible workplace investigations
The CONSTANCE Code™
The Definitive Code for Investigative Due Diligence
As the first Black woman to argue before the Supreme Court, Constance Baker Motley won nine of ten landmark civil rights cases through meticulous preparation, unwavering principle, and a commitment to due process that transformed American jurisprudence. Her legacy demonstrates that true equity requires both rigorous procedure and substantive fairness.
Today, The CONSTANCE Code™ applies these same principles to workplace investigations. We ensure that AI systems analyzing complaints, witness statements, and evidence maintain the procedural integrity and impartiality that Judge Motley championed. This transforms AI-assisted investigations from potential liabilities into benchmarks for defensible, transparent due diligence. The Code provides the comprehensive oversight and normative standards that dictate how investigations should be conducted, from vendor selection and ongoing evaluation to final resolution.
THE ALGORITHMIC THREAT
An investigative AI does not appear in a case file as a biased fact-finder. It appears as a credibility score. The witness who speaks in a register the model was not trained on is rated less believable, and the rating becomes the record.
The same is true for evidence weighting, witness ranking, and case summarization. When the logic is buried in a model, the harm is invisible. The witness has no explanation. The investigator has no record. The employer has no liability on the books.
But the liability is real. And the profession is not ready.
According to a May 2026 MyPerfectResume study, 73% of employers use AI in hiring decisions. 65% of those managers admit the software screens out applicants before a human sees them. The same systems are now being deployed inside workplace investigations. The AI generates the credibility score. The investigator signs the report. The report becomes the record. The record becomes the case.
The EEOC requires investigations to be "sufficiently thorough to arrive at a reasonably fair estimate of the truth." A standard investigation does not require a model version hash. It does not address system prompts, inference logs, or training data lineage. It leaves the investigator without any mechanism to audit human-in-the-loop overrides.
This is not a failure of individual investigators. It is a structural failure of the foundational investigative standards themselves. They were engineered for an era that no longer exists. The profession is navigating blind, through a process that has already been automated.
THE RECURSIVE TRAP
There is a deeper structural failure. Investigators increasingly rely on AI platforms to ingest evidence, flag inconsistencies, and generate summary reports. When an investigator deploys an algorithm to evaluate an allegation of algorithmic bias, the search for truth becomes recursive. It collapses into a hall of mirrors. An un-audited AI analyzing an un-audited AI.
The Failure Stack operates in five steps.
-
Biased HR AI scores the employee.
-
Manager rubber-stamps the output.
-
Investigator AI analyzes the complaint.
-
Investigator AI flags the case as low probability.
-
The final report omits human verification.
No one has independently evaluated the original algorithm. The loop is closed. The truth is lost.
The CONSTANCE Code breaks the loop. It mandates that any automated tool used by the investigator be fully disclosed, its training data and prompt parameters documented, and its analytical outputs independently verified by human review.
THE FOUR DIGITAL ARTIFACTS
Every investigation into an algorithmic decision requires four digital artifacts. None of them appear in legacy investigative checklists.
The Model Version Hash. The cryptographic identifier of the exact algorithm version deployed at the time of the adverse action. Without it, the investigator cannot prove that the model reviewed is the model that made the decision.
The System Prompt and Inference Log. The instructions the employer gave the algorithm, and the score it assigned to the employee at the moment of calculation. This is the investigative equivalent of a manager's hidden intent.
The Training Data Lineage. The provenance, composition, and known biases of the datasets used to train the model. If the training data was biased, the credibility scoring was biased.
The Human-in-the-Loop Record. The metadata proving a human actually reviewed the output, spent time on the file, and had the authority to override the machine. A rubber stamp is not oversight. It is performance art.
These four artifacts are the evidence the CONSTANCE Code demands in every investigation. Without them, the investigation is not thorough. It is a summary of a decision no one examined.
THE CONSTANCE CODE SERVICES
1. Investigative Algorithmic Audit
We audit the evidence weighting and credibility assessment algorithms used in workplace investigations. We verify that selection criteria are job-related, that the model does not default to highest-weight historical patterns, and that the training data is representative. When an investigative system is found to be discriminatory, we document the finding in a litigation-ready format.
Powered by DigitalRAS™.
2. Credibility Scoring Validation
We validate that credibility scoring systems are calibrated against procedural integrity standards. We document whether the scoring model penalizes protected characteristics, speech patterns, or communication registers. We verify that a human investigator's judgment overrides the model's score, not the reverse.
Powered by DigitalRAS™.
3. Human-in-the-Loop Verification
We validate that human review requirements for investigative findings are actually operational. The system should hold irreversible actions for human review while allowing parallel independent actions to proceed. We document whether the verification layer exists, whether it functions, and whether it was bypassed.
Powered by DigitalRAS™.
4. Evidence Provenance Audit
We document the provenance of the data used in the investigation. We verify every source at the point of collection, trace what the model encoded at the weight level, and produce a baseline record that satisfies the evidentiary standards of EEOC complaints and class-action courts.
Powered by COPERNICUS Canon™.
5. Recursive Trap Assessment
We audit the investigation firm's own AI tools. We determine whether the investigative platform introduces a second layer of unverified automation. We document the training data, system prompts, and validation testing behind any AI used to analyze evidence. We block the hall of mirrors.
Powered by DigitalRAS™.
6. Investigation Integrity Report
We produce a forensic report that identifies whether the investigation examined the algorithmic artifacts that shaped the adverse decision. If the artifacts were requested and refused, we document the refusal as evidence of obstruction. If they were never requested, we document the investigation as structurally incomplete.
Powered by DigitalRAS™.
7. The Motley Redress Initiative
For employees and advocacy organizations.
We provide a Probabilistic Harm Audit that uses econometric counterfactual baselines to isolate the Algorithmic Increment of Harm in a specific investigation. This quantifies the harm for EEOC complaints, state enforcement, and impact litigation.
Powered by The Right to Be Probable™.
THE SEVEN EMPLOYEE QUESTIONS
If an AI system touched your hiring, pay, termination, or investigation, you cannot wait for HR to volunteer the evidence. You have to demand it. In writing. On the record.
Here are seven questions to submit during your intake meeting.
1. What specific algorithmic tool or predictive ranking system was used to make the recommendation in my case?
2. Will this investigation review the system prompts, behavioral parameters, or weighted criteria the algorithm used to grade my performance or communication style?
3. Has the company issued a data preservation hold on the cryptographic model version hash and my individual inference logs so the evidence is not altered?
4. Will the investigation examine the training data lineage, including the sources, composition, date range, and known biases of the datasets, to determine whether the AI replicated systemic discrimination?
5. What specific metadata or audit logs prove a human manager meaningfully reviewed the AI output, spent time on my file, and had the authority to override the machine recommendation?
6. Will the investigator use any AI platform to analyze the evidence? If so, what training data, system prompts, and validation testing were used to build that tool?
7. If the company or its software vendor refuses to produce any of these artifacts, I request that the refusal and the specific reason given be documented in the final investigation report.
If HR answers, you get the evidence. If HR ignores you, you create a written record of an incomplete investigation. That record is exactly what a plaintiff's attorney needs.
THE CONSTANCE ECOSYSTEM
The CONSTANCE Code is the workplace investigation configuration of DigitalRAS. It is one sentry in a coordinated system of algorithmic accountability. Three related engines extend its reach across the employment lifecycle.
THE WALLACE INITIATIVE™
The Employment Sentry
Where CONSTANCE audits the investigations that determine who gets believed, WALLACE audits the decisions that determine who gets hired. Hiring, promotion, and compensation algorithms audited against Title VII and state anti-discrimination standards.
→ Explore The WALLACE Initiative™
LA DOCTRINA DE LUISA™
The Supply Chain Labor Sentry
Where CONSTANCE focuses on investigations, LUISA focuses on the workers whose hours, wages, quotas, and terminations are governed by algorithmic management systems. The PAGA litigation practice is the enforcement arm.
→ Explore La Doctrina de Luisa PAGA Practice
SILKWOOD SAFEGUARD™
The Whistleblower & Knowledge Workers Sentry
Where CONSTANCE investigates the complaint, SILKWOOD protects the person who filed it.
REGULATORY JUDO™
The Disclosure Request Engine
LAUNCHING SOON
Ghost jobs are the visible symptom. Regulatory Judo is the tool that exposes them. It generates legally compliant disclosure requests under NYC Local Law 144, FCRA, CCPA, and state pay transparency laws.
→ Join the Regulatory Judo Waitlist
WHO THIS SERVES
The CONSTANCE Code serves both sides of the investigation.
For Employers
Corporate Legal & Compliance Leaders. Chief Human Resource Officers. Workplace Investigators. Boards of Directors.
If you want your investigations to hold up when the report is tested, we build the protocol that mandates the four artifacts before any report is finalized.
For Employees & Advocates
Employees under investigation. Employment Law Attorneys & Plaintiffs' Firms. Worker Advocacy Organizations. State & Local Regulators.
If you were investigated by a system you were never allowed to see, you may have a claim. The record already exists. Let us read it.
THE EMPLOYMENT LIFECYCLE
Hire. Investigate. Employ. Disclose.
Four sentries. One lifecycle. One record.
THE DEFENSIBLE FOUNDATION
THE RECORD SPEAKS FOR ITSELF.
WE MAKE SURE IT IS HEARD.
Every interview. Every credibility score. Every finding. The algorithm generated a record of everything it did. Most of that record has never been examined. Most of it has never been challenged.
We examine it. We challenge it. We litigate it.