Trust & security

Controls you can inspect—not slogans

Quality and risk management are part of the product. We classify data sensitivity, assign reviewers, limit access, and exclude high-impact work until a separate assurance program exists.

Non-negotiable operating principles

  • Customers buy outcomes, not unreviewed trainee hours.
  • Practitioners are paid for productive commercial work under a lawful structure.
  • No graduate receives production access or a senior role based on certification alone.
  • Accumentum retains documented responsibility for scoping, delivery oversight, and QA.
  • A Performance Receipt reports what was actually verified, including limitations.
  • Client data remains private, segregated, and purpose limited.

Mission risk classification

Early offerings emphasize lower-sensitivity, document-heavy work. Expanding access always follows classification and approval—not convenience.

R0

Public / synthetic

Public source material or fully synthetic data; no client account access.

Controls: Standard NDA; QA rubric; ordinary access controls.

R1

Low sensitivity

Redacted internal documents; read-only process documentation.

Controls: Named reviewer; least privilege; scoped retention; customer approval.

R2

Controlled enterprise

Real corporate configuration or internal records; limited read-only access.

Controls: Security approval, MFA, approved workspace, additional monitoring, senior lead.

R3

Restricted / high impact

Production change, sensitive regulated data, privileged access, legal or safety consequence.

Controls: Excluded from early launch offerings. Separate program and security assurance required.

Launch exclusions

  • • 24×7 incident response / MDR as a standing service
  • • Emergency remediation and unrestricted admin access
  • • Production penetration testing as a default catalog offer
  • • Legal, financial, or medical certification opinions
  • • CUI, classified, export-controlled, PCI, or PHI work without a separate compliant environment
  • • Autonomous AI deployments or irreversible production changes
  • • Guarantees of assignment, employment, licensure, or regulatory compliance

Privacy & responsible AI

  • • MFA for privileged users; tenant isolation; encryption in transit and at rest.
  • • Practitioners must not paste client data into personal AI accounts.
  • • AI may assist intake or suggestions with human approval—it does not replace accountable review.
  • • No training on customer or practitioner PII without appropriate rights; no automated hiring decisions from a model alone.

Formal customer terms, DPAs, and security questionnaires are provided during commercial engagement—not as marketing claims on this page.

Need a security conversation before scoping?

Share high-level data classes and access needs. We will route to the right risk posture—or decline work we should not take yet.