Applied research specification
Make sure you're ready before AI, not after.
Ready means you can say what each AI tool or agent may access, decide, and do, who is responsible, where a person must review, and what triggers a stop or another decision.
Rule 52 applies IVA's evidence, ownership, decision, and artifact discipline to 52 operating decisions organizations should make before workplace AI use becomes normal.
AI can now search, connect, generate, move data, call tools, and trigger work at machine speed. Governance has to address that operating reality.
0 to 52
A policy is not enough when tools and agents can act.
A general policy does not resolve every tool, agent, data type, workflow, external output, vendor, exception, or incident. Rule 52 turns those broad promises into specific operating decisions that can be assigned, reviewed, changed, or stopped.
- Decided and risk-proportionate
- Assigned to an owner and authority path
- Supported by evidence and an operating artifact
- Connected to exceptions, incidents, and review
- Maintained when tools, vendors, data, or conditions change
Eight operating domains
The complete readiness field.
Governance + scope
Approved tools, ownership, access, work purpose, experiments, and refresh cadence.
Data, privacy + access
Prohibited and conditional data, minimization, accounts, retention, permissions, and contracts.
Use-case approval
Low-risk uses, approval triggers, prohibited uses, people-affecting decisions, and formal submissions.
Human review + quality
Review triggers, reviewer qualification, facts, sources, bias, accessibility, safety, and responsibility.
Disclosure + IP
Internal and external disclosure, authorship, copyright inputs, output review, and public approval.
Vendors + third parties
Vendor terms, data use, security review, contractor rules, validation, and change review.
Security + incidents
Prompt injection, agent permissions, output handling, logging, incidents, and impersonation.
Training + enforcement
Role-sensitive literacy, acknowledgment, exceptions, remediation, consequences, and refresh.
Relationship to the thesis
Everything is context does not mean every tool receives unrestricted access.
IVA treats context as governed information linked to a decision, source, permitted use, and lifecycle. Rule 52 specifies the governance questions for which AI tools, accounts, roles, data classes, workflows, and outputs may access or use that context. Collection, retention, deletion, access, inference, disclosure, and action are separate governance decisions.
Evidence status
Designed operating structure, not a compliance or incident-reduction guarantee.
Rule 52 currently supports a designed-capability claim: it structures 52 AI-use decisions and the records needed to assign, evidence, review, and maintain them. The current claim concerns the decision structure and entry artifacts, not persistent operational maintenance. Claims that it reduces incidents, compliance cost, risk, or implementation time require repeated measured outcomes and appropriately qualified comparison.
The workflow can support NIST- or ISO-informed implementation and specialist review, but it does not establish legal compliance, NIST conformance, ISO certification, technical model assurance, or universal regulatory coverage.
Proposed applied workflow
Rule52 Continuous is reserved for a proposed maintained recurring workflow.
Rule52 Starter, Pro, and Consultant kits are current entry-artifact routes. Agency access is offered only after billing verification. These routes can generate and organize decision artifacts, but they do not by themselves maintain an organization's AI-use register or monitor continuing operations.
Rule52 Continuous is not in general availability. The name is reserved for a proposed recurring workflow that would maintain ownership, evidence, exceptions, incidents, review triggers, and renewal across AI uses. It is not a certification, compliance guarantee, automated monitoring claim, or currently operating software service.
Rule52 Continuous, Rule52 Operator, and Rule52 Sector remain application-only or design-partner concepts until the required publication, secure operations, support, ownership, record persistence, and evidence gates are satisfied.
Applied bridge
Start with the AI use or agent that needs a real answer.
Free Rule 52 management check
Check whether access, authority, human review, responsibility, exceptions, incidents, and stop conditions are written and owned.
Read this page $500 triageRule 52 AI Policy Triage
Review a bounded AI use, workflow, or policy and identify the access, action, ownership, human-review, evidence, exception, incident, and stop rules it needs.
Read this page Bounded setupRule 52 Implementation Sprint
Create practical operating rules, owners, review points, and records for a defined group of AI uses or agents.
Read this pageRecurring AI oversight
A maintained recurring workflow remains a gated design concept until its operations, security, support, and evidence requirements are real.