Executive slide: https://ambradan.github.io/IFX-vs-KQR/
Organizations will not be judged on whether they used AI. They will be judged on whether they knew what they were doing with its outputs.
In regulated environments, the absence of qualification is itself a risk.
KQR and IFX do not evaluate correctness or truthfulness of outputs. They evaluate whether outputs may be treated as admissible knowledge, under defined traceability and governance constraints.
Knowledge Qualification Regime (KQR) is a governance framework that defines when outputs produced in non-deterministic environments (e.g., LLM-generated content) can be treated as admissible knowledge, and under which constraints.
IFX (Inference Forensics Execution) is the forensic execution layer that implements KQR rules at runtime, ensuring traceability, auditability, and policy compliance.
KQR defines the rules.
IFX enforces the rules.
Neither produces AI outputs.
| Aspect | KQR | IFX |
|---|---|---|
| Nature | Governance regime | Enforcement layer |
| Function | Defines admissibility criteria | Applies criteria at runtime |
| Output | None (rules only) | SignedOutput, ledger events |
| Determinism | Deterministic (fixed rules) | Deterministic (rule application) |
| Audit | Conceptual framework | Complete forensic trail |
- A knowledge qualification and control layer
- A forensic enforcement infrastructure
- A governance primitive for non-deterministic outputs
- An AI product or service
- A prompt engineering framework
- A development methodology
- An automation or decision engine
- A truth-verification system
Non-Deterministic Sources Qualified Outputs
(AI, tools, human input) (traceable, auditable)
│ ▲
└──────────► IFX ───────────────┘
│
implements
│
KQR
(governance rules)
Consumers (applications, workflows, products) are explicitly out of scope. They consume IFX outputs but are not governed by this specification.
Risk containment: Does not promise truth. Does not promise automation. Promises control, audit, and explicit limits.
Regulatory fit: Designed for environments where errors are not acceptable — healthcare, public administration, legal, finance, enterprise AI governance.
Modularity: IFX can be adopted without exposing KQR internals. KQR can evolve without rewriting IFX. Consumers can change without affecting the regime.
| File | Purpose |
|---|---|
| WHY-KQR.md | Executive rationale |
| IFX-VS-KQR.md | Detailed comparison |
| TERMINOLOGY.md | Fixed glossary |
| DIAGRAM.md | Architecture explanation |
| BOUNDARIES.md | Scope and limitations |
IFX enforces the rules defined by the Knowledge Qualification Regime.
KQR defines what is admissible knowledge.
IFX ensures that this decision is traceable, auditable, and defensible.
IFX is infrastructure. KQR is governance.