The Evaluation Process

The model is real.
The results are reproducible.

ENHANCE operates a proprietary multi-stage structural extraction and kernel evaluation architecture. Every engagement follows the same protocol. The output is a verdict — derived from document evidence, not analyst judgment.

Evaluation Architecture

Four stages.
One output.

01
Document Submission
You submit the project document — a GCF Concept Note, funding proposal, corporate transition plan, or equivalent. ENHANCE receives it. Nothing else moves until the scope and engagement terms are confirmed.

All submitted materials are treated as confidential by default. They are not disclosed to third parties, referenced in public research, or used to train any AI model. Where institutional requirements demand it, a mutual NDA can be executed before any document is submitted.
02
3AI Structural Extraction (SCI Protocol)
Three independent AI extractors run simultaneously under ENHANCE's Structured Inference Mode (SIM) protocol. Each extractor operates in isolation — without access to the outputs of the other two.

Extraction is governed by a controlled protocol that prohibits inference beyond document evidence, narrative judgment, and domain-knowledge substitution. Verbatim citation is required for all extracted values.

After extraction, the Orchestrator applies Structured Convergence Inference (SCI): three-of-three agreement is required for a confirmed result. Two-of-three agreement is recorded as split. Divergence across all three extractors triggers a flag — not a guess.
03
Kernel Evaluation
Converged extraction outputs are passed to the ENHANCE evaluation kernel. Each of the four frameworks evaluates a different structural dimension of the project. Frameworks operate independently — the output of one does not determine the input of another.

The kernel produces numerical scores, tier classifications, and gate determinations based on the converged extraction. No narrative interpretation is applied at the kernel stage. The document is the evidence. The score is the result.
04
Diagnostic Output and Engagement Conclusion
The evaluation produces a structured diagnostic report covering: structural scores and tier classifications per framework; gate status and verdict conditions; identified structural conditions and gaps; evidence coverage and unknown field inventory.

The diagnostic is delivered exclusively to the instructing party. ENHANCE does not guarantee approval outcomes, investment performance, or regulatory acceptance. The diagnostic is an independent analytical verdict — not an approval recommendation.
The Four Frameworks

Each dimension.
Independently verified.

VERA-TCV™ is not a single score. It is four structurally distinct evaluations — each targeting a different failure mode in climate project architecture.

VERA™ v2.2
Viability & Intent Fidelity
Evaluates three dimensions: Reality Anchor (is the target grounded in reality?), Causal Integrity (does the design embed the goal?), and Intent Fidelity (is the motivation uncontaminated?). Nine weighted indicators across RA, CI, and IF axes.
ToC:f(x)™
Structural Coherence
20 binary structural conditions evaluated under three question groups: Explicit Presence (Group A), Structural Demonstration (Group B), and Set-Theoretic Causal Evaluation (Group C). Score 0–1. Tier I–IV. KES Gate determines whether conditions block the verdict.
C-FAIR™ v3.0
Deployment Environment & Finance
Evaluates the environment into which the project is deployed — not the project itself. Climate hazard exposure, adaptive capacity, macro-financial fragility. VaR-based grant allocation model. Risk Class A–E. Finance Readiness grade.
VORTA™ v2.5
Impact Value Quantification
Translates ecological, social, and economic impact into investment language. VBI (Value of Benefit to Investment), PSC (Participant Structural Contribution), IRRᵥ and NPVᵥ. Supports private capital mobilisation and benefit-sharing design.
Data Handling

Your documents.
Our analysis. No further movement.

Confidentiality

All submitted documents are treated as confidential by default. Not disclosed to third parties. Not referenced in public research. Not used outside the scope of the engagement.

ENHANCE is willing to execute a mutual NDA before engagement where formal confidentiality protection is required.

AI & Model Training

Submitted documents are not used to train, fine-tune, calibrate, or improve any machine learning model, language model, AI system, or external platform.

AI-assisted extraction within ENHANCE's evaluation workflow is used solely for the commissioned engagement. It does not create any persistent training dataset.

Independence

ENHANCE operates without a financial stake in the outcomes of the projects it assesses. That independence is not incidental — it is the foundation of analytical authority.

No Approval Guarantee

ENHANCE does not guarantee project approval, investment returns, or regulatory acceptance. Engagement outputs are structural diagnostic verdicts, not approval recommendations or investment advice.

Full Privacy Policy → enhance-institute.org

Ready to verify
whether the structure holds?

All engagements begin with a conversation. Reach us directly.

Request a Diagnostic View Service Tracks →