1. Internal training becomes a problem when it creates a hybrid role

A non-independent model often works like this:

• The organization designs the method,
• trains the assessors,
• They teach them best practices.
• transmits the official interpretation to them.
• entrusts them with the support of the projects,
• then also entrusts them with the evaluation.

This combination creates a hybrid actor that:

• advises,
• optimizes,
• preliminary note,
• influence,
• then judge.

Structural consequence:

An accompanying evaluator cannot be impartial.

It's not a question of competence, but of governance.

2. In an accreditable model, training ≠ influencing if the roles remain separate

International standards, ISO 17065, 17020, 17029, 19011, impose a simple rule:

The organization can train the actors, but it must strictly separate their functions.

In a compliant system:

• Training for facilitators = production of evidence.
• The training of evaluators = application of the procedure,
• the two functions never intersect.

Training is only a problem when it leads the same people to support and evaluate the same project.

3. IRICE trains, but never mixes roles

At IRICE, training is functional, never hierarchical or doctrinal.

3.1. IRICE trains Biodiversity Partners (support)

Objective :

• prepare the evidence.
• document the choices,
• structure the file,
• to support the project owner.

They do not:

• They don't note this down.
• do not evaluate,
• do not decide.

→ Accompaniment only.

3.2. IRICE trains the Assessors (assessment)

Objective :

• apply the ISO procedure.
• verify the evidence.
• to verify compliance,
• to ensure reproducibility.

They do not:

• They do not recommend it.
• do not optimize,
• do not participate in the design.
• They do not prepare the file.

→ Evaluation only.

3.3. No actor is a hybrid

That is exactly what the standards require: no dual roles, no contamination of roles.

4. Why hybrid models are not eligible for accreditation

When the assessor:

• accompanied,
• optimizes,
• pre-evaluation,
• prepare the file,
• then evaluates the project.

SO :

• Impartiality is impossible.
• the decision is influenced,
• The evidence is not neutral.
• Governance is not independent.
• The assessment is not legally binding.
• The device is not accreditable.

This is the methodological breaking point that distinguishes institutional models from self-managed private models.

5. Internal training: compliant if and only if the functions are independent

The institutional requirements are based on three simple principles:

5.1. Functional Separation

Designer ≠ Facilitator ≠ Evaluator ≠ Decision-maker.

5.2. Operational independence

An evaluator must not have contributed to the project. A mentor must not have a decision-making role.

5.3. Lack of influence

Training should not guide decisions, but only the evaluation method.

IRICE checks these three boxes.

6. The IRICE architecture: controlled training, strictly separated roles, guaranteed impartiality

The IRICE model is structured as follows:

1. Design: Documented reference frameworks, not influenced by evaluators.
2. Biodiversity Partners provides support; they are trained in evidence production but have never been evaluators.
3. Assessment: Assessors trained in the procedure, but never as guides.
4. Decision taken by an independent body.
5. Governance Compliant with ISO 17065 accreditation requirements.

Result :

Internal training that is perfectly compatible with independence, because the roles never overlap.

Conclusion

Training actors is not an obstacle to independence. It is the mixing of roles—support, optimization, pre-evaluation, decision-making—that destroys impartiality and renders a system unaccreditable.

By strictly separating:

• Biodiversity Partners (support),
• The Assessors (assessment),
• the Decision (independent body),

IRICE implements an architecture that conforms to international standards, ensuring an impartial, enforceable and institutionally reliable assessment.