Organisations are not all built the same way
Organisations differ in structure as well as capability. A formal board and a founder-led company answer very differently to the same governance question, and a matrix organisation coordinates work in a way a twenty-person team never needs to. Public sector bodies, multinational corporations, charities, family-owned businesses and start-ups organise themselves in different ways, so not every organisational mechanism exists in every setting. ORAVORN recognises that variation rather than treating it as an exception.
Absence is not always weakness
Suppose an organisation is asked how AI governance is overseen by its board. For many organisations, that is an appropriate and answerable question. For a founder-led business with no formal board, it is not. Selecting the lowest response would wrongly imply weak governance, when the issue is that the mechanism being assessed does not exist in that context. ORAVORN separates a capability gap from a structural feature of the organisation, rather than folding both into the same low score.
Two different reasons a question might not apply yet
It helps to be precise here. "Not applicable" and "hasn't started yet" are not the same, although both sit outside the usual Low-to-Strong scale.
Some conditions genuinely do not exist for a given organisation and are unlikely to appear in their current form: a founder-led business without a board, for example, or a company small enough that formal functional silos are not a meaningful feature. These are structural facts, not failures of development.
Other conditions have not been tested because the underlying activity has not yet begun. An organisation that has not deployed any AI systems has nothing to hold accountable or monitor. A leadership team where AI has rarely been discussed has no settled alignment to assess. That is different from a structural non-applicability. It is more provisional, and it says something useful about the organisation's current position in its AI journey, closer to early exploration than to an established weakness.
Both options serve the same purpose: they prevent a misleading score when the scale is being applied to a situation it was not designed to judge.
Why this improves accuracy
Consider two organisations. Organisation A has deployed several AI systems but has no named accountability for any of them. Organisation B has not deployed any AI systems. If both selected the lowest available response, the results would make them look the same. They are not. Organisation A has a real accountability gap. Organisation B has no deployed systems that require that form of accountability yet. Treating those cases as equivalent would distort the diagnosis for both.
A genuine "this does not apply yet" response preserves that distinction instead of losing it in the scoring.
"Not Applicable" is not a way to avoid a difficult question
This option should be selected only when the underlying mechanism genuinely does not exist, or when the relevant situation has not yet occurred. It is not appropriate simply because the answer feels uncertain, the information is difficult to find, the organisation hopes the condition will exist later, or people disagree about the honest response.
There is a specific trap worth naming. A real, low-scoring condition can feel more comfortable to describe as "not applicable" than as weak, especially where the truth is closer to "we have not built this yet" than "this genuinely does not apply to us." If AI adoption is not recognised anywhere in the performance management system, that is a real answer about the organisation's current state, not an exemption from the question. Where there is uncertainty about which situation applies, the more accurate choice is usually the plain answer that best reflects current reality, not the option that allows the question to be set aside.
Why only some questions include these options
Only certain questions offer a not-applicable or not-yet option. That is deliberate. Some organisational conditions are close to universal: every organisation has some form of communication, leadership and strategy, even where those conditions are weak. Other conditions depend on context or on activity that has not started. Board oversight cannot exist where there is no board, and operational monitoring cannot be meaningfully assessed where no AI system has been deployed to monitor. In those cases, forcing every organisation onto the same scale would make the assessment less accurate. ORAVORN includes these options only where their absence would weaken the diagnosis.
Fair assessment requires this flexibility
ORAVORN was built for organisations that vary widely in size, sector and structure. A single scale that ignored those differences would appear consistent while producing unreliable results for organisations that do not fit its assumptions. Used carefully, not-applicable and not-yet options help avoid that problem. They keep recommendations grounded in the organisation as it actually is, rather than in a generic model of how organisations ought to be built.