ORAVORN is built on the SLGTC Framework, a model of organisational AI transformation across five domains: Strategy, Leadership, Governance, Transformation and Change Management. This document sets out the research grounding each domain and the logic that connects them, for readers who want to understand why the framework is built as it is.
The problem with treating domains as independent scores
Most AI readiness instruments trace their structure to capability maturity model traditions that originated in software process improvement during the 1980s and 1990s (Humphrey, 1989). These models assess an organisation against a fixed set of dimensions, score each independently, and combine the scores into an overall maturity level. The approach carries an assumption worth naming directly: that organisational capability accumulates additively, and that the same capability contributes the same value to readiness regardless of what else is present in the organisation.
Configurational theory in organisational research has long challenged this assumption, arguing that organisational attributes interact rather than sum, and that identical capabilities produce different outcomes depending on the configuration they sit within (Doty, Glick & Huber, 1993; Meyer, Tsui & Hinings, 1993). SLGTC applies this logic to AI transformation specifically. An organisation with strong AI governance and no defined strategic direction cannot meaningfully be described as partially ready simply because one strong score and one weak score average towards the middle. Its governance is protecting a direction that has not been chosen, a specific organisational condition that an overall readiness score would obscure.
Strategy
The Strategy domain draws on the dynamic capabilities tradition, which treats a firm's capacity to sense change, seize opportunity and reconfigure its resources accordingly as a defining strategic asset (Teece, Pisano & Shuen, 1997; Teece, 2007). It assesses whether an organisation can state, in terms specific enough to be tested, what it intends to use AI for, what competitive posture it is pursuing, and whether its ambition is matched by the resources available to deliver it. This final consideration reflects the resource-based view's core claim that strategy is constrained by the resources a firm actually holds (Barney, 1991; Hamel & Prahalad, 1989). An organisation can hold genuine strategic clarity and pursue an ambitious posture while being resourced for a far more modest one. That combination, clear direction paired with inadequate resourcing, constitutes a recognisable organisational condition that an average of the two scores would fail to represent.
Leadership
Upper echelons theory establishes that organisational outcomes reflect the background, values and cognitive frames of the people leading the organisation (Hambrick & Mason, 1984; Hambrick, 2007). The Leadership domain asks whether a top management team is aligned on AI's role, pace and acceptable risk, drawing on research distinguishing productive cognitive conflict from affective conflict, which consistently degrades decision quality (Amason, 1996). It also asks how a leadership team treats AI-generated input once it has it, since research on algorithmic decision-making shows people can swing towards dismissing a system after a single visible error or towards uncritical deference to outputs they cannot interrogate (Dietvorst, Simmons & Massey, 2015; Logg, Minson & Moore, 2019). A leadership team can be well aligned by conventional measures and still be collectively prone to one of these failure modes, requiring an intervention beyond alignment work.
Governance
The Governance domain's five principles, fairness, transparency, accountability, privacy and human oversight, are drawn from the OECD's Recommendation of the Council on Artificial Intelligence (OECD, 2019/2024) and the European Commission's Ethics Guidelines for Trustworthy AI (High-Level Expert Group on Artificial Intelligence, 2019), with its risk classification informed by the tiering structure of the EU Artificial Intelligence Act (Marcinek et al., 2024). Both instruments specify at length what good governance should contain. The question facing an individual organisation is how reliably those principles operate in practice. The domain also treats accountability diffusion, AI decisions with no clearly named accountable individual, as an extension of agency theory's analysis of the costs that arise when a principal delegates authority without adequate alignment of interests (Jensen & Meckling, 1976).
Transformation
The Transformation domain rests on the socio-technical systems tradition, which holds that technology introduced into a social system not designed to receive it will be distorted or rejected rather than adopted as intended (Trist & Bamforth, 1951). It treats data maturity as a precondition for AI deployment and approaches the capability pipeline as an organisational learning problem, since organisational capability accumulates through experience and codified routine more than through acquiring individual expertise from outside (Argote & Miron-Spektor, 2011). In most organisations, operating model readiness, data maturity and capability pipeline sit with three different functions and are rarely assessed together. Each may therefore be reported as adequate even when the resulting configuration remains unworkable.
Change Management
Edmondson's (1999) study of fifty-one teams found that psychological safety, a shared belief that a team is safe for interpersonal risk-taking, was associated with learning behaviour, which in turn mediated team performance. SLGTC treats psychological safety as a precondition for adoption rather than placing it alongside communication and adoption infrastructure as an equivalent lever, as change management practice typically does (Kotter, 1996, 2012; Rogers, 2003; Davis, 1989). The basis for this departure is specific to AI: employees more often experience it as a judgement on their own competence than they experienced earlier technology rollouts. Employees who do not feel safe raising concerns are unlikely to engage honestly with communication or training infrastructure, however well either is designed.
How the domains interact
The framework's contribution lies in the relationship it specifies between the five domains. Each domain has an established basis in existing literature, and a reader familiar with strategic management, leadership or change management scholarship will recognise the grounding in every case. SLGTC provides an explicit account of how weakness in one domain constrains or distorts the others, with some domains exerting greater influence on the configuration than others.
Strategy occupies a particular position among the five. Of logical necessity, no other domain can be properly configured without a defined strategic direction. Governance has nothing concrete to govern, leadership has nothing collective to align around, and change management has no stable narrative to communicate. This is a definitional claim about direction, which the framework reads as strategic clarity. It does not extend to competitive posture or to capacity.
The framework therefore distinguishes between prerequisite conditions and consequential conditions. Strategic direction functions as a prerequisite because it establishes the objectives governance must oversee, leadership must interpret and change efforts must communicate. The absence of strategy does not simply weaken these domains. It changes the purpose they are being asked to serve.
The remaining dependencies between domains are established with varying degrees of confidence. Some follow by extension from established literature, while others reflect patterns currently supported by industry evidence without yet having a comparable body of peer-reviewed research specific to AI transformation. The framework distinguishes between these types of claim because presenting both with equal certainty would give the weaker claims more evidential weight than they can currently support.
This asymmetric structure allows the framework to move from identifying weak domains to identifying which weak domain is actually binding the others. An organisation with comprehensive governance protecting a strategic direction it has not yet committed to presents a different problem from an organisation with the same governance score protecting a strategy it has clearly defined, even though a conventional scoring instrument would read both organisations as equally governance-mature. Reading the domains through their dependency structure enables the assessment to diagnose the organisational condition rather than merely describe the five measurements. It can therefore identify where an organisation is weak, what to address first and why addressing it first changes what becomes achievable elsewhere.
The same logic underlies the framework's typology of organisational archetypes. SLGTC constructs the typology according to the methodological criteria Doty and Glick (1994) set out for typologies to count as genuine theoretical contributions rather than classification devices: the theoretical assertions connecting the dimensions are made explicit, the same dimensions describe every archetype, and the resulting typology is built to be operationalised and tested empirically. ORAVORN identifies an organisation with comprehensive governance protecting an unresolved strategic direction as the Governed Skeptic, and an organisation whose strategic ambition and leadership conviction have outpaced the governance built to contain them as the Ambitious Disruptor. Each archetype carries its own primary risk and sequencing logic, derived from the dependency structure described above. The derivation of these archetypes from the framework is what gives the typology theoretical substance beyond the labels themselves.
SLGTC produces no general maturity tier. It identifies which condition is currently limiting an organisation's AI transformation, given everything else present in that organisation, and explains why addressing that condition first changes what becomes achievable elsewhere. An organisation's binding constraint can shift as external conditions change and as the organisation acts on what it learns. The framework is therefore built to be applied repeatedly, with each ORAVORN assessment providing a reading of the organisation at a particular moment rather than a fixed judgement of its capability.
References
Amason, A. C. (1996). Distinguishing the effects of functional and dysfunctional conflict on strategic decision making. Academy of Management Journal, 39(1), 123–148.
Argote, L., & Miron-Spektor, E. (2011). Organizational learning: From experience to knowledge. Organization Science, 22(5), 1123–1137.
Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340.
Dietvorst, B. J., Simmons, J. P., & Massey, C. (2015). Algorithm aversion. Journal of Experimental Psychology: General, 144(1), 114–126.
Doty, D. H., & Glick, W. H. (1994). Typologies as a unique form of theory building. Academy of Management Review, 19(2), 230–251.
Doty, D. H., Glick, W. H., & Huber, G. P. (1993). Fit, equifinality, and organizational effectiveness. Academy of Management Journal, 36(6), 1196–1250.
Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383.
Hambrick, D. C. (2007). Upper echelons theory: An update. Academy of Management Review, 32(2), 334–343.
Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reflection of its top managers. Academy of Management Review, 9(2), 193–206.
Hamel, G., & Prahalad, C. K. (1989). Strategic intent. Harvard Business Review, 67(3), 63–76.
High-Level Expert Group on Artificial Intelligence. (2019). Ethics guidelines for trustworthy AI. European Commission.
Humphrey, W. S. (1989). Managing the software process. Addison-Wesley.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm. Journal of Financial Economics, 3(4), 305–360.
Kotter, J. P. (1996). Leading change. Harvard Business School Press.
Kotter, J. P. (2012). Leading change: With a new preface by the author. Harvard Business Review Press.
Logg, J. M., Minson, J. A., & Moore, D. A. (2019). Algorithm appreciation. Organizational Behavior and Human Decision Processes, 151, 90–103.
Marcinek, K., Stanley, K. D., Smith, G., Cormarie, P., & Gunashekar, S. (2024). Risk-based AI regulation: A primer on the Artificial Intelligence Act of the European Union. RAND Corporation.
Meyer, A. D., Tsui, A. S., & Hinings, C. R. (1993). Configurational approaches to organizational analysis. Academy of Management Journal, 36(6), 1175–1195.
OECD. (2019, revised 2024). Recommendation of the Council on Artificial Intelligence. OECD/LEGAL/0449.
Teece, D. J. (2007). Explicating dynamic capabilities. Strategic Management Journal, 28(13), 1319–1350.
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533.
Trist, E. L., & Bamforth, K. W. (1951). Some social and psychological consequences of the longwall method of coal-getting. Human Relations, 4(1), 3–38.