The rules governing professional excellence have quietly shifted beneath our feet, and most organisations haven’t noticed yet. Competence, once a relatively stable benchmark defined by credentials, technical knowledge, and measurable outputs. Has become something far more dynamic and contextually dependent in today’s rapidly evolving landscape.
Despite significant changes in the nature of work, most companies still rely on traditional frameworks built for a different era.. Leaders designed them for a world where roles were predictable. Knowledge stayed relevant longer, and people made decisions without the constant influence of intelligent systems. That world no longer exists.
In this analysis, we examine what competence truly means in today’s environment. Explain why legacy models cannot adequately capture it, and identify the requirements of a more accurate framework. Whether you are leading talent strategy, building organisational capability. Or simply trying to understand your own professional standing, the assumptions you hold about competence deserve serious scrutiny. What follows is not a surface-level critique. It is a rigorous examination of why the gap between how we measure capability and what capability actually demands has never been wider. And what that means for the decisions you make today.
What Competence Has Traditionally Meant
The modern understanding of competence did not emerge organically from organisational practice. Researchers and practitioners deliberately constructed the competence model, grounding it in a specific intellectual moment. In 1973, Harvard psychologist David McClelland published his landmark paper, Testing for Competence Rather Than for Intelligence. Arguing that traditional aptitude tests and academic credentials poorly predicted actual job performance and systematically disadvantaged underrepresented groups. He offered a pointed alternative: underlying personal characteristics, which he termed competencies, predicted outstanding performance. He defined these competencies as clusters of behaviours that enable superior job performance. That single conceptual move set the trajectory for over fifty years of HR and L&D practice.
From Theory to Organisational Tooling
Richard Boyatzis extended McClelland’s framework in 1982. And through the 1980s competency thinking migrated rapidly from academic psychology into mainstream talent management. Organisations began building structured competency models that translated strategies, goals, and role expectations into specific, observable behaviours. These frameworks served a clear practical purpose. HR and L&D teams used them to define hiring criteria, structure performance appraisals, design development programmes, and make promotion decisions. They gave organisations a shared language where none had existed. The model offered something genuinely valuable. Consistency, transparency, and a defensible basis for talent decisions that had previously depended on manager intuition alone.
McClelland’s Iceberg Model captured the architecture of competence neatly. Visible competencies, such as skills and knowledge, sit above the waterline. Below it lie the harder-to-assess attributes: motives, traits, and self-concept. McClelland’s theory of competencies at work operationalised this by distinguishing what a person knows from what they are driven to do. This distinction mattered enormously for selection and development design.
The Individual as the Unit of Analysis
One assumption ran through every major competency framework without exception. The individual was always the primary unit of analysis. Frameworks assessed what a person knows, does, and visibly demonstrates. Nothing in the traditional framing asked what a team dynamic enables. What organisational culture constrains, or what systemic factors amplify or suppress individual capability. This was not an oversight; it reflected a coherent philosophical position. Competence was a personal attribute. Development was a personal responsibility. Performance was a personal outcome.
For many years, this approach appeared sufficient. A competency framework developed in 2005 could often describe relevant behaviours in 2015 with only minor revisions. As a result, the lag between environmental change and framework updates remained largely tolerable. However, the model’s limitations were present from the outset.
In fact, Spencer and Spencer warned as early as 1993 that generic competency dictionaries could never achieve true precision. By seeking universal applicability, they inevitably sacrificed role-specific accuracy. Consequently, organisations often overlooked capabilities that differentiated performance in particular contexts. Spencer and Spencer estimated that role-specific competencies could account for more than 20% of a job’s actual requirements, yet standard frameworks routinely excluded them.
At the same time, traditional task-based job analysis lacked the flexibility required to capture success as work became increasingly dynamic, complex, and context dependent. In other words, organisations built the model for predictability rather than adaptation. As long as the operating environment evolved gradually, this limitation remained manageable. However, once the pace of change accelerated, the cracks began to appear. Ultimately, the model starts to fracture when the operating environment evolves faster than any framework committee can reconvene, review, and respond.
The AI Paradox: More Assistance Does Not Mean More Competence
The investment numbers are striking. Organisations using AI for talent development report 40% faster skill acquisition rates. And 81% of professionals believe AI-enhanced learning reduces time to leadership competency by at least 25%. These figures drive boardroom enthusiasm and L&D budget allocations. But they contain a dangerous assumption: that speed of acquisition is equivalent to depth of competence. It is not, and conflating the two may be one of the most consequential errors organisations make in this decade.
When Output Quality Outpaces Actual Capability
AI fundamentally changes the relationship between what a person produces and what they actually understand. When AI handles execution tasks, visible output quality improves regardless of whether the human directing it genuinely understands the work. A leader can use AI to synthesise market analyses. Draft strategic communications, and structure decision frameworks while producing polished and credible outputs.
Yet their underlying judgement, domain knowledge, and critical reasoning may remain entirely underdeveloped. The gap between assisted performance and actual capability does not show up in deliverables. It shows up when the AI is absent, or when a decision requires genuine independent reasoning under pressure.
The medical evidence is sobering. The ACCEPT trial, which studied AI-assisted colonoscopy procedures, found that after doctors began using AI support on alternating cases. Their adenoma detection rate on non-AI cases dropped from 28.4% to 22.4%. Patients were approximately 21% less likely to have dangerous polyps detected when AI was not present, compared to the pre-AI baseline. Output quality on assisted cases remained high. The human capability behind it had measurably eroded. Leadership development organisations should sit with that finding carefully.
The Competence Illusion and Its Cognitive Roots
This is the competence illusion in operational terms. An MIT study compared participants writing essays using AI assistance versus other methods. Using EEG scans, researchers found that AI tool users showed reduced neural connectivity, particularly in networks associated with memory and creativity. Memory retention dropped immediately after task completion. The outputs were strong. The internal understanding was not. This is not a productivity story. It is a capability degradation story disguised as a productivity story.
The calculator analogy offers partial illumination. Students who over-relied on calculators lost the cognitive scaffolding that makes higher-level mathematics accessible. They could no longer detect obviously wrong answers because they had not internalised what reasonable numbers look like. But as research on AI’s cognitive impact makes clear, AI goes considerably further. It does not merely compute; it reasons, drafts, synthesises, and advises. The cognitive offloading is qualitatively deeper, and the resulting illusion is far harder to detect from outputs alone. A miscalculation is visible. A deficit in leadership judgement, masked by consistently fluent AI-generated communications, is not.
The Organisational Measurement Problem
Organisations now face a diagnostic problem that most development programmes cannot solve. If AI consistently elevates output quality, then output-based measures of competence become unreliable. Promotion decisions, performance ratings, and development assessments that rely on deliverable quality will systematically overestimate genuine capability in an AI-assisted environment. The two variables, output quality and actual competence, can diverge significantly. And the gap between them is precisely what conventional measurement misses. Organisations need to ask a harder question: are their development programmes measuring genuine capability growth. Or are they measuring the quality of AI-amplified outputs and attributing that quality to the human?
The Redefinition: Human Competence in an AI-Augmented World
The 2026 Work Trend Index reframes the entire competence conversation with a single, clarifying insight: AI is not reducing the cognitive demands placed on humans; it is elevating them. As AI agents absorb execution tasks at scale, the human role shifts decisively toward directing, judging, and owning outcomes. This is not a minor recalibration. It represents a structural change in what it means to be capable at work. Raising the competence bar precisely at the moment when many organisations assume technology is lowering it.
Three Categories of Competence That Now Matter
Understanding this shift requires moving beyond a monolithic definition of competence toward three distinct and interdependent categories. Technical competence means understanding how AI tools function, where they produce unreliable outputs, and how to interrogate their reasoning rather than simply accept it. Leadership competence means directing AI-assisted work with clarity, contextualising outputs against organisational and ethical realities, and taking full accountability for decisions that AI informed but did not make. Human-AI collaborative competence is perhaps the most novel of the three; it is the capacity to work alongside AI systems as a genuine thinking partner, contributing the judgement, framing, and critical interrogation that AI cannot generate independently. The Business-Higher Education Forum’s AI-Enabled Professional Framework, developed with over 100 cross-industry leaders in 2026, validates this tripartite structure, explicitly designing for both human and technical competencies to use AI “thoughtfully, responsibly, and effectively.”
The Framework Gap Organisations Cannot Ignore
The urgency here is institutional, not merely conceptual. Sixty-eight percent of executives believe AI will reshape leadership competencies, yet organisations still use frameworks that designers created before AI became operationally central to knowledge work. The World Economic Forum projects that 70% of the skills used in most jobs will change by 2030, with approximately 60% of the current workforce requiring significant reskilling. Organisations are, in effect, measuring performance against a standard designed for a world that no longer exists. Traditional workforce planning cycles cannot keep pace with AI’s rate of evolution, and the 2026 Work Trend Index data reinforces this with a striking structural finding: organisational factors, including culture, manager support, and talent practices, account for more than twice the AI impact of individual effort alone, 67% versus 32%. People, in many cases, are ready. The systems around them are not.
What AI Cannot Replicate
The competencies that carry the greatest strategic weight in this environment are precisely those AI handles least well. Ethical reasoning requires holding competing values in tension and making defensible choices under genuine uncertainty. Contextual judgement means reading a situation through accumulated experience, relational knowledge, and organisational history that no model can fully encode. Relational intelligence involves building trust, reading unspoken dynamics, and sustaining commitment through complexity. The ability to ask better questions, rather than simply process answers, is perhaps the most undervalued capability of all; it determines the quality of everything AI produces in response. These are not soft skills relegated to the margins of performance frameworks. They are the core competencies of the AI-augmented era.
Redefining competence is, ultimately, a leadership culture decision. It changes what organisations hire for, what they fund in development, how they design performance measurement, and what behaviours they visibly reward. Without that redefinition, organisations risk building sophisticated AI capability on a competence foundation that was never designed to support it.
Competency Theatre vs. Real Competence
Most organisations adopt competency frameworks not because the evidence shows they predict performance, but because adoption has become an industry norm. The frameworks proliferate through imitation and audit logic: competitors use them, regulators recognise them, and they produce documentation that signals developmental rigour. Yet the academic literature reveals a structural flaw at the heart of this approach. Research validating competency frameworks typically measures whether large samples of practitioners consider specific competencies important, not whether competency scores actually predict job performance. A landmark study surveying over 25,000 participants across 30 countries confirmed the perceived importance of graduate competencies at scale, but importance ratings and predictive validity are fundamentally different claims. Organisations have largely proceeded as though the distinction does not matter.
The Theatre of Assessment
This is where competency theatre takes hold. The term describes the organisational tendency to treat the completion of a framework assessment, a development programme, or an annual review cycle as evidence of competence itself. What these processes actually evidence is more limited: the ability to perform competence within a structured assessment context. Individuals learn, often without conscious intent, to demonstrate the behaviours that assessors expect within the window of evaluation. The assessment context becomes the performance, not the job context. Selection and occupational psychologists have long recognised this as a validity threat. Construct validity erodes when researchers measure behaviour under conditions that differ significantly from the conditions practitioners encounter in real-world performance. Organisations that use competency frameworks as HR review instruments rather than as live performance tools expose themselves to precisely this problem.
Capability management systems are increasingly designed to capture contextualised evidence rather than completion records, reflecting a market acknowledgement that audit-readiness and performance insight are not the same thing. Yet the default organisational position still leans toward the former. Completion rates remain the dominant proxy because they are measurable, comparable, and defensible to senior stakeholders. EF Corporate Learning’s 2026 L&D Trends report identifies this directly, noting that the sector is shifting from learning access and usage metrics toward demonstrated skill outcomes. Completion rates, the report implies, are no longer acceptable as evidence that capability has been built.
From Attendance to Demonstrated Outcome
The shift from asking “did they attend the programme” to “can they demonstrate the outcome in a real context” is methodologically sound and organisationally uncomfortable in equal measure. It requires leaders and HR functions to accept that significant investment in development activity may not have produced the competence it claimed to produce. That is a politically difficult conclusion. It challenges prior decisions, surfaces accountability questions, and disrupts the reassurance that structured frameworks provide. The strategic case for competency frameworks rests on their capacity to raise performance across the organisation, but this claim is typically asserted rather than measured against outcome data.
Closing the gap requires a deliberate reorientation. Leaders and HR professionals need to move measurement toward observable, contextualised performance evidence gathered from real work, not assessment performance. This requires organisations to identify the specific outcomes a development intervention should produce, define how those outcomes appear in practice, and create feedback mechanisms that capture observed behaviour and drive action. Frameworks remain useful as organising structures. They become theatre only when organisations mistake the map for the territory, and treat evidence of framework engagement as evidence of genuine capability.
Competence as an Organisational Property
The argument that competence resides primarily in individuals is not just incomplete. It is strategically misleading. The 2026 Work Trend Index, drawing on surveys of 20,000 workers across 10 countries and analysis of trillions of productivity signals, establishes that organisational conditions, specifically culture, manager behaviour, and talent practices, drive measurably greater performance outcomes than individual effort alone. The directional implication is significant: however carefully an organisation develops individual capability, that capability is systematically under-deployed when the surrounding conditions are poor. Competence does not simply live inside people. It is activated, constrained, or compounded by the environment in which people work.
Frontier Firms and the Learning System Model
Microsoft’s research identifies a category of organisations it calls Frontier Firms. These are not simply organisations that have adopted AI aggressively. They are organisations that have redesigned themselves as learning systems, where capability compounds over time rather than accumulating in isolated individuals. Frontier Firms are engineering the conditions under which competence transfers more easily across teams, accelerates at the organisational level, and compounds rather than stagnates. This represents a fundamentally different operating model from the competency framework approach, which treats capability development as an individual transaction between a person and a behavioural standard. The competitive implication is direct: organisations that treat competence as a system property will outpace those that treat it as a personal attribute.
Deloitte’s 2026 Human Capital Trends research arrives at the same conclusion from a different angle. Deloitte’s research distinguishes between additive human-machine models, which produce incremental improvement, and multiplicative models, which produce compounding value. The same logic applies to capability development. Moving from developing individuals in isolation to building systems where human and organisational capability interact and reinforce each other is now the dominant direction in high-performing organisations. Senior leaders in 2026 are simultaneously concerned about AI-driven productivity and the adequacy of their talent pipelines; Deloitte’s framing suggests these are not separate problems but symptoms of the same structural gap.
Culture as Multiplier or Suppressor
Culture functions as either a competence multiplier or a competence suppressor. A genuinely capable individual operating in a psychologically safe environment, with strong manager support and access to information, will perform at a higher level than the same individual in a culture that penalises experimentation or withholds knowledge. Research by Amy Edmondson on psychological safety consistently demonstrates that team conditions, not just individual ability, determine whether people apply their full competence. Microsoft’s finding that most organisations are failing to keep pace with employee potential, despite that potential being at an all-time high, is entirely consistent with this: culture and management practice are acting as suppressors of capability that already exists.
The leadership agenda that follows from this analysis is a demanding one. Building organisational competence requires leaders to attend to the conditions of performance, not simply the capabilities of individuals. This means designing for knowledge transfer, constructing feedback environments that accelerate learning, and treating manager behaviour as a capability lever rather than a secondary concern. Most competency frameworks do not support this agenda. They focus on the individual unit of analysis and leave the systemic conditions largely unaddressed. Leaders who want to build genuine organisational competence must be willing to operate at a level of analysis that most conventional frameworks were never designed to reach.
Why Competency Frameworks Are Struggling to Keep Pace
Most competency frameworks operate on update cycles of three to five years. That rhythm made reasonable sense in a relatively stable operating environment. It no longer does. The capability landscape for knowledge workers is now shifting on a shorter cycle, driven by successive waves of AI development that are reshaping what skilled work actually involves at the task level. A framework reviewed in 2021 almost certainly predates the generative AI inflection point entirely. One reviewed in 2023 may already be misaligned with the agentic AI capabilities now entering enterprise workflows. The gap between when frameworks are written and when the work they describe has already changed is no longer a minor calibration issue. It is a structural problem.
The Vulnerability of Technical Skill Inventories
Frameworks built around role-specific execution tasks face particular pressure. AI is actively absorbing many of the technical capabilities these frameworks were designed to assess: data interpretation, report drafting, process documentation, structured analysis. The competencies that once marked a proficient analyst or mid-level manager are being compressed or displaced at the task level, yet most frameworks have no mechanism to retire or reweight those competencies in response to real performance evidence. They continue to describe a job architecture that the operating environment has already started to dismantle.
The DDI 2026 Leadership Trends report makes this exposure explicit. It calls for organisations to redesign leadership competency models for an AI-augmented environment, which is an implicit acknowledgement that current models were not built with this context in mind. DDI specifically identifies a distinct cluster of AI-era leadership capabilities that simply do not appear in pre-2023 frameworks, covering how leaders direct AI systems, sustain trust through change, and exercise judgment where automated outputs fall short.
The Legitimacy Problem
The credibility gap is also visible from the people frameworks are supposed to serve. 78% of professionals now prefer AI-personalised learning paths over standardised leadership programmes. That figure is not just a preference for convenience. It signals that the one-size-fits-all architecture underlying most competency frameworks is losing legitimacy with the very population it is designed to develop. When the majority of professionals actively prefer a different development model, the structural assumptions of the existing framework deserve scrutiny.
The answer is not to discard frameworks wholesale. They still provide a common language for performance conversations and a scaffold for development planning. The more productive response is to treat them as living instruments, subject to regular stress-testing against actual performance outcomes and revised whenever the evidence demands it. That requires a different governance posture than most organisations currently apply, one where framework maintenance is treated as an ongoing analytical function rather than a periodic administrative event.
What This Means for Leaders and Organisations
The analysis in previous sections points toward an unavoidable practical conclusion: knowing that competence frameworks are outdated is not the same as doing something about it. Leaders need to act, and the starting point is a two-question audit of every framework currently in use. First, was it built for the environment that now exists, one defined by AI-augmented decision-making, elevated stakes, and accelerating role change? Second, is there actual evidence that it predicts performance, rather than simply describing observable behaviours that someone once decided were desirable? Most frameworks will fail both tests. That failure is not a minor administrative problem. It is a strategic liability that compounds every time the organisation uses that framework to hire, develop, or evaluate its people.
The Investment Problem
87% of organisations plan to increase AI investment in learning and development, and the commercial logic is understandable. AI-powered tools offer speed, personalisation, and scale that traditional programmes cannot match. The risk lies in assuming that faster delivery of the same content closes the competence gap. It does not. If the underlying definition of competence being measured remains rooted in pre-AI behavioural descriptors, the organisation accelerates its ability to measure the wrong things more efficiently. Investment without definitional reform produces more sophisticated competency theatre, not more capable people. The tools change. The problem stays.
From Programmes to Performance Conditions
Organisations that want to build genuine competence, rather than produce evidence of it, need to redirect attention away from programme design and toward performance conditions. Culture, manager behaviours, and structural supports determine whether learning translates into actual capability on the job. Research consistently shows that 88% of change initiatives fail to meet their objectives, and the primary cause is not poor strategy but the neglect of the human and environmental conditions that make change real. Line managers are particularly critical here. A manager who models judgment under uncertainty, creates space for reasoned dissent, and connects learning to real decisions does more for team competence than any structured programme alone. Organisations that treat performance conditions as secondary to curriculum design will continue to confuse completion rates with capability.
The Right Strategic Question
72% of HR leaders believe AI will be critical for leadership development within three years. The leverage, however, is not in using AI to deliver faster versions of existing content. It is in using AI diagnostically, to surface where genuine competence gaps exist versus where the organisation has simply been measuring the wrong things for a long time. This reframing also sharpens the most strategically valuable question a leader can ask. The question is not “are our people competent?” That question anchors assessment to a static, inherited standard. The right question is whether people are competent enough to direct, judge, and own outcomes in an AI-augmented environment where the complexity and consequences of those outcomes have materially increased. Answering that question honestly requires better frameworks, better performance conditions, and a willingness to treat competence as something that must be continuously earned, not periodically certified.
The Question Organisations Should Be Asking
The real question is not whether organisations are developing competence. Almost all of them believe they are. The harder question is whether what they measure and reward actually constitutes it. Sixty-eight percent of executives believe AI will reshape leadership competencies by 2025, yet most organisations continue running assessment and reward systems built for a fundamentally different operating environment. That gap is not a minor calibration problem. It is a strategic exposure.
Competence has always been contextual. What made a leader capable in one era has never automatically transferred to the next. But the contextual shift now underway is not incremental. Frameworks built for a pre-AI operating environment are benchmarking against a version of capable that the work itself has already outpaced. Measuring task completion, behavioural indicators, and structured learning outcomes tells organisations very little about whether their leaders can direct AI agents, exercise sound judgement under uncertainty, or build the organisational conditions that multiply capability across teams.
Leaders who take this seriously will do four things. They will challenge the frameworks their organisations use, not assume that currency equates to validity. They will interrogate how competence is actually measured, and whether those measurements track capability or activity. They will invest in the organisational conditions, culture, manager quality, talent practices, that research shows account for twice the performance impact of individual effort alone. And they will resist the institutional comfort of mistaking a completed training programme for a developed leader.
The competence bar has moved. It will keep moving. The organisations that acknowledge this honestly, and act on it with the same rigour they apply to financial or operational performance, are the ones that will build the leadership depth the next decade demands. Belief in one’s own development programme is not the same as evidence that it works.
Leave a Reply