Skills Can No Longer Be Assumed

Every organization claims to value skills. Far fewer can demonstrate, with discipline and precision, which skills they actually have, which ones are missing, which ones are becoming obsolete, and which capabilities will matter most for the strategy ahead.

That gap between assumption and evidence has become one of the defining weaknesses of modern workforce management. In markets being reshaped by artificial intelligence, automation, data, and new operating models, organizations can no longer rely on job titles, resumes, manager impressions, or self reported confidence as sufficient proof of capability. The executive question is no longer simply, “Who do we employ?” It is, “What can our people demonstrably do, at what level, with what evidence, and for what future?”

A recent MIT Sloan article, drawing on research from the MIT Center for Information Systems Research, frames this challenge through the use of AI to identify and close workforce skills gaps. The article notes that, in a 2022 survey, executives estimated that 38% of their workers would need fundamental retraining or replacement within three years to address skills gaps. It also highlights a broader reality: AI and data skills are in high demand, talent is difficult to attract and retain, and training has become a strategic path to performance, retention, customer experience, and innovation.

The New Frontier of Skills Assessment

The most important concept in the MIT Sloan piece is skills inference. In practical terms, this means using AI to analyze employee data and estimate proficiency in specific skills. The objective is not to replace human judgment with algorithmic authority. It is to create a more structured evidence layer that can support development, internal mobility, and strategic workforce planning.

That distinction matters. Assessing competence is not the same as asking people what they believe they can do. It is also not the same as relying entirely on managers, whose judgments may be shaped by visibility, bias, proximity, outdated expectations, or incomplete information. Mature assessment requires taxonomy, evidence, criteria, validation, and governance. AI can expand scale and speed, but the quality of the result depends on the quality of the methodology behind it.

The Johnson & Johnson example discussed by MIT Sloan illustrates this point. The company began an initiative in 2020 with approximately 4,000 technologists before expanding it to other business units. Its process included three core steps: defining a future oriented skills taxonomy, selecting sources of skills evidence, and using an assessment model to estimate proficiency. The company identified 41 future ready skills, grouped into 11 capabilities.

The sequence is instructive. The organization did not begin with the tool. It began with the question of which capabilities would matter for the future. That is the difference between a superficial technology deployment and a serious assessment discipline. Before measuring, leaders must know what deserves to be measured. Before applying AI, they must define a competence model that reflects the organization’s strategic ambition.

The Risk of Measuring Poorly at Scale

AI can make a flawed assessment faster, broader, and more convincing. That is precisely why governance matters. When an organization lacks clear criteria, reliable data, ethical boundaries, and a well defined purpose, technology does not solve the problem. It automates confusion.

Skills inference should therefore be treated as a governance matter, not merely as an HR experiment or technology pilot. Data quality, model explainability, employee privacy, intended use, and the ability to challenge or contextualize results are all essential. In the Johnson & Johnson case, leadership made clear that skills insights would not be used in performance reviews, that information would be deidentified, and that insights would be used at an aggregate level to support strategic workforce planning. Employees were also allowed to opt out.

This distinction is decisive. When professionals believe a skills system will be used for punishment, opaque ranking, or unilateral career decisions, trust erodes quickly. When they understand that the purpose is development, mobility, and growth, engagement becomes more plausible. Competence assessment is not only a technical process. It is also a trust architecture.

The MIT Sloan article also notes that Johnson & Johnson compared AI inferred scores with employee self assessments on the same 0 to 5 scale. If the two scores differed by less than one point, the company considered the scores usable. This is a valuable methodological lesson: validity rarely comes from a single signal. It emerges from convergence across multiple forms of evidence.

From Individual Gaps to Strategic Capability

The value of a robust skills assessment system is that it connects two levels that are often managed separately: the professional’s development path and the organization’s strategic capacity.

For individuals, a sound assessment can clarify current strengths, development priorities, and possible career moves. For the organization, it can reveal concentrations of expertise, succession risks, regional weaknesses, investment needs, and internal mobility opportunities. According to MIT Sloan, after the first round of skills inference, use of Johnson & Johnson’s professional development ecosystem increased by 20%. As of March 2024, 90% of the company’s technologists had accessed the learning platform.

That result points to a broader principle: people are more likely to engage in development when they receive useful, personalized, and actionable insight. Generic training rarely changes performance at scale. Evidence based development has greater force because it connects learning, professional ambition, and business need.

This has direct implications for professional certification. In a dynamic skills environment, credentials cannot remain static symbols of participation or completion. They must represent verifiable evidence of capability, relevance, and application. The certification of the future will be less decorative and more diagnostic. It will need to answer not only whether a professional completed a program, but what that professional can demonstrate, at what level, under which criteria, and in what kind of context.

Careers as Networks, Not Ladders

Another important insight in the MIT Sloan article is the shift from a career ladder to a career lattice. Not every professional will progress through traditional vertical promotion. Some will move laterally into new business areas. Others will pursue emerging roles that require new combinations of technical and business judgment. Some may intentionally redesign their careers around flexibility, specialization, or different forms of contribution.

This changes the role of assessment. A strong assessment system should not merely confirm whether someone is ready for the next hierarchical position. It should reveal adjacent capabilities, transition potential, trainable gaps, and forms of talent that may be invisible in the formal structure. In other words, measuring competence can expand the organization’s field of decision making.

For boards, executives, and transformation leaders, this is especially relevant. Many companies claim to face talent scarcity while underusing the capabilities they already have. The issue is not always a complete absence of qualified people. Often, it is the absence of reliable visibility into what those people know, what they can learn, and where they could create value next.

Methodology Is the Real Advantage

Artificial intelligence will play an increasingly important role in identifying skills gaps. Yet its most valuable contribution will not be prettier dashboards or faster reports. Its real value will be to force organizations to clarify criteria, improve data discipline, rethink work models, and treat human capability as a strategic asset that can be assessed with rigor.

For Exent, this discussion goes to the heart of certification and professional assessment. Validated competence is not reputation, intuition, or self declaration. It is evidence organized under clear standards. When properly designed, assessment does not reduce professionals to a score. It gives them a more precise understanding of their value, their development path, and their readiness for consequential work.

The future of skills management will not belong to organizations that simply adopt AI. It will belong to those that combine technology with governance, trust, and methodological rigor. The competitive advantage will not be only in discovering skills gaps. It will be in building credible systems capable of turning those gaps into development, mobility, recognition, and professional excellence.