Human-AI Levelling Framework (HALF)

Executive Summary

Introduction and Purpose

The Human-AI Levelling Framework (HALF) is a comprehensive, global job levelling methodology designed for the modern, AI-augmented workplace. Its purpose is to provide a systematic, analytical, and legally compliant process for determining the relative contribution of all jobs, whether performed by humans, AI, or in collaboration. By establishing a transparent and common language for value, HALF serves as the strategic foundation for critical people programs, including workforce planning, talent management, career pathing, and, most importantly, fair and equitable compensation.

Core Architecture and Key Differentiators

HALF is built on a dual-core architecture that integrates enduring legal principles with the realities of 21st-century work.

• Legally Compliant Foundation: The framework is built upon the four globally recognized, gender-neutral compensable factors: Skills, Effort, Responsibility, and Working Conditions. This ensures inherent alignment with major pay equity legislation, including the EU Pay Transparency Directive and the U.S. Equal Pay Act.
• Future-Ready Overlays: Layered upon this core are two innovative differentiators to measure new forms of value:
  • The Automation Stewardship Overlay: This provides a sophisticated lens for evaluating how a role interacts with intelligent systems, moving beyond basic use to encompass collaboration, orchestration, and architectural design.
  • The Critical Thinking Dimension: This explicitly values the non-automatable cognitive skills—such as nuanced judgment, ethical reasoning, and systemic analysis—that are becoming the paramount differentiators of human contribution.

Alignment with Global Equal-Value Legislation

HALF is explicitly designed to meet and exceed the requirements of major international pay equity laws. Its architecture directly addresses the core tenets of the EU's Pay Transparency Directive (2023/970), the International Labour Organization's (ILO) Equal Remuneration Convention (C100), and the U.S. Equal Pay Act of 1963. By grounding its methodology in these universally recognized legal principles, HALF provides a globally consistent and legally resilient foundation for all compensation-related decisions.

1. Executive Summary

Framework Purpose and Strategic Value

The Human-AI Levelling Framework (HALF) is a comprehensive, global job levelling methodology designed to provide a strategic foundation for an organization's talent and reward architecture in the age of artificial intelligence. It serves as a systematic and analytical process to determine the relative contribution of all jobs—whether performed by humans, AI, or in collaboration—thereby creating a coherent and defensible internal value hierarchy. The primary purpose of the HALF is to establish a common, transparent language for understanding and assessing value, which is essential for the effective design and administration of critical people programs. These programs include, but are not limited to, strategic workforce planning, talent acquisition, career enablement, learning and development, and, most critically, fair and equitable compensation. By moving beyond simple job classification, the HALF provides the core infrastructure necessary to manage a modern, dynamic, and globally distributed workforce.

Key Differentiators: Integrating AI on a Legally Compliant Core

The principal innovation of the Human-AI Levelling Framework lies in its dual-core architecture, which marries enduring legal principles with the realities of modern work. The framework is built upon the four legally mandated, gender-neutral compensable factors: Skills, Effort, Responsibility, and Working Conditions. This foundation ensures that the methodology is not only robust and objective but also inherently aligned with global pay equity legislation from its inception.

Layered upon this compliant core are two modern, forward-looking differentiators designed to accurately measure the new forms of value created in an AI-augmented workplace. The Automation Stewardship Overlay provides a sophisticated lens for evaluating how a role interacts with intelligent systems, moving beyond basic use to encompass collaboration, orchestration, and architectural design. Concurrently, the integration of a Critical Thinking Dimension within the core factors ensures that the framework explicitly values the non-automatable cognitive skills—such as nuanced judgment, ethical reasoning, and systemic analysis—that are becoming the paramount differentiators of human contribution. This integrated approach contrasts sharply with legacy systems, which often struggle to evaluate technology beyond a simple skill or tool, failing to capture its transformative impact on the very nature of work and responsibility.

Alignment with Global Equal-Value Legislation

The HALF is explicitly designed to meet and exceed the requirements of major international and national pay equity laws. Its architecture directly addresses the core tenets of the European Union's Pay Transparency Directive (2023/970), which mandates the use of objective, gender-neutral criteria to assess work of equal value. Similarly, the framework aligns with the International Labour Organization's (ILO) Equal Remuneration Convention (C100), the foundational global treaty requiring objective job appraisal to ensure equal pay for work of equal value. Furthermore, the four core factors of the HALF mirror the criteria of "skill, effort, responsibility, and working conditions" established by the U.S. Equal Pay Act of 1963, providing a defensible basis for pay decisions in that jurisdiction. By grounding its methodology in these universally recognized legal principles, the HALF provides a globally consistent and legally resilient foundation for all compensation-related decisions.

Core Design Principle

The Human-AI Levelling Framework operates on a single, guiding principle that informs every aspect of its design and application: "As routine cognitive and physical tasks are progressively automated, the primary differentiators of human value shift to critical thinking, contextual understanding, and the effective stewardship of intelligent systems, in addition to core domain expertise." This principle recognizes that in a world where AI can generate answers, the premium is on the ability to ask the right questions, validate the outputs, and integrate them into a broader strategic and ethical context. This philosophy is embedded throughout the framework's factor definitions, scoring anchors, and innovative overlays, ensuring it remains relevant and effective for the next generation of work.

2. Design Principles and Architecture

Foundational Principles: Objective, Gender-Neutral, and Globally Consistent

The integrity and effectiveness of the Human-AI Levelling Framework are rooted in a set of non-negotiable design principles. These principles ensure that every job evaluation is conducted in a manner that is fair, consistent, and defensible across all functions and geographies.

• Objectivity: The framework is designed to evaluate the role and its requirements, not the individual performing it. All assessments are based on the standard of fully acceptable performance, disregarding the incumbent's personal characteristics, performance level, or current compensation. This focus on the job's intrinsic demands is crucial for eliminating subjective bias.
• Gender Neutrality: The factors, definitions, and scoring anchors are constructed to be inherently gender-neutral. They focus on the nature of the work to be performed, avoiding language or concepts that may be stereotypically associated with a particular gender. This principle is fundamental to achieving pay equity.
• Global Consistency: The framework uses a common set of factors and a universal point scale to evaluate all jobs worldwide. This creates a single, coherent architecture that allows for meaningful comparisons of job value across different business units, functions, and countries, providing a solid foundation for global talent management and reward programs.
• Analytical Rigor: The HALF employs a point-factor methodology. This approach, which breaks jobs down into defined factors and assigns numerical points to different levels of contribution, is recognized as a more analytical, detailed, and objective method for supporting pay equity compared to less structured methods like whole-job ranking or slotting.

The Four Pillars of Value: Skills, Effort, Responsibility, Working Conditions

The architecture of the HALF is built upon four foundational pillars. These are not arbitrary constructs but are derived directly from the core principles of major pay equity legislation around the world, including the ILO Convention C100, the EU Pay Transparency Directive, and the U.S. Equal Pay Act. By using these legally recognized compensable factors as its primary structure, the framework ensures its fundamental compliance and defensibility.

The selection of these four pillars is a deliberate architectural choice. An analysis of established, proprietary job evaluation methodologies—such as the Hay Group's "Know-How, Problem Solving, Accountability," Mercer's "Impact, Communication, Innovation, Knowledge," and Aon/Radford's "Knowledge, Problem-Solving, Interaction, Impact"—reveals that their factors are, in essence, customized interpretations and combinations of the four legal pillars. By building the HALF directly on the legal "source code" of equal value, the framework becomes more fundamental, transparent, and universally applicable.

The four pillars are:

• Skills: The knowledge, competencies, and experience required to perform the role.
• Effort: The mental, cognitive, and physical exertion required.
• Responsibility: The accountability for outcomes, resources, and decisions.
• Working Conditions: The physical and psychosocial environment in which the work is performed.

Innovative Overlays: The Automation Stewardship Tier and Critical Thinking Dimension

To ensure the framework accurately captures the nuances of modern, AI-augmented work, two innovative analytical layers are integrated with the four pillars:

• The Automation Stewardship Overlay (A0-A4): This overlay provides a structured way to assess the sophistication of a role's interaction with AI and automation. It creates a spectrum ranging from passively using predefined tools to actively architecting new human-machine systems. This element is not scored separately but acts as a critical lens for evaluating the Skills and Responsibility factors, ensuring that the framework recognizes and values the increasing demand for effective human oversight and direction of intelligent technologies.
• The Critical Thinking Dimension: This dimension is not a standalone factor but a meta-skill woven into the behavioral anchors of the Skills, Effort, and Responsibility pillars. As AI automates routine information processing, the human capacity for deep analysis, evaluation, verification, and self-regulated judgment becomes a primary driver of value. The framework explicitly measures the application of these cognitive skills in a work context, ensuring they are a key differentiator in determining job value.

The Dual-Track Architecture: Professional/Expert and Managerial/Leadership Paths

The HALF is explicitly designed to support dual career ladders, a recognized best practice for attracting and retaining high-value technical and subject-matter experts who may not desire or be suited for a traditional management path. This structure provides parallel and equally valued progression opportunities for both individual contributors and people leaders.

A critical feature of the HALF architecture is that both the Professional/Expert track and the Managerial/Leadership track are evaluated using the exact same four factors and point system. This ensures internal equity and a shared understanding of value across the organization. The differentiation between the tracks emerges naturally from the scoring process:

• Professional / Expertise Track: Roles on this track create value through the depth of their knowledge, analysis, design, or innovation. They typically score highest on the Skills factor (reflecting deep or broad mastery) and the cognitive components of the Effort factor. Their Responsibility is primarily for the quality, originality, and correctness of their outputs.
• Managerial / Leadership Track: Roles on this track create value through the coordination and orchestration of others (both human and digital agents). They score highest on the Responsibility factor (reflecting the scope of resources, teams, and business outcomes they are accountable for) and the interpersonal components of the Skills factor.

This unified approach demonstrates that an organization values both deep expertise and broad orchestration, allowing a Principal Engineer and a Director of Engineering to be placed at the same job level, with equivalent compensation potential, because their total contribution to the organization is deemed equal, albeit achieved through different means.

3. Factor Definitions and Scoring Methodology

The core of the Human-AI Levelling Framework is its point-factor evaluation system. Each of the four pillars—Skills, Effort, Responsibility, and Working Conditions—is scored on a scale of 0 to 25 points. The following sections provide the modernized definition for each factor and the detailed scoring matrix with behavioral anchors that explicitly incorporate the context of an AI-augmented workplace.

Factor 1: Skills

Modern Definition: Consistent with legal definitions, "Skills" encompasses the experience, ability, education, and training required for fully competent performance of the job. The HALF modernizes this definition to recognize that in a contemporary work environment, skills are not merely about possessing knowledge (domain expertise) but also about the ability to interact with, manage, and leverage complex knowledge systems. This includes the technical and cognitive capabilities required to effectively question, guide, validate, and synthesize outputs from AI agents and other intelligent tools. The skill lies in the application of knowledge, whether human- or machine-derived, to solve problems and create value.

Measurement Indicators: The level of skill is measured by its depth (specialization), breadth (range of subjects), and the complexity of its application. This includes interpersonal and communication skills required for influencing and collaborating with others. The Automation Stewardship Overlay provides a critical lens for assessing the sophistication of technology-related skills.

Table 1: Factor Scoring Matrix - SKILLS

Factor 2: Effort

Modern Definition: Legally defined as the amount of physical or mental exertion needed to perform a job. The HALF places significant emphasis on the cognitive and emotional dimensions of effort. As AI and automation absorb an increasing share of routine, repetitive, and physically demanding tasks, the nature of human effort shifts decisively toward sustained, high-stakes mental exertion. This "invisible" effort includes the deep concentration required for strategic analysis, the cognitive load of managing and synthesizing vast streams of information, the creative energy for innovation and complex problem-solving, and the emotional labor involved in leadership, negotiation, and high-stakes stakeholder management.

Measurement Indicators: Effort is measured by the intensity, duration, and frequency of the required exertion. It considers the complexity of the mental processes involved (e.g., analysis, synthesis, evaluation), the pressure of the environment (e.g., deadlines, consequences of error), and the need for emotional regulation and interpersonal engagement.

Table 2: Factor Scoring Matrix - EFFORT

Factor 3: Responsibility

Modern Definition: Consistent with legal definitions, "Responsibility" is the degree of accountability required in the performance of the job. The HALF expands this concept to encompass not only accountability for final outcomes (e.g., financial results, project delivery) but also for the processes, tools, and systems used to achieve them. In an AI-driven organization, this includes accountability for the ethical deployment of automated systems, the integrity and accuracy of algorithmic outputs, the management of data privacy and security risks, and the overall governance of human-machine decision-making processes.

Measurement Indicators: Responsibility is measured by the scope and impact of the role's decisions and actions. This is assessed through dimensions such as freedom to act (autonomy), the magnitude of resources controlled or influenced (e.g., budget, assets, revenue), and the nature of the impact on the organization's objectives.

Table 3: Factor Scoring Matrix - RESPONSIBILITY

Factor 4: Working Conditions

Modern Definition: This factor covers the physical surroundings and hazards of the job, as required by law. The HALF modernizes this pillar by formally including the psychosocial and cognitive environment as a compensable condition. This acknowledges that in modern knowledge work, significant hazards are not only physical but also psychological. These can include the stress from constant digital surveillance, the cognitive overload from managing incessant information flows, the pressure of an "always-on" culture, and exposure to sensitive, distressing, or harmful digital content.

Measurement Indicators: Working conditions are measured by the frequency and severity of exposure to unpleasant or hazardous elements. This includes physical risks as well as psychosocial risks (e.g., high-stress environment, exposure to traumatic material, risk of digital harassment).

Table 4: Factor Scoring Matrix - WORKING CONDITIONS

4. The Automation Stewardship Overlay (A0-A4)

The Automation Stewardship Overlay is a critical, modernizing component of the Human-AI Levelling Framework. It is not a separate, scorable factor. Instead, it serves as an analytical lens through which the core factors of Skills and Responsibility are evaluated. It provides a standardized vocabulary to describe the sophistication and proactivity of a role's interaction with AI, automation, and other intelligent systems. As organizations increasingly deploy these technologies, the ability to effectively manage, direct, and improve them becomes a key differentiator of job value. The A-Tier of a role directly influences the degree level selected in the Skills and Responsibility scoring matrices.

Defining the Tiers of Human-AI Collaboration

The overlay consists of five distinct tiers, representing a maturity model of human-AI collaboration:

• A0: User: The role operates predefined tools and systems as instructed. The primary activity is consumption of AI-driven outputs or following system-guided processes.
• A1: Collaborator: The role actively partners with AI tools to co-create novel outputs. This involves using advanced techniques like sophisticated prompt engineering, chaining multiple tools, and iteratively refining AI-generated content.
• A2: Orchestrator: The role manages and directs a portfolio of human and digital agents to achieve a complex outcome. This involves delegating tasks to the most appropriate agent and integrating their outputs.
• A3: Coach: The role is responsible for training, refining, and improving the performance of AI models and automated systems. This involves providing high-quality feedback, curating training data, and fine-tuning models.
• A4: Architect: The role designs, commissions, and is ultimately accountable for new, integrated human-AI systems to solve fundamental business problems.

Table 5: The Automation Stewardship Tier Matrix

5. Critical Thinking as a Core Competency

In the Human-AI Levelling Framework, Critical Thinking is not treated as a standalone, scorable factor. Instead, it is recognized as a foundational meta-skill—a core competency that is embedded within the behavioral anchors of the Skills, Effort, and Responsibility factors. The increasing prevalence of AI makes critical thinking more, not less, important. As AI systems handle the "what" (information retrieval, data processing, content generation), the premium on human value shifts to the "so what" and "now what," which are the domains of critical thought.

The framework integrates a comprehensive model of critical thinking, with a particular emphasis on three components:

1. Questioning: The ability to formulate precise, insightful, and challenging questions to probe for deeper understanding, uncover hidden assumptions, and effectively guide AI tools.
2. Verification and Tracking: The discipline of scrutinizing AI-generated outputs for accuracy, relevance, and bias. This includes cross-referencing sources, identifying "hallucinations," and maintaining an audit trail of the reasoning process.
3. Self-Regulation (Metacognition): The capacity for introspection—monitoring one's own thinking for cognitive biases (like confirmation bias or automation bias), assessing the limits of one's own knowledge, and adjusting one's judgment and strategy accordingly.

Behavioral Anchors for Critical Thinking by Level

Foundational Levels (L1-L3): At these levels, critical thinking is focused on immediate tasks and outputs.

• Skills Anchor Example: "Identifies and flags obvious errors or inconsistencies in data provided by standard reporting tools or AI assistants (demonstrates basic Verification)."
• Responsibility Anchor Example: "Is accountable for escalating issues or outputs that do not seem correct based on established procedures (demonstrates emerging Self-Regulation)."

Professional Levels (L4-L6): At these levels, critical thinking becomes more analytical and proactive.

• Skills Anchor Example: "Analyzes and synthesizes information from multiple sources, including AI-generated reports, to identify underlying trends and relationships. Formulates clarifying questions to refine AI queries and improve output quality (demonstrates Analysis and Questioning)."
• Effort Anchor Example: "Requires sustained mental concentration to evaluate the logic and evidence supporting an AI's recommendation before acting on it."
• Responsibility Anchor Example: "Is accountable for the validity of the analysis presented, including a justification of the methods and sources used, both human and machine (demonstrates Explanation)."

Expert/Leadership Levels (L7-L10): At the highest levels, critical thinking is systemic, strategic, and highly self-aware.

• Skills Anchor Example: "Pioneers novel analytical frameworks that integrate human expertise and machine intelligence, defining the critical questions the organization should be asking (demonstrates advanced Questioning and Inference)."
• Responsibility Anchor Example: "Makes high-stakes strategic decisions based on AI-driven insights after rigorously evaluating the model's limitations, potential biases, and second-order ethical and business consequences. Is accountable for mitigating systemic risks associated with the use of automated decision systems (demonstrates advanced Evaluation and Self-Regulation)."

6. The Global Job Level Matrix (L0-L10)

The Global Job Level Matrix is the culminating output of the Human-AI Levelling Framework. It translates the analytical point scores derived from the factor evaluations into a clear, consistent, and transparent hierarchy of job levels that is applicable across the entire organization. The total point score for a job is calculated by summing the points from the four factors: Skills, Effort, Responsibility, and Working Conditions.

Table 6: The Global Job Level Matrix (L0-L10)

7. Applying the Dual Career Tracks

The dual-track architecture is a cornerstone of the Human-AI Levelling Framework, designed to provide equitable and parallel career progression for employees on both the Professional/Expert and Managerial/Leadership paths. Both tracks are measured against the same four compensable factors, ensuring that the organization's definition of "value" is applied consistently.

Valuing Depth vs. Breadth: The Professional/Expert Track

The Professional/Expert track is designed for individuals who create value through deep subject matter expertise, innovation, analysis, and the creation of intellectual property. A high-level role on this track achieves a high total score primarily through exceptional ratings in the Skills factor, reflecting their mastery of a complex domain. They also score highly on the cognitive aspects of the Effort factor.

Valuing Orchestration: The Managerial/Leadership Track

The Managerial/Leadership track is for individuals who create value by setting direction, allocating resources, and developing the capabilities of others—both human and digital. An equivalent high-level role on this track achieves a high total score primarily through exceptional ratings in the Responsibility factor, reflecting their accountability for significant business outcomes.

Illustrative Scoring Scenarios

To demonstrate how the framework equitably values different contributions, consider two roles at Level 5 (L5):

Both roles fall within the L5 range (80-94 points), demonstrating how different contributions can be of equal value to the organization.

8. Implementation Toolkit

To ensure the consistent and effective application of the Human-AI Levelling Framework, the following tools and guidelines are provided.

Job Evaluation Scoring Template

9. Compliance and Legal Alignment

The Human-AI Levelling Framework is architected with global legal compliance as a core design principle. Its structure, factors, and methodology directly align with the foundational requirements of major pay equity legislation, providing a robust and defensible basis for compensation decisions worldwide.

Mapping to EU Directive 2023/970

The EU Pay Transparency Directive requires employers to establish and maintain pay structures based on objective, gender-neutral criteria. The HALF directly satisfies this mandate:

• Objective, Gender-Neutral Criteria: The framework's four pillars are the explicit criteria cited within the directive's supporting principles for assessing the value of work.
• Transparency and Justification: The point-factor methodology generates a quantitative, documented output for every job evaluation, providing the objective evidence required to justify any pay differentials.
• Comparability of Dissimilar Jobs: The framework's ability to assign a value score to any job allows for the comparison of dissimilar jobs, which is central to identifying and remedying systemic pay gaps.

Satisfying ILO C100 Principles

The International Labour Organization's Convention C100 is the global standard for equal remuneration. Its central tenet is the promotion of an "objective appraisal of jobs on the basis of the work to be performed". The HALF is a direct implementation of this principle through its structured methodology, detailed factor definitions, and graduated behavioral anchors.

Parallels with the U.S. Equal Pay Act

The U.S. Equal Pay Act of 1963 (EPA) prohibits sex-based wage discrimination for jobs that require "substantially equal skill, effort, and responsibility, and which are performed under similar working conditions". The four pillars of the HALF are a direct reflection of these legally defined factors.

Guidance for Joint Pay Assessments

Under the EU Directive, if a company's pay reporting reveals an average gender pay gap of 5% or more within any category of workers that cannot be justified by objective, gender-neutral factors, a "joint pay assessment" with employee representatives is mandatory. The HALF provides the ideal analytical tool for conducting this assessment through data-driven analysis, identification of work of equal value, and objective justification documentation.