From Job Description to Purpose Description An Organizational Theory of Dynamic Role Design and Future Value Creation in the Age of AI Naoki Kadowaki Founder & CEO, VURA Capital Innovation Holdings Inc. · Lecturer, Keio Business School ORCID 0009-0009-5295-1506 This version: August 2026 · Working Paper Version 1.1 “A job description allocates present tasks. A purpose description distributes the creation of future value.” Abstract Artificial intelligence (AI) is rapidly expanding the scope of analysis, prediction, generation, decision support, and task execution, rendering human roles within firms increasingly fluid. In many organizations, however, roles are still defined around the Job Description (JD)—a collection of assigned tasks, areas of responsibility, required competencies, and reporting relationships. While this static allocation model serves stable divisions of labor well, it cannot adequately express the purpose of a role or the value it creates in environments where the mapping between tasks and humans changes continuously and the human-AI collaborative system itself becomes an object of redesign. Connecting work design theory with dynamic capabilities research, this paper proposes the Purpose Description (PD), a concept of dynamic role design positioned above the JD. A PD is an organizational design unit that articulates—in light of the future expectations shared by markets and society and the future the firm seeks to realize—why an individual or team exists, for whom it creates what future value, and on the basis of which questions and decision principles it acts. The paper’s central thesis is that whereas a JD allocates present tasks, a PD distributes the creation of future value. The PD is positioned as structural middleware that 1 connects the market’s expectations of the future captured by Future Value Theory (Kadowaki, 2026a) and the continuous redesign of the enterprise addressed by Enterprise Redefinition (Kadowaki, 2026b) to the actions of individuals and teams. The paper presents a Two-Tier Dynamic Role Architecture that integrates PD and JD as complements, together with a synchronization protocol. It further sets out the causal mechanism by which purposeless local optimization in AI adoption is avoided, nine testable propositions, directions for measurement, legal and labor governance, safeguards against over-adaptation, and human authority to halt AI systems. The creation of future value depicted by the PD and the execution of PD-defined roles are carried out by integrated human-AI systems, but the determination of purpose and decision principles, exception handling, halt authority, and legal and ethical accountability remain vested in humans. The PD does not substitute for working conditions, authority, or accountability; it preserves them while placing existential purpose and the connection to future value above ever-changing means. Keywords: Purpose Description, Job Description, Future Value Theory, Enterprise Redefinition, human-AI collaboration, dynamic organization design, complementary assets, dynamic capabilities JEL codes: M10, M12, M54, O33, D23, J24 This is a working paper presenting a conceptual framework in development. Comments are welcome. Companion working papers, “Future Value Theory: A Management Framework for Enterprise, Capital, and Society in the Age of AI” (Kadowaki, 2026a), available at SSRN (https://ssrn.com/abstract=7120980), and “Enterprise Redefinition: Toward an Enterprise Evolution Theory for the Age of AI” (Kadowaki, 2026b), available at SSRN (https://ssrn.com/ abstract=7210118). 1 Introduction Firms have long designed their organizations by decomposing work into jobs and assigning each job responsibilities, authority, required competencies, and reporting relationships (Chandler, 1962; Hackman & Oldham, 1976). The Job Description is the emblematic institution of this approach. By clarifying who is in charge of what, who is responsible for what, and what outcomes are expected, the JD has supported the division of labor, specialization, hiring, compensation, and equitable evaluation. The development of generative AI and autonomous AI agents, however, is eroding the stability of task composition and job boundaries that the JD has implicitly presupposed. What matters is not that particular jobs vanish all at once. Rather, the boundaries of the fine-grained tasks that constitute a job—and the appropriate From Job Description to Purpose Description VURA Working Paper Series 2 allocation of those tasks among humans, AI, and external partners—keep changing continuously and nonlinearly. Empirical research shows that generative AI assistance raises productivity in customer support by 15% on average, with especially large effects for less experienced workers (Brynjolfsson, Li, & Raymond, 2025). In knowledge work, AI’s capability boundary is not uniform but forms an irregular “jagged frontier” across tasks (Dell’Acqua et al., 2023). At the same time, technology adoption does not by itself raise firm value automatically. Exploiting AI as a general-purpose technology requires redesigning business processes, human capital, organizational structure, decision rights, and complementary assets (Bresnahan, 2024; Iansiti & Lakhani, 2020). The challenge, therefore, is not merely the frequency with which Job Descriptions are updated. The more fundamental problem is that while a Job Description primarily defines “what is to be done,” it does not adequately address “why the role exists,” “what future it is meant to realize,” “for whom it creates what value,” or “by what principles the means—including AI—are to be re-selected.” The more AI expands the means of execution, the less stably human roles can be defined by “the residual tasks AI cannot perform,” because AI’s capability boundary keeps shifting. To define human roles durably, the starting point must be purpose, questions, judgment, responsibility, and the value to be created—not tasks. This paper therefore proposes the shift from Job Description to Purpose Description. This is not a renaming exercise: it is a transition from organizations that position humans as executors of predefined work to organizations that position humans as the definers of, and accountable principals for, the future and the value to be realized. 2 Research Questions and Contributions This paper poses three research questions. First, when AI renders task structures fluid, by what should the roles of individuals and teams be defined? Second, how can the purpose and future value a firm espouses be connected to frontline judgment and to the division of labor between humans and AI? Third, how do the clarity and alignment of a Purpose Description affect whether AI adoption stops at technical success or advances to success in terms of future value? The paper makes four contributions. First, it defines roles not as bundles of fixed tasks but as connections to the creation of future value. Second, while integrating work design, meaningful work, job crafting, and organizational purpose research, it presents a formal organizational design unit distinct from all of them. Third, it provides microfoundations that translate the firm-level concept of dynamic capabilities into frontline questions and decision principles. Fourth, it builds in a synchronization From Job Description to Purpose Description VURA Working Paper Series 3 protocol that maintains dynamic alignment between JD and PD, together with legal and labor guardrails, thereby increasing implementability. In addition, it distinguishes the PD from OKR and MBO, positioning it not as a goal-management technique but as a unit of dynamic organization design and governance. The governance principle running through this paper is unambiguous. It is humans and AI together that create future value and execute PD-based roles. But the final responsibility for deciding what to aim for, which principles to uphold, what to decide in exceptional situations, when to halt AI, and who explains the results remains with humans. Value creation can be shared; accountability cannot be transferred to AI. 3 Theoretical Background 3.1 Job Descriptions and Work Design In practice, a Job Description organizes a job’s purpose, principal responsibilities, duties, required competencies, authority, and reporting relationships. This paper therefore does not claim that all Job Descriptions lack purpose. The problem is that institutional practice tends to center on assigned tasks and boundary management, with purpose reduced to a brief annotation. Work design research has never treated jobs as mere task lists. Hackman and Oldham’s (1976) job characteristics model showed that skill variety, task identity, task significance, autonomy, and feedback shape the experienced meaningfulness of work and sense of responsibility, which in turn affect motivation. Morgeson and Humphrey (2006) offered a broader measurement framework encompassing not only task characteristics but also knowledge, social relationships, and work context. Organization design theory, moreover, formulates the fundamental problem of organizing as the division of tasks and the integration of effort (Puranam, 2018), with role definition at its core. These literatures ground the Purpose Description, but they deal primarily with the characteristics and experience of existing jobs. The Purpose Description goes further, continuously reconstituting the job itself from the starting point of changing future value. 3.2 Meaningful Work, Calling, and Job Crafting Research on meaningful work concerns the effects on individuals and organizations of experiencing work as significant and meaningful. Blustein, Lysova, and Duffy (2023) distinguish decent work from meaningful work and position the former as an important precondition for the latter. Calling research addresses the experience of work as a vocation, and job crafting research addresses the process by which From Job Description to Purpose Description VURA Working Paper Series 4 individuals actively alter task, relational, and cognitive boundaries (Wrzesniewski & Dutton, 2001). The Purpose Description overlaps with these constructs but is not identical to them. Whereas meaningful work concerns primarily the subjective experience of work, calling concerns individual orientation, and job crafting concerns individual-initiated modification of an existing job, the Purpose Description integrates into a single descriptive unit not only personal meaning but also beneficiaries, corporate purpose, the future expectations of markets and society, decision principles, and accountability. It also applies to teams as well as individuals and is created dialogically as part of formal organization design. 3.3 Organizational Purpose and Strategic Alignment Organizational purpose expresses why a firm exists beyond profit and what it brings to whom (Gartenberg, Prat, & Serafeim, 2019; George et al., 2023). Yet firm-level declarations do not translate automatically into frontline roles. Unless an abstract purpose is converted into individual judgments, priorities, human-AI allocations, and evaluation metrics, it remains a symbolic expression. The Purpose Description is the intermediate layer that translates corporate purpose into the purposes and future value of individuals and teams. 3.4 AI, Complementary Assets, and Organizational Redesign The value of information technology is realized not by the technology alone but through complementarities with business processes, skills, decision rights, incentives, and organizational capital. The same holds for AI. The Stanford Digital Economy Lab’s synthesis of 51 enterprise cases attributes performance differences not merely to model selection but to organizational readiness, processes, leadership, and transformation capability (Pereira et al., 2026). This finding indicates that AI adoption must be understood not as “fitting available AI capabilities onto existing operations” but as “redesigning work and roles from the value the organization seeks to realize.” Complementary-assets reasoning alone, however, offers only a weak normative starting point for what the redesign should aim at. The Purpose Description supplies that starting point by making explicit the future value, beneficiaries, questions, and decision principles. 3.5 Theoretical Grounding in Future Expectations, Resources, and Strategy Formation Future Value and Enterprise Redefinition as used in this paper do not rest solely on the author’s prior working papers. Future Value can be connected to signaling, whereby From Job Description to Purpose Description VURA Working Paper Series 5 present actions convey uncertain future quality to markets (Spence, 1973); to organizational learning, whereby the allocation between exploration and exploitation shapes future capabilities (March, 1991); and to expectation formation, whereby imagined but uncertain futures orient present economic action (Beckert, 2016). This paper integrates these strands and operationalizes future value as the change in state that beneficiaries will attain, the new options the organization acquires, and shifts in the expectations of markets and stakeholders. Enterprise Redefinition is theoretically grounded in the Resource-Based View, which addresses the reconfiguration of valuable resources and complementary assets (Barney, 1991); in Dynamic Capabilities, which concern sensing, seizing, and reconfiguring resources in response to environmental change (Teece, 2007; Eisenhardt & Martin, 2000); and in the interplay of intended and emergent strategy (Mintzberg & Waters, 1985). The PD is a microfoundation that translates these firm-level processes of resource reconfiguration and strategy formation into the questions, decision principles, and human-AI allocations of individuals and teams (Helfat & Peteraf, 2015). Kadowaki (2026a, 2026b) are therefore complementary references that trace the developmental lineage of the concepts, not grounds for circularly validating this paper through the author’s own definitions. The case observations of eighteen firms within the same research program (Kadowaki, 2026c, 2026d) are likewise cited as complementary illustration. Institutional theory shows that JDs and appraisal systems are stabilized not only by efficiency but by professional norms, legal institutions, and mimetic pressure from industry peers (DiMaggio & Powell, 1983). The PD must therefore be institutionalized incrementally, maintaining consistency with contracts, professional norms, and safety regimes rather than declaratively replacing existing institutions. Sensemaking theory (Weick, 1995), in turn, explains how organizational members under ambiguity iterate between action and interpretation to form shared meaning. Defining Questions and the Decision Log convert this meaning-making into an observable and correctable organizational process. Cognitive Systems Engineering further shows that safety is not secured merely by formally retaining human responsibility; the ensemble must be designed as a joint cognitive system in which humans can understand the situation, track change, and intervene (Hollnagel & Woods, 2005). From this perspective, the PD is derived as a roledesign unit that binds together purpose, decision principles, visibility, intervention authority, and responsibility boundaries within the Human-AI System. The paper’s core logic is thus constructed directly from established theory; Kadowaki (2026a, 2026b) serve only as complementary references marking the lineage of the research program. From Job Description to Purpose Description VURA Working Paper Series 6 3.6 Comparison with OKR (Objectives and Key Results) and MBO Any discussion of role fluidity and agile goal setting must clarify the differences from Management by Objectives (MBO: Drucker, 1954) and from Objectives and Key Results, which originated at Intel and diffused to Google and beyond (OKR: Grove, 1983; Doerr, 2018). MBO and OKR are goal-management approaches that align organizational and individual action by setting objectives to be pursued over a given period together with the results or criteria that indicate their attainment. Whereas OKR and MBO are execution- and goal-management techniques that stipulate “what is to be achieved within a given period” (What/Target), the PD defines, for organizations in which humans and AI collaborate, “why the unit exists and what future value it creates for whom” (Why/Beneficiary/Future Value), and it is the minimal unit of dynamic organization design (a Structural Governance Unit) that governs the division of labor and responsibility boundaries between humans and AI in environments where tasks are fluid. OKR and MBO can run in parallel with the JD as a separate goal-management layer. The PD, by contrast, is the core of a two-tier dynamic role architecture that bridges the upper purpose layer and the lower contractual-control layer (the JD) through a synchronization protocol. The PD is therefore not an alternative goal-setting technique to OKR or MBO but a superordinate organizational-structure and governance concept that can subsume and orient them as needed. 4 Defining the Purpose Description This paper defines the Purpose Description as follows. A Purpose Description is a statement of why an individual or team exists, for whom it creates what future value—in light of the future expectations shared by markets and society and the future the firm seeks to realize—and of the questions it sets and the principles by which it judges in order to do so. This definition has three features. First, it situates purpose not solely in the individual’s inner meaning but in relations with the firm, beneficiaries, and society. Second, it conceives of performance not as current workload but as the state changes to be realized and the contribution to future value. Third, it does not fix the means of execution, continuously updating the combination of humans, AI, technology, and external partners. Accordingly, the executing agent that creates future value, explores the questions, and carries out the PD can be an integrated Human-AI System. However, the setting and guardianship of Decision Principles, exception handling, halt authority, From Job Description to Purpose Description VURA Working Paper Series 7 and accountability for outcomes are retained by a human as the Accountable Principal. The six components are organized into inputs (Purpose, Beneficiary), mediating processes (Defining Questions, Decision Principles), and outputs/outcomes (Desired Future, Future Value). 4.1 Comparing JD, OKR/MBO, and PD The PD is reducible neither to the traditional JD nor to goal-management techniques such as OKR and MBO. Note also that whereas a company-wide purpose statement expresses the raison d’être of the organization as a whole, the PD is the formal roledesign unit that translates it into the beneficiaries, questions, decision principles, and responsibility boundaries of individuals and teams. Table 1 summarizes the institutional differences among the three. Dimension Traditional JD OKR/MBO Purpose Description (PD) Fundamental nature A contractual and control document defining duties, responsibilities, and authority A goal- and progressmanagement technique aligning objectives and results An organizationalstructure and governance unit designing human-AI collaboration Primary question What is one in charge of and responsible for? (What) What is to be achieved within a given period? (Target) Why does the unit exist, and what future value does it create for whom? (Why/Who/Future Value) Temporal axis Present duties and scope of responsibility Short- to mid-term goals with deadlines Dynamic updating of questions and means, anchored in mid- to longterm purpose and principles Tasks and means Enumerated and assigned relatively statically Actions updated in pursuit of goal attainment Reallocated across humans, AI, and external resources on the basis of Defining Questions Relation to AI Assistance with or substitution of existing tasks A means of execution and productivity improvement supporting goal attainment Roles and allocations redesigned with the Human-AI System as the initial condition Responsibility and control Directly tied to job authority, direction and supervision, and Centered on goal evaluation and progress management; legal Humans designated as Accountable Principals; decision principles, exception handling, and From Job Description to Purpose Description VURA Working Paper Series 8 contractual responsibility responsibility handled by other institutions halt authority made explicit Institutional linkage A single contractualcontrol layer Runs in parallel with the JD as a separate goal-management layer Connects the purpose layer (PD) and the contractual-control layer (JD) via a synchronization protocol Table 1. Three Approaches to Role Management and Definition 4.2 Four Sources of Theoretical and Institutional Distinctiveness The PD presented in this paper is not merely an individual- or team-level version of an OKR document. It possesses theoretical and institutional distinctiveness on the following four counts. First: the shift from Target-Driven to Query/Principle-Driven. On a jagged frontier where uncertainty is high and AI’s capability boundary shifts continuously, targets set at the start of a period can themselves become obsolete within a short time. The PD anchors not on numerical targets but on Defining Questions that keep tracking environmental change and on Decision Principles that persist even as means change. Where OKRs are used, their Objectives and Key Results are subordinate to the PD’s purpose, questions, and principles and are updated as learning proceeds. Second: dynamic role reallocation within the Human-AI System. OKR and MBO deal primarily with goal attainment and alignment by human actors. The PD takes the integrated human-AI system as the initial condition of design and, when advances in AI automate tasks, provides organizational direction as to which untapped Future Value the freed human time and cognitive resources are to be reinvested in. Third: institutional embedding of responsibility boundaries. Goal and performance management under OKR or MBO alone does not settle questions of ethics, safety, compliance, authority to halt AI, or the locus of legal responsibility. The PD synchronizes with the subordinate JD and control layer and preserves the Accountable Principal norm that “value creation can be shared between humans and AI, but final accountability remains with humans.” Fourth: preventing the hollowing-out of purpose through the Query Translation Protocol. From Job Description to Purpose Description VURA Working Paper Series 9 Merely decomposing corporate purpose or corporate goals into numbers risks regression to KPI management at the front line, where purpose is reduced to a means. The PD propagates the corporate Future Value Hypothesis by translating it into the constraints the front line must remove and into Defining Questions. This connects managerial intent with frontline autonomy and curbs both symbolic purpose and passive waiting for instructions. 4.3 The Six Components Component What is described Purpose (raison d’être / input) Why does the unit exist? Which elements of the corporate purpose does it directly carry? Beneficiary (beneficiary / input) Whose problems does it solve, and to whom does it deliver value? Defining Questions (questions / process) What core questions must be asked continuously in order to track environmental change and update the role? Decision Principles (decision principles / process) What standards of ethics, quality, safety, customer priority, and responsibility persist even as means change? Desired Future (desired future / outcome) What state changes is the unit committed to realizing? Future Value (future value / outcome) What capabilities, options, expectations, and structural changes does it generate for beneficiaries, markets, and society? Table 2. The Six Components of the Purpose Description Specific tasks, technologies employed, staffing, and the division of labor between humans and AI are treated as variable elements subordinate to these six components. From Job Description to Purpose Description VURA Working Paper Series 10 Figure 1. An Integrated Governance and Operating Architecture Centered on the PD Note: OKR and JD do not stand in a hierarchical relationship to each other. Relative to the PD, the OKR synchronizes short-term results and learning, while the JD synchronizes contractual, authorityrelated, safety, and labor boundaries. Updates center on deltas and exceptions; the three documents are not fully revised each time. 4.4 Reconciling the Formal PD with Job Crafting Job crafting is a bottom-up act in which individuals voluntarily alter the task, relational, and cognitive boundaries of their work. The PD, by contrast, is a formal role design connected to superordinate future value, beneficiaries, decision principles, decision rights, and JD-level responsibility boundaries. The two are not opposed concepts, but individuals do not hold the authority to rewrite the PD unilaterally. Reconciliation proceeds through the following procedure: (1) the individual or team submits a Crafting Proposal grounded in frontline evidence; (2) the proposal’s implications for superordinate questions, beneficiaries, decision principles, capabilities and resources, and JD-level responsibilities are made visible; (3) the role owner decides whether to approve, run a time-limited experiment, modify, or reject; (4) changes affecting safety, law, budgets, or the job boundaries of others require prior approval; (5) experimental results are reflected in both the PD and the JD. This preserves frontline autonomy while preventing unauthorized abandonment of roles, shifting of responsibility, and local optimization. Corporate Strategy and Enterprise Purpose Future value hypotheses, beneficiaries, resource allocation, institutional guardrails Strategic translation Bidirectional updating from environmental change and field evidence Purpose Description (core) Purpose / Beneficiary / Defining Questions / Decision Principles / Desired Future / Future Value Outcome cycle Contract-and-control cycle OKR / KPI Short-term; updated at and within each period JD, work rules, authority, qualifications, safety Updated by formal procedure upon material role changes Results, learning, exceptions, objections — decision logs and leading indicators flow back to the PD and strategy From Job Description to Purpose Description VURA Working Paper Series 11 4.5 The Two-Tier Dynamic Role Architecture The PD does not abolish the JD outright. Role design is divided into an upper Purpose Layer and a lower Task & Accountability Layer. The upper layer defines purpose, beneficiaries, questions, decision principles, the desired future, and future value, providing directional stability. The lower layer defines current tasks, the human-AI division of labor, authority, legal obligations, safety standards, reporting relationships, and minimum performance standards, and is updated nimbly as technology advances. While the PD layer sets the direction of value creation by humans and AI, the lower control layer concretizes the accountable human agents, approval authority, monitoring duties, and halt procedures that humans retain. To contain divergence between the two layers and the costs of dual administration, a synchronization protocol is instituted. Synchronization is triggered by the introduction of new AI, the automation of major processes, significant regulatory or market change, or quarterly and semiannual reviews. The review examines whether “the current tasks and human-AI allocation contribute most to pursuing the questions defined in the PD and to creating future value,” and reflects in the lower layer the removal of unnecessary tasks, delegation to AI, and the addition of new human judgments and exception handling. That said, the joint formulation of PDs, their periodic updating, and inter-layer synchronization entail nontrivial costs of dialogue, information gathering, and consensus building. PDs should not be updated at the same frequency and precision for all jobs; operational intensity must vary with the pace of change in AI capabilities, markets, regulation, and risk. Conversely, the more AI lowers the costs of executing and coordinating routine tasks, the more an organization can redirect part of the time and resources thereby freed toward dialogue about purpose, resetting of questions, confirmation of decision principles, and redefinition of future value. PD operation should be designed not as additional administration but as organizational reinvestment made possible by falling execution costs. This two-tier structure avoids both the abstraction of purpose and the blurring of responsibility. A Purpose Description alone can leave unclear who bears final responsibility, what must not be done, and which regulations apply. Conversely, a Job Description alone struggles with changes beyond its stipulations. The two stand not in a relationship of substitution but in a complementary relationship of superordinate purpose and subordinate control. From Job Description to Purpose Description VURA Working Paper Series 12 5 Connecting to Future Value Theory and Enterprise Redefinition Future Value Theory conceives of firm value not merely as the accumulation of past performance but as market-shared expectations about “what future the firm can bring into being” (Kadowaki, 2026a). Market here is not limited to capital markets; it includes expectation formation by customers, employees, business partners, local communities, and other stakeholders. Expectations, however, are not realized by declaration alone. The firm must continuously and concertedly redesign Purpose, Business, Organization, Capital, and Leadership in line with the future to be realized. This is Enterprise Redefinition (Kadowaki, 2026b). The Purpose Description translates that firm-level redefinition to the unit of teams and individuals. The theoretical chain runs as follows. Markets and society form expectations about the future. The firm reads those expectations and simultaneously envisions a desirable future. Enterprise Redefinition redesigns the firm as a whole. The Purpose Description defines each individual’s and team’s raison d’être and contribution to future value. Humans and AI reconfigure the means and act. Realized outcomes and learning form new expectations. Within this chain, Future Value Theory explains “what constitutes firm value,” Enterprise Redefinition explains “how the firm as a whole is changed,” and the Purpose Description explains “who carries future value, and for what purpose.” In sum: the Job Description allocates present tasks; the Purpose Description distributes the creation of future value. 5.1 From Corporate Purpose to Frontline PDs: The Query Translation Protocol When the directions indicated by corporate purpose, Future Value Theory, and Enterprise Redefinition are connected to the PDs of individuals and teams, a rupture of translation arises. Handing down a highly abstract corporate purpose verbatim leaves it a symbolic slogan; conversely, rushing to concretize it into existing tasks and KPIs causes the vision of future value to regress into a static JD. This paper presents the Query Translation Protocol as the core mechanism for closing this Abstraction Gap. Basic principle: deploy questions, not tasks. The minimal unit that connects the future value sought by Enterprise Redefinition to the front line is not work instructions or numerical targets but the inheritance and translation of the core questions to be solved. Upper levels do not fix answers or means; From Job Description to Purpose Description VURA Working Paper Series 13 they present future value hypotheses, strategic intent, decision principles, and the boundaries that must not be crossed. Lower levels use customer contact, expert knowledge, frontline data, and AI capabilities to translate those questions into more specific, testable Defining Questions. This is therefore a cascade of future value hypotheses and questions, not a cascade of KPIs. The three-level structure of query translation. First, the Corporate Level Query asks: “In the age of AI, what essential value and desirable future does our firm generate for markets and society?” This links the market and stakeholder expectations captured by Future Value Theory to the desirable future that management chooses. Second, the Division / Strategic Level Query asks: “To realize that future, which customer, societal, and organizational constraints must this business or function remove?” This converts the redesign of Purpose, Business, Organization, Capital, and Leadership in Enterprise Redefinition into value hypotheses specific to each business and function. Third, the Team / Individual Defining Questions ask: “To remove those constraints, whom do we serve as beneficiaries, under which decision principles do we mobilize humans and AI, and what must we keep asking?” This final translation is written into the six components of the PD. These three levels also constitute a hierarchical deployment of value hypotheses running from Corporate Future Value through Business / Strategic Future Value and Capability / Platform Future Value to the Team / Individual PD. Each level, however, is not a scaled-down copy of the wording above it. While preserving the intent and constraints expressed by the superordinate question, each lower level contextualizes them into beneficiaries, constraints to be removed, desired states, and testable questions. Bidirectional governance: Top-down Guidance × Bottom-up Refinement. Management presents the corporate Future Value Hypothesis, Strategic Priorities, Decision Principles, ethical, safety, and legal boundaries, and resource-allocation priorities. The front line, drawing on evidence from the customer, technology, and operational frontier, concretizes the superordinate questions and, where necessary, returns refutations of and proposed revisions to the superordinate value hypotheses themselves. This bidirectionality is consistent with the classic finding that resource allocation proceeds through the interaction of frontline initiatives and the structural context set by management (Bower, 1970). Management and leaders synchronize periodically with the front line on whether frontline PD questions respond most appropriately to superordinate strategic intent and, at the same time, whether new frontline learning should update the superordinate questions. Approval is not formal assent to wording but an Endorsement of the logical coherence of questions, decision principles, beneficiaries, and future value. From Job Description to Purpose Description VURA Working Paper Series 14 Connection to AI allocation. Query translation also precedes the allocation of roles between humans and AI. After management identifies the Commoditized Execution to be consolidated and automated with AI, it presents as a question which of the untapped future values sought by Enterprise Redefinition the freed time and cognitive resources should be reinvested in. Teams translate that question into their PDs and redesign the human roles of customer understanding, normative judgment, exception handling, and final accountability, and the AI roles of exploration, generation, analysis, and execution. Workload reduction is not the endpoint; it becomes an intermediate outcome that enables the movement of resources toward the exploration of new value. Criteria for translation quality. A translated PD is evaluated against five criteria: (1) its causal connection to superordinate future value can be explained; (2) it identifies concrete beneficiaries and the constraints to be removed; (3) it takes the form of explorable questions that fix neither answers nor tasks; (4) it preserves decision principles and responsibility boundaries; and (5) its questions can be updated in light of outcomes and learning. These criteria make it possible to reconcile managerial intent with frontline autonomy while curbing both the hollowing-out of purpose and passive waiting for instructions. Figure 2. The Bidirectional Flow of the Query Translation Protocol Note: This is not a one-way goal cascade. Upper levels do not fix answers or means, and lower levels return evidence and refutation. Updates are limited to deltas and exceptions so that waiting for approval does not become the normal state. Downward translation Upward refutation and learning 1 Management: presents future value hypotheses, beneficiaries, and inviolable principles Reflects market and societal change into hypotheses 2 Divisions: translate into queries that release strategic constraints Surface cross-unit contradictions and resource constraints 3 Team PD: sets verifiable Defining Questions and decision principles Records customer evidence, AI capability, exceptions, and failures in the decision log 4 Connects to OKRs, experiments, and JD boundaries for execution Maintains, revises, or rejects hypotheses; updates the PD and strategy From Job Description to Purpose Description VURA Working Paper Series 15 6 The Purpose Description in AI Adoption 6.1 Three Kinds of Success in AI Adoption Success in AI adoption is not singular. The first kind is technical success: the model or system operates as required and meets standards for accuracy, speed, availability, and safety. The second is operational success: improvements in time, cost, quality, throughput, employee experience, and the like. The third is success in terms of future value: creating new value for customers, the firm, and society, and raising expectations of—and capabilities for—the future the firm seeks to realize. Even without a PD, technical or operational success can sometimes be achieved in limited routine operations. This paper therefore does not claim that “with a PD, AI adoption succeeds.” The claim is a falsifiable conditional relationship: the structural clarity and alignment of the PD raise the probability that operational outcomes are connected to future-value outcomes, mediated by dynamic role redesign, human-AI reallocation, and psychological ownership. 6.2 The Mechanism of Purposeless Means Optimization AI adoption without a Purpose Description usually begins with the question “which of our current operations can be automated with AI?” This starting point is useful for discovering candidates, but it does not ask whether the existing operation is truly necessary, whose problem is being solved, what the desirable customer experience is, or which judgments and responsibilities humans should retain. The possible consequences include the acceleration of unnecessary processes, divergence between departmental KPI gains and customer value, non-use in the field, blurred responsibility boundaries, and a bias toward measurable workload reduction. With a Purpose Description, the order is reversed. First, the future to be realized, the beneficiaries, the future value, the questions, and the decision principles are defined. Next, the judgments, actions, and capabilities that the purpose requires are clarified. On that basis, allocation follows: to humans, purpose setting, exception judgment, ethics, supervision, halting, and accountability; to AI, exploration, generation, prediction, and automated execution. AI can be an executing partner (Co-agent) or a co-creator (Co-creator) of future value, but it does not become an accountable agent. Finally, the necessary business processes and systems are designed. AI is then not the starting point but a means selected for the realization of purpose. Failures of AI adoption therefore do not arise solely from insufficient AI performance. They also begin from the absence of a defined future to be realized through AI. AI adoption without a Purpose Description optimizes means without defining purpose. From Job Description to Purpose Description VURA Working Paper Series 16 6.3 Relation to Human-on-the-Loop Operation The Purpose Description does not presuppose only Human-in-the-Loop operation, in which humans approve every step of processing. It is consistent with Human-on-theLoop operation, in which AI executes autonomously within a defined scope while humans own purpose, boundaries, evaluation criteria, exceptions, and redesign. Here execution and accountability are distinguished. AI can execute tasks and create future value together with humans, but it cannot be an agent that assumes legal or ethical responsibility. The further humans move from direct execution, the more clearly their roles must be defined as deciders of purpose, supervisors of the system, final judges in exceptions, holders of halt authority, and ultimate bearers of responsibility for outcomes. 7 Conceptual Model and Theoretical Propositions The paper’s conceptual model posits a mediating chain: exogenous changes in AI capability and task fluidity operate through the design quality of PDs and the translation of questions to generate shared purpose understanding, psychological ownership, and human-AI role reallocation, which in turn affect customer value, organizational capability, and Future Value. These effects are not automatic. Psychological safety, interdepartmental trust, leaders’ cognitive capabilities, clarity of decision rights, data and IT infrastructure, and institutional fairness are made explicit as moderators or confounders, and empirical studies should control for organization size, industry, AI investment intensity, initial productivity, and digital maturity. 7.1 Propositions P1 (Environmental fit). The faster the rate of change in AI’s potential to substitute for and complement tasks, the faster the content validity of roles defined solely by static JDs declines over time, even controlling for initial digital maturity and AI investment intensity. P2 (Dynamic reallocation). The structural clarity of the PD increases the speed of role reallocation between humans and AI, mediated by shared purpose understanding. This relationship is stronger where data and IT infrastructure and employees’ AI literacy exceed a threshold level. From Job Description to Purpose Description VURA Working Paper Series 17 P3 (Vertical alignment). Query translation fidelity enhances Future Value outcomes, mediated by vertical alignment across the corporate, division, and team levels. This effect is stronger where leaders’ cognitive capabilities are high and institutions exist for accepting refutation from the front line. P4 (Horizontal alignment). Horizontal alignment of PDs across teams increases the interdepartmental sharing of AI and data assets. This relationship is stronger where interdepartmental trust and data governance are high, and weaker where resource competition is intense. P5 (Self-regulating capability). Under high environmental uncertainty, PD clarity increases the speed and quality of decentralized role redesign. This effect holds, however, only where decision rights, psychological safety, and escalation boundaries are clear. P6 (Psychological ownership). Substantive participation in PD formulation increases psychological ownership, mediated by perceived procedural fairness. Where job-loss anxiety is high, the effect disappears or reverses if participation is merely formal. P7 (Shift in evaluation axes). Organizations that incorporate the PD into their appraisal systems increase—relative to their pre-adoption state and to comparable control organizations—the proportional use of customer-value, learning, and capability-building indicators beyond workload reduction. This effect is stronger where executive compensation design and measurement capability are aligned. P8 (Curbing low-value automation). Controlling for digital maturity, IT literacy, process standardization, and management capability, organizations using the PD and the Query Translation Protocol exhibit lower continuation rates of automation projects that do not contribute to beneficiary value than organizations relying solely on goal and task management. This difference is mediated by the frequency of hypothesis-rejection and resource-reallocation decisions; it is stronger the higher the psychological safety and non-punitive learning from failure, and weaker the stronger the demerit-oriented appraisal and sunk-cost pressure. From Job Description to Purpose Description VURA Working Paper Series 18 P9 (Institutional boundary conditions). The positive relationship between PD use and organizational outcomes holds where beneficiaries, decision principles, decision rights, JD-level responsibility standards, intervention protection, and appeal procedures are made explicit. The relationship is stronger the higher the psychological safety, procedural fairness, and non-retaliation norms; where excessive demerit-oriented appraisal, steep authority gradients, and responsibility-authority misalignment prevail, it turns null or negative through role ambiguity, silence, and unbounded role expansion. 8 Writing and Implementing the Purpose Description 8.1 A Descriptive Template A Purpose Description must be a short statement usable in decision making, not an abstract slogan. The recommended basic template is as follows. We exist to enable [beneficiaries] to realize [the desired future / change]. We create [the future value to be created] and contribute to [the purpose and future value of the firm as a whole]. To that end we ask [the questions we continuously pose] and, guided by [decision principles], continuously redesign the combination of humans, AI, and external resources. Performance is evaluated by [future value indicators] and [the necessary operational and control indicators]. 8.2 Implementation Process Implementation proceeds in six stages. First, articulate the expectations of markets and society and the future the firm seeks to realize as a Corporate Level Query. Second, translate that query into Division / Strategic Level Queries asking which customer, societal, and organizational constraints each business or function must remove. Third, teams—drawing on beneficiaries, frontline data, expert knowledge, and AI capabilities —translate the superordinate questions into concrete, testable Defining Questions and draft their Purpose Descriptions dialogically. Fourth, connect each individual’s Purpose Description to the team’s questions. Fifth, dynamically allocate the activities the purpose requires among humans, AI, and external partners, updating Job Descriptions and the associated authority and responsibility. Sixth, feed outcomes, market responses, and frontline learning back to the upper levels, periodically redefining the questions and Purpose Descriptions at each level.