Theories and Models of Organizational Change

Introduction and Context

Organizational change refers to a significant alteration in an organization’s strategy, structure, technology, processes, culture, capabilities, or relationships with stakeholders. Change may be planned, emergent, incremental, transformational, internally driven, or imposed by external conditions. Organizational development (OD) is concerned not merely with changing structures but with improving organizational effectiveness, human well-being, adaptability, and the quality of relationships through behavioural-science-informed interventions.

Change models are useful because they provide language, sequencing, diagnostic questions, and intervention choices. However, no model is universally valid. A model is a conceptual lens rather than a mechanical recipe. Its usefulness depends on the nature of the change, the degree of uncertainty, organizational power relations, stakeholder readiness, cultural context, and the quality of evidence supporting decisions.

For HR professionals, this distinction is central. HR may be responsible for workforce planning, leadership development, communication, consultation, capability building, employee relations, and evaluation. These responsibilities require more than selecting a fashionable framework. They require critical judgement about which assumptions fit the situation and how a model should be adapted.

Learning Objectives

By the end of this chapter, readers should be able to:

1. Explain the assumptions and stages of major organizational change models.
2. Compare planned, participative, individual, interpretive, institutional, and complexity-based perspectives.
3. Evaluate the empirical and practical strengths and limitations of each approach.
4. Select or combine models according to context, evidence, ethics, and implementation requirements.
5. Apply change theory to diagnose and manage a complex organizational situation.

1. How to Understand a Change Model

A change model normally answers four questions:

• What is changing? This may involve behaviour, routines, meaning, power, systems, or institutional legitimacy.
• Why does change occur? Explanations may emphasize managerial intention, learning, crisis, social influence, identity, or environmental selection.
• How does change unfold? It may occur in stages, cycles, episodes, negotiations, or unpredictable interactions.
• How should change be evaluated? Possible criteria include performance, adoption, employee well-being, equity, sustainability, and legitimacy.

Models differ in their underlying assumptions. A planned-change model assumes that a purposeful agent can diagnose a problem, design an intervention, and guide movement toward a desired state. An emergent model assumes that change develops through experimentation, local adaptation, sensemaking, and interaction. A political perspective emphasizes competing interests and power. An institutional perspective asks why organizations adopt similar practices, sometimes for legitimacy rather than efficiency.

A sophisticated change diagnosis should therefore distinguish between:

• Change content: what is changing;
• Change process: how change is implemented;
• Change context: the internal and external conditions shaping it; and
• Change outcomes: intended and unintended consequences.

2. Lewin’s Three-Step Model

Kurt Lewin’s model is commonly expressed as unfreezing, changing (or moving), and refreezing. It emerged from field theory and group dynamics, which view behaviour as a function of the person and the environment. The model is best understood as a high-level logic of transition rather than a detailed project plan.

Unfreezing

Unfreezing involves weakening the forces that maintain existing behaviour and creating readiness for alternatives. It may include establishing a case for change, surfacing dissatisfaction, challenging assumptions, building psychological readiness, and involving affected groups. Effective unfreezing is not equivalent to creating fear. Ethical unfreezing uses credible information and meaningful participation rather than manipulation or manufactured crisis.

Changing or moving

The movement stage involves developing and trying new behaviours, structures, processes, or meanings. Learning, experimentation, training, coaching, participation, and revised systems can support this stage. Change is often non-linear: individuals may move forward, regress, or adopt different aspects at different speeds.

Refreezing

Refreezing means stabilizing useful new patterns through reinforcement, systems alignment, role clarity, leadership behaviour, recruitment, rewards, measurement, and social norms. The concept does not require permanent rigidity. In a volatile environment, refreezing may mean creating a temporary, sufficiently stable platform from which further adaptation can occur.

Critical evaluation

Lewin’s strengths include simplicity, attention to group forces, and recognition that lasting change requires reinforcement. Its limitations arise when it is interpreted as a universal linear sequence. Contemporary organizations often experience continuous change, making a fully frozen state unrealistic. Moreover, the original field-theory tradition is richer than the three-word summary often presented in textbooks. The model is most useful as a diagnostic scaffold and least useful when treated as a rigid timetable.

3. Action Research

Action research is both a change methodology and a way of producing knowledge. It links diagnosis, action, and learning through collaboration between organizational members and a researcher, consultant, or change practitioner. A typical cycle includes diagnosing, planning action, implementing action, evaluating results, and learning or redefining the problem.

The process is iterative rather than strictly sequential. Data may include interviews, observation, surveys, operational indicators, focus groups, and documentary analysis. The credibility of the process depends on transparent methods, multiple perspectives, protection of confidentiality, and willingness to revise initial assumptions.

Action research has several distinctive features:

• Participation: those affected contribute to diagnosis and intervention.
• Reflexivity: participants examine how their assumptions and actions shape the situation.
• Iteration: findings from one cycle inform the next.
• Practical and scholarly value: the process seeks improvement while generating transferable learning.
• Power awareness: the practitioner must ask who defines the problem, whose evidence counts, and who benefits.

Action research is especially suitable where the problem is poorly understood, local knowledge is important, and commitment must be developed through involvement. It can be slower than directive implementation, and participation can be symbolic if leaders have already made all substantive decisions. It also requires careful research ethics, including informed consent where appropriate, secure data handling, non-retaliation, and clarity about the practitioner’s role.

4. Kotter’s Eight-Step Process

John Kotter’s process model describes eight broad activities:

1. Create a sense of urgency.
2. Build a guiding coalition.
3. Form a strategic vision and initiatives.
4. Enlist a volunteer army or communicate and mobilize broadly.
5. Enable action by removing barriers.
6. Generate short-term wins.
7. Sustain acceleration.
8. Institute change.

The model emphasizes leadership alignment, broad mobilization, visible progress, and cultural embedding. It is particularly useful for large-scale strategic initiatives in which senior sponsorship and cross-functional coordination are important.

The model’s practical value lies in its attention to often-neglected political and motivational conditions. A technically sound redesign can fail if leaders disagree, middle managers resist, barriers remain in performance systems, or employees cannot see credible progress.

Critiques concern its apparent linearity and top-down emphasis. Change may not begin with urgency, and a guiding coalition may emerge after experimentation rather than before it. “Short-term wins” can also encourage superficial metrics or favour visible results over patient capability building. The process should therefore be treated as a set of mutually reinforcing activities, not a compulsory sequence.

5. ADKAR

ADKAR, associated with Prosci, focuses on individual adoption. Its elements are:

• Awareness of the need for change;
• Desire to support and participate in it;
• Knowledge of what to do and how to do it;
• Ability to perform the new behaviour in practice; and
• Reinforcement to sustain the change.

The model helps managers distinguish different barriers. An employee may understand a new system but lack the desire to use it; may desire change but lack knowledge; or may have knowledge but be prevented by inadequate time, tools, authority, or managerial support.

ADKAR is valuable for designing communication, training, coaching, manager support, and adoption measures. It is particularly useful in technology implementation and role transitions. Its individual focus is also a limitation: collective power, organizational politics, structural constraints, institutional pressures, and conflicting interests may not be adequately explained. Resistance should not automatically be interpreted as an individual deficit. Sometimes it is a rational response to unfair risk, poor design, or lack of voice.

6. Bridges’ Transition Model

William Bridges distinguished change, an external event or situational alteration, from transition, the internal psychological process through which people come to terms with change. His three phases are:

1. Ending, losing, and letting go: people disengage from old roles, routines, identities, or relationships.
2. The neutral zone: old arrangements have weakened while new ones are not yet fully established. This can be confusing but also creative.
3. The new beginning: people develop new identities, purposes, and patterns of action.

This model is useful when the emotional and identity dimensions of change are underestimated. HR practitioners can support endings through honest communication and acknowledgement of loss; support the neutral zone through temporary structures, experimentation, and frequent information; and support new beginnings through role clarity, symbolic recognition, capability building, and visible leadership.

The model’s language can be culturally variable. Not all people experience transition in the same order or with the same emotions. It may also understate material interests and organizational power. It should complement, rather than replace, analysis of systems, resources, and decision rights.

7. Appreciative Inquiry

Appreciative inquiry (AI) is a strengths-based approach that studies what gives life to an organization when it is functioning well and uses those insights to imagine and create a preferred future. Its familiar 4-D cycle is:

• Discover: identify strengths, achievements, and generative conditions;
• Dream: imagine what could be;
• Design: construct structures and practices that support the aspiration; and
• Destiny (or deliver): learn, adapt, and sustain the new patterns.

AI challenges deficit-only diagnosis. It can increase engagement, surface tacit capabilities, and create an affirmative shared narrative. It is useful in culture work, team development, service improvement, and settings where stakeholders are exhausted by problem-focused change.

A critical use of AI does not deny problems. Focusing only on positive stories may silence people experiencing discrimination, conceal risk, or produce unrealistic optimism. The method should include disconfirming evidence, marginalized voices, and safeguards against using “positivity” to avoid accountability.

8. Punctuated Equilibrium

The punctuated-equilibrium perspective proposes that organizations may experience long periods of relative stability interrupted by brief episodes of substantial transformation. Stability is maintained by established routines, power arrangements, cognitive frames, and resource patterns. A crisis, technological shift, leadership change, regulatory disruption, or accumulated contradiction may destabilize the existing equilibrium.

This perspective helps explain why organizations can ignore weak signals for years and then change rapidly. It is useful for strategic analysis, organizational ecology, and understanding episodic transformation. It cautions managers against assuming that gradual improvement is always possible or that a crisis can be deliberately manufactured without ethical consequences.

The model is descriptive rather than a complete intervention guide. It does not specify how leaders should manage the transition, and empirical patterns vary across industries and organizational levels. Some organizations show continuous adaptation while others display episodic change; the appropriate conclusion must be based on evidence rather than the label alone.

9. Institutional Theory

Institutional theory explains how organizations are shaped by regulative, normative, and cultural-cognitive forces. Organizations may adopt practices because laws require them, professional communities view them as appropriate, or widely shared assumptions make them appear natural. DiMaggio and Powell identified three forms of institutional isomorphism:

• Coercive isomorphism: pressure from governments, regulators, funders, or powerful stakeholders;
• Mimetic isomorphism: imitation under uncertainty; and
• Normative isomorphism: influence from professions, education, and standards.

Institutional theory distinguishes technical efficiency from legitimacy. An organization may adopt a practice to appear modern, responsible, or compliant even when implementation is partial. This can produce decoupling, where formal policies differ from everyday practice.

For HR, institutional analysis is important in areas such as diversity policies, competency frameworks, hybrid-work practices, sustainability reporting, and algorithmic recruitment. It asks whether a change is genuinely embedded or merely ceremonially adopted. The perspective also highlights that organizations are not free actors: their choices are constrained and enabled by institutional environments.

10. Sensemaking

Sensemaking concerns how people interpret ambiguous events and construct plausible accounts that guide action. Karl Weick’s work emphasizes that sensemaking is ongoing, social, retrospective, identity-laden, and enacted: people interpret situations partly through the consequences of their own actions.

During change, employees ask: What is happening? Why is it happening? What does it mean for my identity and work? Who can be trusted? What should I do next? Formal communication is only one source of meaning. Informal conversations, leader behaviour, peer interpretation, symbols, and visible resource allocations also shape sensemaking.

A sensemaking approach encourages leaders to provide honest provisional explanations, create forums for dialogue, acknowledge ambiguity, and listen for competing interpretations. It rejects the assumption that resistance is simply irrational opposition. Resistance may indicate that the official story is not credible or that consequences have not been addressed.

Sensemaking is interpretive rather than predictive. It can explain how meanings develop but offers less precise guidance about which interpretation will prevail. It must also be connected to material realities: communication cannot compensate indefinitely for job insecurity, workload overload, or inequitable decisions.

11. Complexity-Based Approaches

Complexity-based approaches view organizations as adaptive systems composed of interconnected agents whose interactions produce outcomes that cannot be fully predicted from individual intentions. Change may emerge through local experimentation, feedback, self-organization, network effects, and adaptation.

A complexity-informed practitioner typically:

• establishes a clear direction without prescribing every action;
• enables small, safe-to-fail experiments;
• monitors patterns and unintended consequences;
• connects diverse actors and information; and
• adapts interventions as conditions change.

This approach is valuable where cause-and-effect relationships are uncertain, multiple stakeholders interact, and local knowledge matters. It is less appropriate as an excuse for managerial inaction. Complexity does not eliminate accountability, governance, prioritization, or ethical obligations. Leaders still decide resource allocations, acceptable risk, and non-negotiable standards.

Complexity approaches also challenge conventional evaluation. A single before-and-after measure may not explain what produced an outcome. Mixed methods, contribution analysis, network evidence, qualitative learning, and repeated feedback may be more appropriate.

12. Comparative Analysis and Model Selection

The models differ in the level and unit of analysis they privilege.

Model selection should begin with diagnosis, not preference. The following questions are useful:

1. Is the desired outcome known, or must it emerge through inquiry?
2. Is the main barrier individual capability, collective meaning, political conflict, institutional legitimacy, or system interdependence?
3. Is the environment relatively stable or highly turbulent?
4. Is participation ethically and practically necessary?
5. What evidence will show adoption, effectiveness, equity, and sustainability?
6. Which stakeholders bear the costs and which receive the benefits?

Models can be combined, but combination should be theoretically coherent. For example, an organization may use action research to diagnose a service problem, Kotter-like coalition work to mobilize leaders, ADKAR to design role-based support, Bridges to address identity loss, and complexity principles to run local experiments. This is not a claim that all models are simultaneously true in every respect; it is a division of analytical labour across levels and questions.

13. Evidence, Evaluation, and Ethical Governance

Change evaluation should avoid relying only on completion of activities. A training programme may be delivered without producing capability; a new policy may be announced without changing practice; and a short-term performance gain may conceal burnout or inequity.

A balanced evaluation can examine:

• Implementation: Was the intervention delivered as intended?
• Adoption: Are people using the new practice?
• Outcomes: Did performance, service, quality, or safety improve?
• People effects: What happened to workload, inclusion, well-being, and perceived fairness?
• Sustainability: Does the change persist and adapt appropriately?
• Distribution: Who gains, who loses, and whose voice is missing?

Evidence should be triangulated across quantitative indicators, qualitative accounts, observation, and relevant comparison points. HR analytics can identify patterns, but correlation does not establish causation. Employee surveys require attention to sampling, response bias, confidentiality, and psychological safety.

Ethically responsible change requires truthful communication, meaningful consultation where appropriate, respect for dignity, protection from retaliation, fair selection criteria, accessibility, and attention to disparate impacts. Digital and AI-enabled change adds concerns about privacy, explainability, bias, surveillance, data quality, and human oversight. An algorithmic recommendation should not be treated as neutral merely because it is computational. Governance should specify accountability, appeal routes, audit arrangements, and limits on automated decisions.

14. Applied Case: Introducing an AI-Assisted Recruitment System

This is an illustrative case. Meridian Services, a 2,500-employee organization, wants to introduce an AI-assisted tool to screen applications and schedule interviews. The business case claims faster recruitment and more consistent shortlisting. Recruiters are concerned about job displacement, candidates question fairness, and employee representatives request evidence that the system will not reproduce historical bias.

Background and problem

The initial proposal treats the project as a technology installation. Early pilot data show reduced processing time, but shortlisted candidates from some demographic groups decline disproportionately. Recruiters also report that the system’s recommendations are difficult to explain.

Analysis through multiple models

A Lewinian diagnosis identifies forces supporting change—volume, vacancy costs, and service-level pressure—and restraining forces—low trust, weak explainability, capability gaps, and perceived threat to professional judgement.

Action research suggests involving recruiters, hiring managers, candidates, accessibility specialists, and employee representatives in diagnosis and iterative testing. Their participation may reveal problems not visible in aggregate efficiency data.

Kotter’s perspective highlights the need for a credible coalition involving HR, technology, legal and compliance specialists, operational leaders, and employee representatives. A short-term win should not be defined only as faster processing; it should include demonstrably safe and fair use.

ADKAR distinguishes awareness of the business problem from desire to use the tool, knowledge of appropriate interpretation, ability to challenge or override recommendations, and reinforcement through audits and manager accountability.

Bridges draws attention to the ending of familiar recruiter expertise, the uncertain neutral zone during piloting, and the need to construct a new professional identity in which recruiters become accountable users and reviewers of decision-support technology.

Institutional theory asks whether adoption is driven by genuine improvement or imitation of competitors. Sensemaking analysis examines whether employees interpret the tool as assistance, surveillance, or covert downsizing. A complexity approach recommends controlled experiments across different recruitment contexts rather than a single organization-wide launch.

Approach and solution

Meridian pauses full deployment and establishes governance rules: human decision authority remains mandatory; candidates receive an accessible explanation of relevant process; data use is limited and documented; performance is monitored for subgroup disparities; recruiters can override recommendations with reasons; and an independent review route is established. The organization runs small pilots, compares results with existing practice, gathers qualitative feedback, and publishes limitations.

Training addresses both technical use and ethical judgement. Recruiters participate in redesign workshops. Evaluation covers time-to-shortlist, quality of hire, candidate experience, adverse-impact indicators, override patterns, and employee workload. Results are reviewed at regular intervals, and deployment is withdrawn if agreed safety thresholds are not met.

Outcome

The revised approach produces a smaller efficiency gain than the original business case, but it improves transparency, increases recruiter trust, and identifies a data-quality problem that would otherwise have been scaled. The system is retained for administrative support and structured information retrieval, while final screening remains human-led pending further evidence.

Learning points

1. A technology change is also a change in identity, power, governance, and institutional legitimacy.
2. Faster implementation is not equivalent to successful change.
3. Combining models can clarify different levels of the same case.
4. Evidence must include distributional and ethical outcomes, not only average productivity.
5. Experiments and reversibility are valuable where consequences are uncertain.

15. Best Practices for HR Change Practitioners

First, diagnose before selecting a model. Second, make assumptions explicit: what is believed about people, power, learning, and predictability? Third, distinguish communication from participation; informing people is not the same as sharing decision authority. Fourth, align formal systems—rewards, recruitment, performance management, technology, and leadership routines—with the desired change. Fifth, treat resistance as data as well as an obstacle. Sixth, protect dissent and test whether benefits and burdens are fairly distributed. Seventh, use staged experimentation where uncertainty is high. Finally, evaluate both intended outcomes and unintended consequences, and revise the intervention when evidence warrants it.

Chapter Summary

Organizational change theories provide different explanations of how and why change occurs. Lewin offers a broad logic of readiness, movement, and stabilization; action research links participation, intervention, and learning; Kotter emphasizes coalition-building, mobilization, and institutionalization; and ADKAR focuses on the conditions of individual adoption. Bridges highlights psychological transition, while appreciative inquiry develops change from strengths and generative possibilities.

Punctuated equilibrium explains episodic transformation after periods of stability. Institutional theory examines coercive, mimetic, and normative pressures and the distinction between legitimacy and technical effectiveness. Sensemaking focuses on how people interpret ambiguity and construct plausible action. Complexity-based approaches emphasize emergence, interdependence, experimentation, and adaptation.

No model should be applied mechanically. Selection should follow diagnosis of the change content, context, process, stakeholders, uncertainty, power relations, and ethical risks. Models may be combined when their distinct purposes are clear. Evidence should assess implementation, adoption, outcomes, well-being, equity, and sustainability. For HR professionals, effective change is therefore both a managerial and a governance responsibility.

Key Terms

Action research: A participative, iterative process linking diagnosis, action, evaluation, and learning.

ADKAR: A model of individual change involving Awareness, Desire, Knowledge, Ability, and Reinforcement.

Appreciative inquiry: A strengths-based approach that develops change by studying and extending what works well.

Change content: The substance of what is altered in an organization.

Change context: The internal and external conditions influencing change.

Change process: The activities and interactions through which change unfolds.

Complex adaptive system: An interconnected system in which interactions produce emergent and partly unpredictable patterns.

Decoupling: A gap between formal organizational structures or policies and actual practice.

Institutional isomorphism: The tendency of organizations to become similar through coercive, mimetic, or normative pressures.

Legitimacy: Social acceptance that an organization or practice is appropriate or proper.

Neutral zone: Bridges’ transitional period between old arrangements and new beginnings.

Punctuated equilibrium: A pattern of long stability interrupted by relatively brief episodes of major change.

Refreezing: Stabilizing and reinforcing useful new patterns after movement.

Sensemaking: The social process of interpreting ambiguous events and constructing meaning for action.

Short-term win: A visible, credible early result used to build confidence and momentum.

Transition: The psychological process of adapting to an external change.

Unfreezing: Creating readiness by loosening attachment to existing patterns.

Safe-to-fail experiment: A bounded intervention designed to generate learning without exposing the whole system to unacceptable risk.

Review Questions

Short Answer Questions

1. What are the three stages in Lewin’s model, and what is the purpose of each?
2. How does action research differ from a one-time organizational diagnosis?
3. What are the five elements of ADKAR?
4. Distinguish change from transition in Bridges’ model.
5. What are coercive, mimetic, and normative institutional pressures?

Long Answer Questions

1. Critically evaluate the usefulness and limitations of Lewin’s three-step model in a continuously changing organization.
2. Compare Kotter’s process with action research in relation to participation, leadership, and learning.
3. Explain how sensemaking and Bridges’ transition model can improve the human management of organizational change.
4. Discuss the contribution of institutional theory to understanding why organizations adopt HR innovations.
5. Compare appreciative inquiry and complexity-based approaches, including situations in which each may be inappropriate.

Analytical / Application Questions

1. A hospital is introducing electronic records under a strict regulatory deadline, while clinicians distrust the system and workload is already high. Select and justify a combination of change models for diagnosis, implementation, and evaluation.
2. An organization reports that a diversity policy has been fully implemented, but employee interviews reveal little change in promotion decisions. Analyze the situation using institutional theory, sensemaking, and action research.
3. Design an evidence and governance plan for introducing an AI-enabled HR system. Specify how you would evaluate adoption, effectiveness, fairness, employee well-being, and sustainability.

Multiple Choice Questions with Answers

Question 1

Which statement best captures the main limitation of treating Lewin’s model as a rigid prescription?

A. It ignores the importance of organizational groups entirely.

B. It may imply a linear sequence and a permanent frozen state in environments characterized by continuous change.

C. It focuses exclusively on individual training.

D. It rejects the need for reinforcement after change.

Correct Answer: B. It may imply a linear sequence and a permanent frozen state in environments characterized by continuous change.

Question 2

What is distinctive about action research?

A. It separates diagnosis completely from intervention.

B. It relies only on numerical performance indicators.

C. It combines participation, action, evaluation, and iterative learning.

D. It requires consultants to make all decisions independently.

Correct Answer: C. It combines participation, action, evaluation, and iterative learning.

Question 3

Which ADKAR element concerns the capacity to perform a new behaviour in practice?

A. Awareness

B. Desire

C. Knowledge

D. Ability

Correct Answer: D. Ability

Question 4

In Bridges’ transition model, the neutral zone is best described as:

A. The period before leaders announce change.

B. The period when old arrangements have weakened but new ones are not yet fully established.

C. The final stage in which all change stops permanently.

D. A formal legal review of the change.

Correct Answer: B. The period when old arrangements have weakened but new ones are not yet fully established.

Question 5

Which activity is most closely associated with Kotter’s process?

A. Building a guiding coalition and generating short-term wins.

B. Measuring only individual psychological readiness.

C. Rejecting all planned communication.

D. Assuming that institutions exert no influence.

Correct Answer: A. Building a guiding coalition and generating short-term wins.

Question 6

What is a central risk of using appreciative inquiry uncritically?

A. It necessarily prevents employee participation.

B. It may use positive narratives to avoid difficult problems or silence marginalized voices.

C. It cannot identify organizational strengths.

D. It requires a stable environment before any inquiry begins.

Correct Answer: B. It may use positive narratives to avoid difficult problems or silence marginalized voices.

Question 7

Mimetic institutional isomorphism occurs when organizations:

A. Adopt practices because legislation requires them.

B. Adopt practices through professional norms and education.

C. Imitate other organizations, especially under uncertainty.

D. Reject practices that competitors have adopted.

Correct Answer: C. Imitate other organizations, especially under uncertainty.

Question 8

A sensemaking perspective would most strongly encourage leaders to:

A. Provide a single explanation and discourage dialogue.

B. Treat all resistance as irrational.

C. Acknowledge ambiguity, create dialogue, and examine competing interpretations.

D. Rely exclusively on formal policy documents.

Correct Answer: C. Acknowledge ambiguity, create dialogue, and examine competing interpretations.

Question 9

Which intervention best reflects a complexity-based approach?

A. A fixed organization-wide sequence with no adaptation.

B. Small bounded experiments, feedback, monitoring, and local adaptation.

C. A decision to eliminate all governance.

D. A one-time communication event followed by no evaluation.

Correct Answer: B. Small bounded experiments, feedback, monitoring, and local adaptation.

Question 10

In the illustrative AI recruitment case, why was human decision authority retained?

A. Because technology can never assist with recruitment.

B. To preserve accountability, allow professional judgement, and address fairness and explainability risks.

C. Because efficiency was the only relevant outcome.

D. To ensure that no data would be collected.

Correct Answer: B. To preserve accountability, allow professional judgement, and address fairness and explainability risks.