← Papers

Treat analytics as the interpretive bridge: a five-layer Human-AI decision framework argues that combining AI outputs with business-analytics interpretation, managerial judgment and feedback loops can raise decision quality, reduce algorithmic bias and boost employee confidence; however, the framework is conceptual and requires empirical validation to quantify productivity and labor impacts.

Designing Human–AI Collaborative Decision Analytics Frameworks to Enhance Managerial Judgment and Organizational Performance

Jannatul Ferdousi, Md Shokran, Md Saiful Islam · Fetched March 15, 2026 · Journal of business and management studies

semantic_scholar theoretical low evidence 7/10 relevance DOI Source PDF

The paper proposes a five-layer Human-AI Collaborative Decision Analytics Framework that positions analytics as the interpretive bridge between AI outputs and managerial judgment to improve decision quality, accountability, and workforce confidence while mitigating automation bias.

The fast spread of artificial intelligence (AI) in the United States organizations has radically altered the managerial decision-making process, but on the other hand, it has augmented the complexity and uncertainty in decision, and the accountability stresses. Despite the high-level predictive and prescriptive potentials of AI-based analytics, most organizations have difficulties converting algorithmic results into sustainable managerial decisions. Low levels of trust, lack of explanation, and poor integration between AI systems and human judgment have been caused by over reliance on automation, weak explain ability, and poor organizational outcomes. Current literature has majorly focused on automation-based views of decision support, with a severe lack of insight into the coordinated manner in which human experience and AI intelligence can be systematically integrated with the assistance of analytics. This paper fills this gap by outlining a Human-AI Collaborative Decision Analytics Framework that could be beneficial to improve managerial decisions and organizational performance. Following a conceptual research design, the study integrates interdisciplinary literature in the field of managerial decision-making theory, business analytics, and governance of AI in its attempt to establish an integrative framework where analytics becomes the focal interpretive intercession between AI outputs and human decision-makers. The framework has five overlapping layers such as data, AI analytics, business analytics interpretation, human judgment, and feedback learning that combine to facilitate transparency, accountability, and contextual decision-making. The framework is depicted in the most important areas of the organization with the main focus on the strategic management and workforce decision-making and the secondary focus on the finance, operations, and marketing. The framework minimizes the effects of the algorithmic bias, automation bias, and enhances workforce confidence through embedding managerial control and ethical reasoning and contextual evaluation frameworks into the workflows of AI-assisted decision-making. The contribution of the study to the theory is that it develops human-grounded decision analytics and to practice by providing practical advice to executives and analytics leaders. The presented framework contributes to the responsible use of AI, productivity, and economic competitiveness in the United States in the long term.

Summary

Main Finding

The paper proposes a Human–AI Collaborative Decision Analytics Framework that positions business analytics as the interpretive mediator between AI outputs and managerial judgment. The framework — composed of five overlapping layers (data, AI analytics, business-analytics interpretation, human judgment, and feedback learning) — is designed to improve decision quality, transparency, accountability, and workforce trust, especially for strategic and workforce decisions. The framework claims to reduce algorithmic and automation bias, embed managerial control and ethical reasoning into AI-assisted workflows, and thereby enhance organizational performance and long-term competitiveness.

Key Points

Problem addressed: Weak integration of AI outputs into managerial discretion creates underuse, misuse, automation bias, and reduced accountability.
Core proposal: Treat analytics not just as reporting but as an interpretive, sense-making layer that translates AI outputs into actionable, contextualized guidance for managers.
Five-layer framework:
Data (quality, provenance, bias mitigation)
AI analytics (prediction, classification, prescription)
Business-analytics interpretation (explainability, KPI alignment, scenario analysis)
Human judgment (ethical reasoning, contextualization, final decision authority)
Feedback learning (monitoring, model refinement, organizational learning)
Primary focus areas: strategic management and workforce decisions (hiring, evaluations, reskilling), where moral/ethical and contextual judgment matter most.
Secondary exemplars: finance (risk), operations (demand, exceptions), marketing (personalization, privacy).
Benefits claimed: reduced automation/algorithmic bias, higher manager and employee trust, better alignment with strategy and regulation, improved decision quality and accountability.
Limitations acknowledged by authors: conceptual/theoretical study (no empirical testing); scope restricted to higher-stakes decisions (not low-risk fully automated tasks); focuses on managerial process design, not algorithm development.

Data & Methods

Research design: Conceptual/theoretical paper using integrative literature review.
Literatures integrated: managerial decision-making (bounded rationality, cognitive biases, intuition), AI and business analytics (predictive/prescriptive models, dashboards, KPIs), governance/ethics of AI.
Method: Synthesis of interdisciplinary theory to build an integrative framework; illustrative application across organizational functions (strategy, workforce, finance, operations, marketing).
Empirical component: None — framework is proposed and motivated by prior literature and conceptual reasoning; no datasets or field experiments included.

Implications for AI Economics

Human–AI complementarity: The framework highlights productivity gains stemming from complementarities between algorithmic outputs and managerial judgment. Economic models of production should treat analytics-mediated human-AI interaction as a distinct input (not pure automation), with potentially lower elasticity of substitution between labor and AI where interpretive human input remains essential.
Investment and returns: Firms will need to invest not only in AI models but in analytics capabilities, interpretability tools, governance, and managerial training. Returns to AI investment depend on these complementary investments; simple measures of AI capital will understate effective productive capacity if analytics/managerial integration is absent.
Labor market effects: For strategic and workforce decisions, embedding human judgment and ethical oversight may moderate displacement effects—shifting from pure substitution to task reallocation and upskilling. The framework implies greater demand for analytics-literate managers and technical roles that bridge AI outputs and decision-making, potentially amplifying skill-biased demand but also creating new mid-skill jobs.
Allocation and firm performance: Better human-AI integration can improve decision quality and allocative efficiency (e.g., more accurate investments, hiring, risk management), raising firm-level productivity and aggregate competitiveness. However, heterogeneity in adoption (due to capabilities or governance costs) can increase productivity dispersion and market concentration.
Externalities & fairness: By embedding bias mitigation and human oversight in workforce decisions, the framework could reduce discriminatory outcomes and associated social welfare losses. Conversely, poorly implemented frameworks might formalize new biases. Economists should treat fairness/ethical governance as an economic input with welfare implications.
Regulatory and transaction costs: Compliance, explainability requirements, and governance introduce both direct costs and potentially binding constraints that shape adoption paths. Policy choices (e.g., transparency mandates) will affect firms’ adoption incentives and the net social benefits of AI.
Measurement challenges: Empirical testing requires new metrics — decision quality, interpretability, managerial trust, feedback-loop effectiveness — that go beyond model accuracy. Structural and reduced-form empirical work should incorporate mediating variables (analytics interpretation, trust, governance) rather than treating AI adoption as a binary.
Research directions:
- Estimate productivity gains from human-AI collaboration vs. AI-only automation using firm-level or plant-level panel data.
- Task-based models that incorporate an "analytics mediation" factor and endogenous investment in interpretability/governance.
- Field experiments/randomized encouragement designs testing managerial training, interpretability tools, or feedback-loop implementations on decision outcomes, hiring fairness, and performance.
- Labor studies measuring wage and employment effects for roles focused on analytics interpretation and AI governance.
- Welfare analyses of policy interventions (transparency mandates, fairness audits) that affect adoption costs and distributional outcomes.

Overall, this paper suggests that economic models and empirical work on AI should explicitly account for the mediating role of business analytics and the institutional/managerial systems that translate algorithmic outputs into organizational decisions. Ignoring this mediation risks overstating substitution effects and understating complementarities, governance costs, and distributional consequences.

Assessment

Paper Typetheoretical Evidence Strengthlow — Conceptual and integrative literature review with no original empirical data, no causal identification strategy, and no quantitative effect estimates; claims are plausibly theorized but not empirically validated. Methods Rigormedium — Presents a coherent, interdisciplinary synthesis and a clearly articulated five-layer framework that is useful for theory-building and practice; however, it lacks a documented systematic review protocol, empirical tests, robustness checks, or operationalization of constructs. SampleNo original dataset; an interdisciplinary synthesis of existing literature in managerial decision-making, business analytics, AI/ML governance, and organizational studies (conceptual sources and prior empirical studies cited but not re-analyzed). Themeshuman_ai_collab org_design adoption productivity skills_training governance GeneralizabilityNo empirical validation — uncertain external validity to firms or sectors, Heterogeneity across industries and business functions not assessed, Unclear applicability across firm sizes and resource constraints, Implementation complexity and costs vary widely across organizational contexts, Cross-country and regulatory differences not addressed, Worker heterogeneity (skills, roles, cultures) and union/regulatory environments not empirically considered

Claims (10)

Claim	Direction	Confidence	Outcome	Details
The fast spread of artificial intelligence (AI) in U.S. organizations has radically altered the managerial decision-making process. Decision Quality	mixed	medium	managerial decision-making process (structure, speed, inputs)	0.04
AI adoption has augmented complexity, uncertainty in decision-making, and accountability stresses for managers. Decision Quality	negative	medium	decision complexity, decision uncertainty, accountability stresses	0.04
Most organizations have difficulties converting algorithmic results into sustainable managerial decisions due to low levels of trust, lack of explanation, and poor integration between AI systems and human judgment. Decision Quality	negative	medium	conversion of algorithmic outputs into sustainable managerial decisions; trust; explainability; system–human integration	0.04
Current literature has primarily focused on automation-based views of decision support and lacks insight into systematic human–AI coordination aided by analytics. Other	negative	medium	coverage of topics in AI decision-support literature (automation-centric vs. human–AI coordination)	0.04
This paper outlines a Human–AI Collaborative Decision Analytics Framework integrating five overlapping layers: data, AI analytics, business analytics interpretation, human judgment, and feedback learning. Other	null_result	high	structure/components of the proposed Human–AI Collaborative Decision Analytics Framework	0.06
The framework is depicted across organization areas with primary focus on strategic management and workforce decision-making and secondary focus on finance, operations, and marketing. Organizational Efficiency	null_result	high	organizational domains targeted by the framework (strategic management, workforce, finance, operations, marketing)	0.06
Embedding managerial control, ethical reasoning, and contextual evaluation in AI-assisted workflows minimizes effects of algorithmic bias and automation bias and enhances workforce confidence. Ai Safety And Ethics	positive	low	algorithmic bias, automation bias, workforce confidence	0.02
Analytics can serve as the focal interpretive intercession between AI outputs and human decision-makers, facilitating transparency, accountability, and contextual decision-making. Decision Quality	positive	medium	transparency, accountability, contextualization in decision-making mediated by analytics	0.04
The study contributes to theory by developing a human-grounded decision analytics perspective and to practice by providing practical advice to executives and analytics leaders. Other	positive	high	theoretical contribution (human-grounded perspective); practical guidance for executives/analytics leaders	0.06
The presented framework contributes to the responsible use of AI, productivity, and long-term economic competitiveness in the United States. Fiscal And Macroeconomic	positive	speculative	responsible AI adoption, organizational productivity, long-term economic competitiveness (U.S.-level)	0.01