Table 3 Research agenda: streams,... | Emerald Publishing

Table 3

Research agenda: streams, propositions, and methodological requirements

Stream	Research question	Propositions	Methodological approach
1. Paradox dynamics and moderators	Under what conditions does the autonomy–ambiguity paradox resolve in favor of augmentation versus overwhelm?	P1a: Organizations automating legacy tasks before expanding AI-generated scenario volume demonstrate lower planner cognitive load than those expanding scenarios first	Longitudinal designs tracking organizations through the GenAI implementation phases (with cognitive load and role clarity as mediating variables)
		P1b: The productivity J-curve duration correlates inversely with clarity of human–AI decision rights
		P1c: Organizations implementing GenAI scenario generation without concurrent legacy task automation exhibit higher planner turnover intention than those implementing both simultaneously
	Under what conditions does GenAI support adaptive responses to genuinely unknowable uncertainty versus predictive responses to quantifiable variation?	P1d: GenAI adoption enables simultaneous deployment of predictive, proactive, reactive, and adaptive uncertainty regulation strategies; augmentation outcomes depend on matching the AI system design to the type of uncertainty faced by planners	Comparative case studies across uncertainty contexts Experimental designs manipulating uncertainty type and AI system configuration
2. Intervention effectiveness	Which cognitive load mitigation strategies produce measurable improvements in planner performance and well-being?	P2a: Scenario-filtering heuristics reduce decision time without degrading decision quality when filtering criteria are co-developed with planners	Experimental and quasi-experimental designs isolating intervention effects and maturity model (Figure 3) as a staging framework for matched comparisons
		P2b: AI-generated narrative explanations reduce cognitive load more effectively than numerical confidence intervals alone
		P2c: Targeted interventions addressing specific transition paradoxes outperform generic change-management approaches
3. Longitudinal outcomes	How does GenAI adoption affect planner decision quality, well-being, and retention over 12–36-month horizons?	P3a: Planners in augmentation configurations report higher job satisfaction and lower turnover intention than those in overwhelm configurations, controlling for workload	Outcome metrics appropriate to SCP contexts (forecast accuracy, exception resolution time, inventory performance), excluding self-reported effectiveness
3. Longitudinal outcomes		P3b: Forecast accuracy and exception resolution time exhibit J-curve patterns with inflection points that are contingent on organizational support mechanisms

Stream	Research question	Propositions	Methodological approach
1. Paradox dynamics and moderators	Under what conditions does the autonomy–ambiguity paradox resolve in favor of augmentation versus overwhelm?	P1a: Organizations automating legacy tasks before expanding AI-generated scenario volume demonstrate lower planner cognitive load than those expanding scenarios first	Longitudinal designs tracking organizations through the GenAI implementation phases (with cognitive load and role clarity as mediating variables)
		P1b: The productivity J-curve duration correlates inversely with clarity of human–AI decision rights
		P1c: Organizations implementing GenAI scenario generation without concurrent legacy task automation exhibit higher planner turnover intention than those implementing both simultaneously
	Under what conditions does GenAI support adaptive responses to genuinely unknowable uncertainty versus predictive responses to quantifiable variation?	P1d: GenAI adoption enables simultaneous deployment of predictive, proactive, reactive, and adaptive uncertainty regulation strategies; augmentation outcomes depend on matching the AI system design to the type of uncertainty faced by planners	Comparative case studies across uncertainty contextsExperimental designs manipulating uncertainty type and AI system configuration
2. Intervention effectiveness	Which cognitive load mitigation strategies produce measurable improvements in planner performance and well-being?	P2a: Scenario-filtering heuristics reduce decision time without degrading decision quality when filtering criteria are co-developed with planners	Experimental and quasi-experimental designs isolating intervention effects and maturity model (Figure 3) as a staging framework for matched comparisons
		P2b: AI-generated narrative explanations reduce cognitive load more effectively than numerical confidence intervals alone
		P2c: Targeted interventions addressing specific transition paradoxes outperform generic change-management approaches
3. Longitudinal outcomes	How does GenAI adoption affect planner decision quality, well-being, and retention over 12–36-month horizons?	P3a: Planners in augmentation configurations report higher job satisfaction and lower turnover intention than those in overwhelm configurations, controlling for workload	Outcome metrics appropriate to SCP contexts (forecast accuracy, exception resolution time, inventory performance), excluding self-reported effectiveness
3. Longitudinal outcomes		P3b: Forecast accuracy and exception resolution time exhibit J-curve patterns with inflection points that are contingent on organizational support mechanisms

Source(s): Authors’ own elaboration