Journal of Service Research

This article presents the 2026 Service Research Priorities (SRPs), developed via a hybrid AI–human agenda-setting approach. We construct large-scale concept networks from more than 10,500 service-related articles in core service journals and Financial Times (FT) 50 outlets, and apply machine-learning link prediction to identify high-probability, but underexplored, concept pairs. Fifteen forward-looking themes emerge that fall into four topical clusters: (a) Human–AI Agency, Interactions, and Service Reconfiguration; (b) AI-Enabled Omnichannel Engagement and Ecosystem Orchestration; (c) Responsible, Inclusive, and Resilient AI Service Systems; and (d) Psychological Dynamics of Service Experiences. Comparing the themes’ relevance in the service outlets vis-à-vis the FT50 journals helps identify where service research currently leads, where it imports from adjacent disciplines, and where cross-fertilization is most promising. Furthermore, we reflect our themes against the 2021 SRPs, distinguishing “emerging,” “evolved,” and “continuing” priorities. Importantly, we also survey the Journal of Service Research leadership to further calibrate and enrich the AI-derived SRPs. Next, we introduce a new descriptive statistic, the “Interdisciplinary Influence Index,” to map which disciplines “drive” a particular service theme within the literature broadly defined. Finally, we provide the “Interactive Service Scholarship Incubator” app that enables scholars to explore the underlying concept network, predicted links, and themes, in light of their own interests and skill sets. Together, these SRPs offer a scalable roadmap for high-impact future service scholarship and practice.

Graphical Abstract

This is a visual representation of the abstract.

The rapid growth of artificial intelligence (AI) reflects its widespread adoption across various industries and occupations. Yet, much of the discourse still conceptualizes AI narrowly as a property of computational machinery. Within marketing and service scholarship, such mechanistic framings obscure the relational and systemic dimensions of intelligence that become salient in AI-enabled service interactions. Drawing on service-dominant (S-D) logic, this paper reconceptualizes intelligence as an emergent property of adaptive, self-organizing service ecosystems, foregrounding the systemic processes and relational configurations through which actors and their capacities take form. As intelligent technologies become increasingly entangled in service exchange, they render more visible and consequential the ongoing interactions through which human and non-human actors are co-constituted, underscoring the need for a clearer relational-ontological grounding of S-D logic. We conceptualize adaptation and intelligence as a generative duality: adaptation captures the adjustments enacted within a service ecosystem, whereas intelligence denotes the systemic capacity enabling those adjustments through feedback, learning, and institutional co-evolution. This perspective strengthens S-D logic’s relational foundations by showing that AI exemplifies—rather than disrupts—the emergent, relational processes through which service ecosystems coordinate, evolve, and create value.

Graphical Abstract

Artificial intelligence (AI)-powered chatbots offer a cost-effective solution for customer service, but often fall short in delivering personalized or complex interactions. In response, many merchants invest in AI training and explore human–AI collaboration to leverage the strengths of both automation and human touch. However, this introduces a strategic tradeoff between service quality and operational efficiency. Using a game-theoretic framework, this study examines how merchants can optimally choose service strategies with AI involvement. Our analysis reveals a critical collaboration trap. Contrary to the prevailing belief that increased collaboration consistently enhances service, consumers’ sensitivity to the identity of the service agent erodes profitability, and greater collaboration will worsen this effect by accelerating task delegation to AI, thereby amplifying negative consumer perceptions. Furthermore, the study shows that a moderate-cost collaboration trap emerges: human–AI collaboration underperforms compared to human-only service when labor costs are at intermediate levels. Collaboration yields benefits only when labor costs are either very low or prohibitively high. In competitive markets, merchants can gain a strategic advantage not by enhancing their own service quality, but by capitalizing on rivals’ inefficient AI deployment. These findings challenge the assumption that more human–AI collaboration is always better and provide actionable insights for managing hybrid service strategies and optimizing AI investments.

Graphical Abstract

The latest advancements in artificial intelligence (AI) not only highlight exceptional intelligence quotient but also emphasize remarkable emotional intelligence or emotional quotient. The current research investigates how interactions with empathetic AI affect individuals’ propensity to engage in prosocial behavior. Across six experiments, including a real-time human-AI interaction study and an incentivized behavioral study, we demonstrate that interactions with high- (vs. low-) empathy AI significantly increase individuals’ inclinations to donate, the amount donated, and their intentions to volunteer. Drawing on observational learning theory, we show that this effect occurs because high-empathy AI enhances perceptions of AI humanness. This heightened humanness facilitates empathetic influence, whereby the AI’s modeled empathic behaviors elevate users’ own empathic states, ultimately promoting greater prosocial action. Furthermore, this positive effect of high-empathy AI is strengthened by AI’s human-identification marks (i.e., human name and icon). The research findings offer theory-based, actionable recommendations for effectively leveraging AI in service encounters to promote prosocial behaviors.

Graphical Abstract

Shopper artificial intelligence (AI) presents a striking paradox: while massive investments drive rapid expansion and increasingly sophisticated AI solutions, two-thirds of consumers express dissatisfaction with AI shopping assistants, citing frustrations with pushy upselling, poor understanding, and inaccurate recommendations. This disconnect motivates our development of the Shopper AI taxonomy. To develop our taxonomy, we synthesized insights from multiple disciplines through a design science research process with empirical validation. Grounded in customer experience management (CEM) theory, our taxonomy identifies 14 dimensions within two meta-characteristics: AI capabilities (knowledge, intelligence, autonomy, breadth of use, quality of work, data privacy) and AI parasocial skills (personalization, anthropomorphism, communications mode, emotion recognition, emotion expression, empathy, influence, engagement). The taxonomy advances service research theory in three ways. First, we extend CEM theory by revealing how AI creates value through interrelated but discrete capabilities and parasocial dimensions. Second, we identify how AI capabilities enable autonomous value creation without active customer participation, representing a new form of value pre-creation. Third, we reveal complex dimensional interactions, where improvements in one dimension can enhance or diminish others. This multidimensional taxonomy provides managers with actionable guidance for navigating dimensional trade-offs, designing efficient, balanced AI systems, identifying context-specific investment priorities, and avoiding common pitfalls.

Graphical Abstract

Services are often segmented into discrete events, allowing customers to self-schedule their own itinerary. We term this behavior as self-scheduling. We theorize that customers self-schedule peak events—those they predict will be their most salient—at predictable points, primarily for the bookends (i.e., at the beginning or the end). This behavior can create demand fluctuations and pose challenges for demand management. To examine this phenomenon, we conducted two exploratory studies. First, a survey using a tour context revealed a preference for self-scheduling the peak event at the beginning. A more balanced distribution between bookends emerged when information promoting the peak event was provided. Second, wait-time data from three major theme parks in the United States was collected during the summer of 2024 and validated that customers predominantly self-schedule peak events for the beginning. Next, we hypothesized how information provision may influence customers’ self-scheduling behavior of a peak event. A scenario-based experiment and a conjoint study in a theme park context found that practices enhancing perceived control (i.e., wayfinding and wait line management information) effectively shifted some of the demand from the beginning to later in a visit. We discuss insights to support demand management in self-scheduling service contexts.

Graphical abstract

Developing responsive dynamic marketing strategies can be challenging in the absence of complete customer information, such as share of wallet, limiting the ability of the firm to target promotions and other marketing efforts with a view to optimizing customer lifetime value (CLV). Furthermore, much of the existing research on CLV treats customers as receivers, rather than co-creators, of services. We address these two key challenges by developing a reinforcement learning (RL)-based promotion optimization model to determine which promotion strategies are most suitable for targeting different customer groups. Specifically, using feedback derived from customers’ real-time transactional responses to promotional campaigns, we present an RL algorithm that (a) continually refines the estimated effectiveness of promotions, aligning them to customers’ preferences to maximize their CLV, and that (b) supports value co-creation by involving customers as active participants to enhance their service experience. We demonstrate the effectiveness of the model through simulation scenarios within the context of a ferry travel agency, providing evidence of its real-world potential.

Graphical Abstract

Loyalty programs remain a cornerstone in the relationships between customers and service providers. While most previous research focuses on comparing the effectiveness of loyalty programs by comparing the consumption behaviors of members and non-members, few papers have investigated whether new enrollees are profitable and how such enrollments affect non-members. Results across data from two different firms demonstrate that enrolling new members in a loyalty program can increase profits. However, additional studies provide evidence that increased new member enrollment is associated with decreased non-member spending, with new members seeking more discounts than other customers. We show that non-members may feel that member benefits are undeserved, thus decreasing their own spending. Adopting a service management approach, we argue that non-members should not be neglected as they also form an important asset class. Also, we find that experienced managers and greater clarity about loyalty program rules may mitigate some of the negative effects of higher enrollments. The results suggest that managers of loyalty programs may need to more actively assess the return on investment of established reward members, new enrollees, and even non-members.

Graphical Abstract

Employees with intellectual disabilities remain markedly underrepresented in frontline service roles, despite the social and economic benefits of inclusion. A key barrier is firms’ uncertainty about consumer reactions. To address this gap, we develop and validate the CAFID scale (Consumer Attitudes toward Frontline Employees with Intellectual Disabilities), designed specifically for the service sector—unlike existing, general disability attitude measures. Across multiple studies, we demonstrate the scale’s reliability, dimensionality, and validity. Results show that more favorable consumer attitudes not only support inclusive hiring but also buffer negative responses to service failures, improving repatronage and word-of-mouth intentions. With this practical tool, firms can assess customers’ readiness and support for inclusive hiring practices; policy makers can develop data-driven incentives to promote workplace equity; and researchers can explore dynamic consumer–employee interactions involving people with intellectual disabilities. Ultimately, the CAFID scale contributes to building more equitable, socially responsible service environments.

Graphical Abstract

This multicity, qualitative study provides a nuanced understanding of how social service providers deploy human-centered service design approaches and the respective intended value creation through a diverse set of social services. This work makes three central contributions. First, it offers an initial typology of social service design approaches based on their intended value: (1) triaging to stabilize, (2) incubating to develop, and (3) empowering to liberate. Second, we demonstrate the imperative to focus on the experiences of both vulnerability and resilience for disadvantaged consumers, allowing sensitivity to consumers’ resource deficits and capabilities. Third, we introduce the concept of humanitarian value, a dynamic, adaptive, and encompassing value that addresses immediate and long-term needs, recognizes human strengths and resource deficits, and serves to enhance the well-being of individuals and communities. The findings provide a guide for social service organizations, which often operate in unstable, precarious environments, to address the idiosyncratic needs of their clients and holistically assess the persistent wicked problems and social needs in their respective communities.

Graphical Abstract