Introduction

Artificial intelligence (AI) is playing an increasingly important role in firm production and services (Michalec et al., 2021; Nabavi and Browne, 2023; Zhang et al., 2024). For example, Ant Fortune uses AI to build investment portfolios that are highly aligned with customer needs (Zain et al., 2020), Sensely provides the AI nurse “Molly” to help patients with their health management needs (Kalis et al., 2018). Restaurants and hotels employ service robots to provide consumers with information and services (Prentice and Nguyen, 2020). AI continues to challenge the work of human employees by endowing machines with cognitive functions that are comparable to or even surpass the human brain (Rai et al., 2019). Therefore, some studies have shown AI would pose a huge threat to employees’ jobs (Lee et al., 2018), resulting in the loss of some unique abilities and knowledge (Fügener et al., 2021), and a heightened sense of job insecurity (Yam et al., 2023). These effects may be associated with more maladaptive workplace behaviors, including a reduced perception of career achievement. By contrast, some scholars indicate that the use of AI may improve work efficiency (Tong et al., 2021), reduce error rates, and help employees provide customers with higher-quality and personalized services (Marinova et al., 2017). These effects may significantly enhance employees’ perception of career achievement. Nevertheless, what causes contradictions owing to these viewpoints and empirical results has not been examined adequately in the literature in a thorough and systematic manner.

The perception of career achievement reflects employees’ evaluation and understanding of their work experience and attainments (Gattiker and Larwood, 1990). The formation and development of employees’ intrinsic fulfillment at work are very important, especially for low-level employees working in the front line of service industry (Prentice et al., 2019). It is often difficult for these employees to obtain tangible rewards (e.g., promotions and salary increments; Acker, 2006) because their jobs are highly substitutable (Prentice et al., 2013) and require very little sophisticated professional training and skills (Prentice et al., 2019). In the absence of extrinsic rewards, intrinsic fulfillment is the key factor driving their subsequent work behavior and performance (Babakus et al., 2003; Zapata-Phelan et al., 2009). However, these employees play an imperative role in firm services because their work behaviors and performance directly affect customer evaluation and transaction results (Cheshin et al., 2018; Xiong and King, 2015). Therefore, in AI service environments, choosing an appropriate strategy (i.e., intelligence substitution and collaboration) to enhance employees’ perception of career achievement is an issue that companies cannot ignore. If AI replaces employees completely, that is, companies adopt the intelligence substitution service strategy (Marinova et al., 2017), then, in the absence of a platform for employees to showcase their achievements (Prentice et al., 2019), such replacement is likely to reduce their perception of career achievement. Some companies use the intelligence collaboration service strategy, that is, they focus on services dominated by employees and supplemented by AI, and emphasize the critical position of employees in the service process (Wilson and Daugherty, 2018). This strategy may not destroy employees’ perception of career achievement because their work value is still recognized by the companies (Marinova et al., 2017). However, there is still a lack of attention to the impact of the intelligent service strategy on the employees. Most studies have been conducted from the perspective of consumers rather than employees (e.g., Garvey et al., 2023; Mende et al., 2019; Złotowski et al., 2017). Although some studies have discussed the impact of AI on employees, they pay more attention to their psychological cognition and work performance, ignoring the possible changes in employees’ intrinsic motivation (such as the perception of career achievement). Whether or not there is a difference in the impacts of the two different enterprise service strategies, namely intelligence substitution and collaboration, on the perception of career achievement remains unclear.

To address this knowledge gap, we explore the following research questions:

RQ1: What mechanism explains the differential impacts of two enterprise service strategies on employees’ perception of career achievement? And if so:

RQ2: Would the impacts of the intelligence substitution versus the intelligence collaboration service strategy differ depending on the organizational innovation climate?

To answer these questions, first, drawing on prior research on self-perception theory (Bem, 1972; Teng, 2018), we introduced human–human and human–machine interactivity as the mediation mechanism of intelligent service strategy on employees’ perception of career achievement. From the perspective of the work pattern, the application of AI has gradually changed the form of interaction in the service environment from human–human (employee–customer) to human–machine (employee–AI–customer) (Paschen et al., 2020). However, this change in employees’ intrinsic motivation, especially in the perception of career achievement, remains unclear, and the literature ignores the exploration and explanation of this impact. From the perspective of work efficiency and performance (e.g., Tong et al., 2021), or AI threat theory (e.g., Kopp et al., 2022; Lee et al., 2018), explaining the impact of the application of AI in service work on the employees’ psychology and behavior is very insightful, but cannot fundamentally understand the changes in the intrinsic motivation of employees. According to the self-perception theory, people learn about their internal emotion and state (e.g., perception of career achievement) by observing their external behavior and the situation in which such behavior occurs (Bem, 1972). An individual’s internal reflections on their associated environment and behaviors (e.g., the work pattern based on human–machine interaction) trigger their psychology and behavior (Jiang et al., 2019; Makhija and Stewart, 2002). Therefore, in the AI service environment, an enterprise’s intelligent service strategy can affect employees’ perception of career achievement by triggering changes in the way employees interact.

Second, we drew on the research on the influence of the innovative climate on employees’ cognition and behavior (Newman et al., 2020; Shalley et al., 2004) to argue that enterprises can affect employees’ perception of career achievement by creating an organizational innovation climate. In an enterprise environment with a strong organizational atmosphere, when compared with employees’ communication and interaction skills, enterprises respect employees’ ability to use new technologies to achieve greater service innovation (Hecht and Allen, 2005; Quratulain et al., 2021). The enhancement of human–machine interactivity means that employees need to use more new technologies, such as AI. Therefore, the creation of an organizational innovation climate can affect employees’ perceptions of human–human and human–machine interactivity, which, in turn, will change their perception of career achievement. From this perspective, the introduction of an organizational innovation climate can weaken the negative impact of an enterprise’s implementation of the intelligent service strategy on employees.

We conducted three experiments involving 736 participants and varied types of service firms (experience-based and trust-based services). We have found that the intelligence substitution and collaboration service strategies would affect employees’ perception of career achievement by changing the interaction behavior and pattern (human–human vs. human–machine interactivity) in the service environment. The organizational innovation climate can moderate employees’ perception of career achievement. Our research makes the following contributions. First, to the best of our knowledge, this study is one of the pioneers in exploring the new and important phenomenon of the impact of AI adoption on employees’ perception of career achievement. As AI technology advances, enterprises are gradually replacing front-line employees with AI to provide services to consumers (Huang and Rust, 2018). This represents an unprecedented opportunity for enterprises to create value (Tong et al., 2021), but it has also led to significant changes in service enterprises (Liang et al., 2022; Mende et al., 2019). Existing studies have predominantly focused on consumer attitudes and preferences towards AI (e.g., Garvey et al., 2023; Mende et al., 2019), overlooking the impact of AI adoption on employees, especially on the intrinsic motivation of front-line employees. Thus, this research takes an initial step in extending previous research into AI applications in production and marketing. In this context, drawing on self-perception theory, this research proposes that enterprises’ intelligent service strategy (collaboration vs. substitution) can change the employees’ perceptions of interactivity in the service environment, influencing their understanding of their work. Our second contribution comes from this. AI-related research has mainly focused on the positive effects, such as improving employee productivity (e.g., Tong et al., 2021), or the negative effects, such as reducing employee performance, and triggering employees’ job turnover intention (e.g., Li et al., 2019; Tang et al., 2022b). An unclear question is what causes this difference. By explaining the differing impacts of enterprise intelligent service strategy on the perception of career achievement, this study enriches the existing research on AI adoption. Third, from the perspectives of both work pattern and organizational environment, we construct a theoretical model of the formation of employees’ perception of career achievement. While studies on employees’ perception of career achievement have focused on factors such as organizational climate (Briones et al., 2010; Prentice et al., 2019) and job performance (Miner, 2015; Prentice et al., 2019), these discussions have been fragmented and single-perspective in the traditional work environment. Our study incorporates the concept of organizational innovation climate to examine the key roles played by this factor in the implementation of intelligent service strategy. This work has practical implications, as the integration of AI and services is a crucial factor influencing human resource practices (Larivière et al., 2017).

Theoretical background and hypotheses development

Perception of career achievement

Perception of career achievement refers to employees’ subjective judgments of both themselves and their career achievements, often based on their understanding of events or situations (Gattiker and Larwood, 1990). As intrinsic motivation, the achievement perception is influenced by internal and external variables. The former includes gender, age, self-efficacy (Briones et al., 2010), employees’ job insecurity (Yam et al., 2023), and work overload (Lings et al., 2014). The latter includes the organizational atmosphere (Briones et al., 2010; Prentice et al., 2019) and work methods (Prentice et al., 2013).

Although studies have discussed the formation of employees’ perception of career achievement from various perspectives, they have been fragmented and informed by a single perspective. Most studies have focused on traditional service scenarios. In AI-embedded service environments, it is unclear whether the intelligent service strategy will change an employee’s perception of career achievement. Previous studies have highlighted that the adoption of AI increases uncertainty and job insecurity among employees (Kong et al., 2021), resulting in job burnout (Yam et al., 2023), higher turnover intention (Li et al., 2019), and impacts on their career development (Kong et al., 2021). These studies focus on the impact of AI on employees from the perspective of employees’ perceptions and cognition of AI. To a certain extent, these studies reveal the impact of AI technology on employees’ psychology and behavior but overlook the changes in working methods in the AI-embedded service environment (i.e., from human–human interaction to human–machine interaction), which may influence employees’ intrinsic motivation.

Self-perception theory

The core of self-perception theory holds that an individual’s behavior and the situation in which such behavior occurs can reflect the individual’s inner activity and state (Bem, 1972; Ju et al., 2019). An underlying assumption of self-perception theory is that individuals cannot fully perceive all the psychological states related to their behaviors (Jiang et al., 2019). Therefore, individuals usually infer their inner states from their behaviors (Yan et al., 2013). In the process, they consider the factors that have triggered this behavior by reflecting on it (Doll and Rosopa, 2015). This process resembles social perception, which emphasizes that people understand others’ thoughts and feelings by observing their behaviors (Garnefeld et al., 2011). Self-perception theory assumes the same mechanism that people become aware of their own inner states through the reflection of their own behaviors (Bem, 1972). For instance, Yan and Davison (2013) found that individuals can “know” their intrinsic motivations by observing their overt knowledge-seeking behavior and its context. This line of reasoning seems to be extended to research on employees’ intrinsic motivation (i.e., perception of career achievement) in an intelligent service environment. The internal achievements of employees largely depend on the environment in which the intelligent service strategy is implemented (Larivière et al., 2017). According to the self-perception theory, employees form overt perceptions toward their work patterns and environment. Once work pattern or environment change, they are likely to reflect on these changes, and “know” their current intrinsic motivation by observing and interpreting their behavior and related situations. The enterprise’s intelligent service strategy changes the mode of interaction (i.e., from human–human interaction to human–machine interaction), initiating employees’ self-perception/attribution process and changing their perception of career achievement. The self-perception theory is used to examine employees’ perception of career achievement in the intelligent service environment because it highlights that an individual’s psychological response is a proximate or fundamental outcome of their own behaviors (Bem, 1972). In other words, an individual always enters into psychological reflection after a specific behavior (Bem, 1972; Jiang et al., 2019). The research framework is shown in Fig. 1.

Fig. 1: Theoretical framework.
figure 1

The model shows the relationship between the main variables. The arrows indicate the mechanism of influence between the variables.

Intelligent service strategy and employees’ perception of career achievement

AI changes the sources of competitive advantage in enterprises (Daugherty and Wilson, 2018). Krakowski and colleagues (2023) pointed out that when AI assumes humans’ cognitive capabilities, it can either substitute or complement human conduct. We proposed that the intelligent service strategy implemented by enterprises has one of two directions: substitution or collaboration with human capital with the goal of having AI assume the service capabilities of employees. The intelligence substitution service strategy means that enterprises tend to use AI to dominate services, and employees’ service capabilities are substituted by AI (Marinova et al., 2017). In such an environment, employees no longer have direct, face-to-face communications with customers, and intelligent systems automatically provide customers with appropriate services. The employees are no longer involved in the service process generated by technology (Froehle and Roth, 2004), serving only as back-office personnel who operate machines or use technology. By contrast, the intelligence collaboration service strategy refers to the use of AI to help, supplement, and enhance the service of employees, and employees’ service capabilities are complemented by AI (Marinova et al., 2017). This is the embodiment of technology supporting and assisting human thinking, analysis, and behavior (Larivière et al., 2017). In this environment, employees remain the main force in interacting with customers and providing services to them. AI technology analyses and predicts customer needs accurately, thus helping employees provide higher-quality and more personalized services (Bowen and Morosan, 2018). This strategy emphasizes the maximization of corporate benefits through the joint creation of value by employees and AI (Marinova et al., 2017).

We now turn to the relationship between the intelligent service strategy and employees’ perception of career achievement. Perception of career achievement, an intrinsic motivation, manifests an employee’s psychological state (Gattiker and Larwood, 1990). From the perspective of the self-perception theory, employees’ cognition of their work environment influences their intrinsic motivation (Bem, 1972; Zeng et al., 2023). Therefore, the intelligent service strategy inevitably affects the employees’ perception of career achievement. In an environment with the intelligence substitution service strategy, the original work of employees is replaced by AI, rendering the work of employees in traditional service positions outdated (Froehle and Roth, 2004; Larivière et al., 2017). This may lead to the professional knowledge and skills of employees being underutilized (Froehle and Roth, 2004). In an environment with the intelligence collaboration service strategy, companies realize employees’ work value through human–machine collaboration (Froehle and Roth, 2004; Marinova et al., 2017). Therefore, in an environment with the intelligence collaboration service strategy, employees’ perception of career achievement is enhanced. The following hypothesis is thus proposed:

Hypothesis 1

Intelligence substitution service strategy (vs. intelligence collaboration service strategy) is associated with lower employees’ perception of career achievement.

Employees’ perception of interactivity in the intelligent service environment

The implementation of intelligent service strategy leads to changes in the work patterns of employees. We analysed the impact of this change on employees’ perception of career achievement by introducing the concept of interactivity. Some studies consider interactivity a psychological factor, derived from people’s psychological perception of changes in technology and communication environments (Newhagen et al., 1995). We define it as employees’ understanding and perception of the interaction behavior and form in the organizational service environment.

Interactivity is classified according to the environment and subject (Kiousis, 2002). Considering AI applications in the enterprise environment as the interactive environment and the employees, customers, and AI as the interaction subjects, this research divides interactivity into human–human and human–machine interactivity. The former refers to the employee’s perception that the service environment is dominated by human-to-human and face-to-face sensory communication and interaction, forming employee–customer interactions (Murphy and Sashi, 2018). It focuses on face-to-face communication and coordination among people, with humans as the main objects (Ziegert et al., 2022). The latter refers to the employees’ perception that the service environment is dominated by the interaction between humans and AI machines. AI machines are part of the interaction to construct an employee–AI–customer interaction form (Ischen et al., 2020). Human–machine interactivity specifies the interactions between humans and machines, but it is not the machine that entirely mediates the communication between humans.

In an organizational environment with an intelligence substitution service strategy, AI replaces human labor in traditional service forms (Mcleay et al., 2021). With employees no longer serving customers, direct human–human interactions are reduced or even eliminated. In the retail service of Amazon Go, for instance, in the process of product selection and checkout, the salesperson no longer provides services for customers, and the payment is made directly through the Amazon account (Larivière et al., 2017). Amazon automatically recommends products to customers through the collection and analysis of their purchase records (Marinova et al., 2017). In the financial and legal services industries, there are corresponding applications and AI systems that automatically communicate with customers and provide them with investment information and legal assistance (Larivière et al., 2017; Lee et al., 2023). In service companies that deploy the intelligence substitution service strategy, the interaction form of “employee–customer” is replaced by the interaction form of “employee–AI–customer.” In an organizational environment that deploys the intelligence collaboration service strategy, enterprises assist employees in completing service work by applying AI. Human–human interactions are not replaced by human–machine interactions; instead, they coexist in the service environment (Larivière et al., 2017; Mcleay et al., 2021). For example, Augmedix’s smart glass technology can help doctors collect, organize, update, and analyse patient data in real time, thus effectively improving the efficiency of doctor–patient communication and diagnosis (De Keyser et al., 2019). AI emotion recognition systems (such as affectiva.com) can monitor the emotional state of and psychological changes in customers when employees communicate with them, helping employees adjust communication strategies in a timely manner (Huang and Rust, 2018). In service companies that deploy the intelligence collaboration service strategy, the interaction form of “employee–customer” continues to exist. The human–human interactivity is stronger with the intelligence collaboration service strategy compared to the intelligence substitution service strategy. By contrast, the human–machine interactivity will be lower. Thus, the following hypotheses are proposed:

Hypothesis 2a

Intelligence substitution service strategy (vs. intelligence collaboration service strategy) is associated with lower human–human interactivity.

Hypothesis 2b

Intelligence substitution service strategy (vs. intelligence collaboration service strategy) is associated with higher human–machine interactivity.

Employees’ perception of interactivity and perception of career achievement

Employees’ perception of career achievement depends on the adaptability between their personality characteristics and the environment (Smart et al., 1986). For employees engaged in service work, an environment of high human–human interactivity appears to be a better match for their professional skills because both in traditional and intelligent service environments, human–human interactions are important for employees to display their professional knowledge and skills. They can fully utilize their knowledge, experience, and ability (Prentice et al., 2019) and co-create value with customers (Gazzoli et al., 2013). Based on the development and application of such skills and the ensuing positive results, employees’ perception of career achievement will be enhanced further (Locke, 1970). When employees perform well in the process of human–human interaction with customers, they are likely to garner positive feedback and praise from customers. Such feedback and recognition from others bring employees a sense of inner fulfillment (Prentice et al., 2019).

By contrast, employees’ perception of human–machine interaction in the service environment may reduce their perception of career achievement. Interactions between employees and AI machines are dramatically different when compared to their interactions with other technologies (Daft and Lengel, 1986; Suen et al., 2019) because AI can learn by analyzing employees’ prior work content and decisions (Dunjko and Briegel, 2018). AI has a high degree of autonomous capabilities, which enables it to make judgments and decisions without instructions from employees (Brynjolfsson and Mitchell, 2017). These capabilities weaken the ability of employees to control their work during human–machine interaction (Hu and Judge, 2017). These advanced skills of AI machines reduce the discretion and flexibility of employees in the process of human–machine interaction, limiting their opportunities to develop novel insights and methods to deal with their jobs (Tang et al., 2022b). The enhancement of human–machine interaction makes service employees less relevant to the service generation process (Danaher, 2017), causing them to doubt their value and contribution and weakening their perception of career achievement. AI enables machines to think, process, and perform tasks autonomously (Brynjolfsson and Mcafee, 2017), blurring the otherwise clear responsibilities of employees (Sun and Medaglia, 2019). These machines may rapidly synthesize information and revise work content or methods through the continuous capture and analysis of data (Raisch and Krakowski, 2021; Tang et al., 2022b). Therefore, human–machine interactions may cause employees to experience confusion, anxiety, and self-doubt, reducing their perception of career achievement. Thus, the following hypotheses are proposed:

Hypothesis 3a

Human–human interactivity positively impacts employees’ perception of career achievement.

Hypothesis 3b

Human–machine interactivity negatively impacts employees’ perception of career achievement.

The moderating effect of organizational innovation climate

Employees’ perception of career achievement stems from strong human–human interactivity, which contrasts with enterprises implementing intelligent service strategies expect. Therefore, enterprises must find ways to alleviate the reduction of employees’ perception of career achievement owing to the enhancement of human–machine interactivity. According to the self-perception theory, employees reflect on their behavior and consider the environment in which the behavior manifests (Bem, 1972). Therefore, from the perspective of the organizational environment, an organizational climate that encourages innovation can change employees’ views on new technologies such as AI, and help them adapt to new work patterns and environments of human–machine interactivity.

Organizational innovation climate is employees’ perception of the degree to which the organization supports and recognizes their use of new technologies, tools, and/or methods to provide innovative services to customers (Al-Hawari et al., 2019). In an enterprise with a strong organizational innovation climate, employees can freely control their behavior and are encouraged to participate in generating innovative ideas to provide services to customers (Hon, 2011). In such a work environment, employees become more confident and be equipped to actively address the challenges brought about by the application of AI (Wang and Ma, 2013). Thus, they are better able to use AI to do their jobs. That is, even if the interaction with AI machines makes employees lose control over their work and triggers new changes (Hu and Judge, 2017; Tang et al., 2022b), a strong organizational innovation climate helps employees develop a work manner with their company’s climate and improve the ability to use AI (Agnihotri et al., 2022). This is because companies with a high organizational innovation climate create flexible and open environments for employees to use AI technology (Evans et al., 2007; Marshall et al., 2019), and offer them time to adapt to novel methods (Scott and Bruce, 1994). From the perspective of the self-perception theory, environments with high human–machine interactivity and innovation climate seem to fit. This can enhance employees’ perception of career achievement. An environment with high human–human interactivity embodies the importance of communication among people, and the employees’ interpersonal skills can often be fully reflected in this environment (Prentice et al., 2019; Ziegert et al., 2022). In an environment with a strong organizational innovation climate, companies prioritize new technologies over the employees’ interpersonal skills (Al-Hawari et al., 2019), and the resources, technologies, and training provided are more related to the application of AI (Quratulain et al., 2021). Based on the reflection of such work methods with human–human interactivity and the environment that encourages innovation, employees’ perception of career achievement will be reduced further.

By contrast, organizations with a low innovation climate are less willing to create a free and open organizational climate for employees, and place greater emphasis on absolute obedience to work patterns and processes (Scott and Bruce, 1994). This may force employees to work in line with the company’s established work patterns, and the company will clarify employees’ work content and responsibilities by formulating rules and orders (Noe et al., 2010). Employees engaged in service work have strong interpersonal communication skills (Prentice et al., 2019). Therefore, in an organizational environment with high human–human interactivity, these employees can still leverage their expertise in line with the company’s established work patterns, create value, and thus enhance the employees’ perception of career achievement. However, in an organizational environment with high human–machine interactivity, employees’ service skills are underutilized, forcing them to adapt to new changes brought about by the use of AI machines (Gabriel and Pessl, 2016). Based on the reflection on this human–machine interactivity work method and the organizational environment, employees’ perception of career achievement is reduced. Thus, the following hypotheses are proposed:

Hypothesis 4a

Organizational innovation climate moderates the relationship between human–human interactivity and employees’ perception of career achievement.

Hypothesis 4b

Organizational innovation climate moderates the relationship between human–machine interactivity and employees’ perception of career achievement.

Overview of studies

In services involving experience attributes, such as staying in hotels, dining in restaurants, and shopping in supermarkets, consumers have a relatively clearer risk perception of uncertainty compared to services with credence attributes such as purchasing insurance and financial products (Keh and Sun, 2018; Mcleay et al., 2021). Therefore, Study 1 and Study 2 used companies with experience and credence attributes as experimental materials, respectively. Study 1 examined the main and mediation effects of intelligent service strategy on employees’ perception of career achievement, verifying Hypotheses 1–3. Study 2 aimed to test the moderating effect of the organizational innovation climate on the relationship between human–human (human–machine) interactivity and perception of career achievement. Study 3 retested Hypotheses 1–4 using field data.

Study 1: Methods

Sample, design, and procedures

A total of 223 front-line employees (58.70% female, Mage = 32.38) were recruited online on Credamo to participate in this experiment in exchange for a monetary reward. All participants completed all study measures and passed the attention test. A single-factor (intelligent service strategy: intelligence substitution service strategy vs. intelligence collaboration service strategy) between-subjects design was adopted. All participants were randomly assigned into two groups. They read the descriptions of M Retail Company’s service strategies and were asked to imagine that they were supermarket clerks in a company (see Appendix A). M Retail Company is a fictitious company that implements an intelligent service strategy. To strengthen the effect of manipulation, all participants were asked to spend up to 5 min describing their feelings on the service strategy, without a word limit. Next, they completed the evaluations of human–human and human–machine interactivity, and perception of career achievement. After the experiment, all participants filled in demographic information such as gender, age, education and monthly income levels, and work experience with using AI in their previous jobs. As an experimental manipulation check, all participants were asked to assess the intelligent service strategy on a two-item scale (“In the company’s strategy, AI replaces the work of human employees (clerks),” and “In the company’s strategy, AI assists with and complements the work of human employees (clerks)” 1 = strongly disagree; 7 = strongly agree).

Measures

To evaluate the perception of career achievement, the participants answered six items: “I can still solve work-related problems effectively with my existing knowledge and skills,” “I feel that my work is still important to the business and customers,” “In my opinion, I can still use the knowledge and skills I am good at,” “I can still fully demonstrate the value of my work,” “I can still easily understand and deal with problems at work,” and “I feel that my work can still have an important impact on the business and customers.” These items were adapted from Prentice et al. (2019) and scored on a seven-point Likert-type scale (1 = strongly disagree; 7 = strongly agree).

The measure of human–human interactivity comprised five items: “I feel that the company’s customers frequently communicate with employees (clerks) face-to-face during the consumption process,” “I feel that the company advocates face-to-face communications and interactions between employees (clerks) and customers,” “I feel that communications and interactions between people are the basic service form of the company,” “The interaction form of ‘employee (clerk)–customer’ is common in the company’s service environment,” and “The company provides good service to customers through employees (clerks).” These items were adapted from Yavas et al. (2003) and Yoo and Arnold (2016) scored on a seven-point Likert-type scale (1 = strongly disagree; 7 = strongly agree).

The measure of human–machine interactivity comprised five items: “I feel that the company’s customers frequently use and connect with AI during the consumption process,” “I feel that the company advocates communications and interactions between customers and AI,” “I feel that communications and interactions between people and intelligent devices are the basic service form of the company,” “The interaction forms of ‘employee (clerk)–AI’ and ‘customer–AI’ are common in the company’s service environment,” and “The company provides good service to customers through AI.” These items were adapted from Yavas et al. (2003) and Yoo and Arnold (2016) scored on a seven-point Likert-type scale (1 = strongly disagree; 7 = strongly agree).

Given the possible confounding effects on the perception of career achievement (via human–human and human–machine interactivity), we controlled for gender (0 = male; 1 = female), education (0 = no college; 1 = college-educated), and experience using AI at work (0 = no experience using AI; 1 = with experience using AI), based on the following findings in the literature: First, there may be differences between men and women in their attitudes and evaluations of new technologies such as AI (Rosenthal-Von Der Pütten and Krämer, 2014). Second, education can provide domain-relevant knowledge (Baer et al., 2021), and the use of AI may devalue it. Third, employees who have experience using AI may have a deeper understanding of it (Lu et al., 2019) and may be more aware of the impact of AI on their jobs.

Study 1: Results and discussion

Reliability and validity analysis

Table 1 presents the descriptive statistics and correlations among the research variables. We conducted a Confirmatory Factor Analysis (CFA) to analyse the constructs’ reliability and validity. We estimated a three-factor model (i.e., human–human interactivity, human–machine interactivity, and perception of career achievement) using item-level indicators. Results revealed the following fit statistics: \({\chi }^{2}\) = 136.22 (p = 0.01), \({\chi }^{2}\)/DF = 1.35, GFI = 0.93, CFI = 0.99, IFI = 0.99, RMSEA = 0.04. The standardized factor loading of each variable ranged from 0.67 to 0.92 and the composite reliability (CR) was greater than 0.8, consistent with α (see Table 2). The average variance extraction (AVE) of each variable was greater than 0.5, thus the convergence validity was supported. Each square root of AVE (SAVE) was greater than the correlation coefficient between the corresponding variable and other variables, manifesting that each variable had good discriminant validity. In general, the measurement model was suitable for further analysis.

Table 1 Study 1: Descriptive statistics and correlations.
Table 2 Study 1: Reliability and validity.

Manipulation check

The results of the independent sample t-test showed that when compared to those in the intelligence collaboration service strategy group (M = 2.19, SD = 0.83), the participants in the intelligence substitution service strategy group (M = 5.91, SD = 1.07) had a stronger perception that the organization was implementing the intelligent service strategy to replace employees (t (221) = 29.05, p < 0.001). In contrast, compared to those in the intelligence substitution service strategy group (M = 2.04, SD = 0.93), the participants in the intelligence collaboration service strategy group (M = 5.90, SD = 0.72) had a stronger perception that the organization was implementing the intelligent service strategy to cooperate with employees (t(221) = 34.63, p < 0.001). The manipulation of the intelligent service strategy was successful.

Test of hypotheses

The results of the analysis of variance (ANOVA) showed that when compared with those in the organizational environment with the intelligence collaboration service strategy (M = 5.59, SD = 0.74), the employees in the organizational environment with the intelligence substitution service strategy (M = 4.46, SD = 1.44) had a lower perception of career achievement (F (1, 221) = 52.76, p < 0.001, partial η2 = 0.19), which was consistent with Hypothesis 1 (see Table 3 for the regression analysis).

Table 3 Study 1: Regression results.

The MANOVA results showed that when compared with the organizational environment with the intelligence collaboration service strategy (MI = 5.17, SDI = 0.89; MH = 5.26, SDH = 0.76), the employees in the organizational environment with the intelligence substitution service strategy had lower human–human (MI = 2.68, SDI = 1.32, F (1, 221) = 270.24, p < 0.001, partial η2 = 0.55) and higher human–machine interactivity (MH = 5.94, SDH = 0.66, F (1, 221) = 50.47, p < 0.001, partial η2 = 0.19). These were consistent with Hypothesis 2a and Hypothesis 2b.

The results of regression analysis showed that human–human (b = 0.35, SE = 0.05, p < 0.001) and human–machine interactivity (b = −0.26, SE = 0.11, p < 0.05) had a positive and negative effect on the perception of career achievement, respectively, supporting Hypothesis 3a and Hypothesis 3b.

Finally, PROCESS Model 4 and 5000 bootstrapped resamples were used to test the mediation effects (Hayes, 2018). As shown in Fig. 2, the direct effect of the intelligent service strategy on the perception of career achievement (path: intelligent service strategy → perception of career achievement) was not significant since the confidence interval contained 0 (b = 0.19, SE = 0.21, 95% CI [−0.23, 0.61]). The indirect effects through human–human interactivity (path: intelligent service strategy → human–human interactivity → perception of career achievement; b = 0.76, SE = 0.18, 95% CI [0.41, 1.13]) and human–machine interactivity (path: intelligent service strategy → human–machine interactivity → perception of career achievement; b = 0.17, SE = 0.07, 95% CI [0.06, 0.33]) were both significant. Since the confidence interval of the indirect effects excluded 0, the mediating effect of human–human interactivity and human–machine interactivity were qualified.

Fig. 2: Study 1: Mediating effect of interactivity.
figure 2

Figure shows the mediating role of human–human interactivity and human–machine interactivity.

Discussion

Study 1 provides preliminary support for the effect of enterprises’ intelligent service strategy on the perception of career achievement. The intelligence substitution service strategy (vs. intelligence collaboration service strategy) results in lower human–human and higher human–machine interactivity, indirectly decreasing employees’ perception of career achievement. These results still hold after controlling for gender, education, and work experience with using AI. These findings highlight the negative consequences of increased AI adoption for employees, aligning with Yam et al. (2023) and Tang et al. (2022b). Unlike these studies, which examined the impact of AI on job insecurity and job performance, our study emphasizes its impact on employees’ intrinsic motivation (i.e., perception of career achievement). More importantly, our findings reveal the positive impact of AI technology applications. Intelligence collaboration service strategy can enhance employees’ perception of career achievement. This finding also aligns with research that maintains a positive attitude towards AI applications, such as improving work efficiency (Tong et al., 2021).

Study 2: Methods

Sample, design, and procedures

A total of 232 front-line employees (63.40% female, Mage = 30.32) were recruited online on Credamo to participate in this experiment in exchange for a monetary reward. All participants completed all study measures and passed the attention test. They were randomly assigned to a 2 (intelligent service strategy: intelligence substitution service strategy vs. intelligence collaboration service strategy) × 2 (organizational innovation climate: high vs. low) between-subjects experimental design. All participants were asked to imagine they were insurance consultants at M Insurance Company. M Insurance Company is a fictitious company that implements an intelligent service strategy. They read a description of M company’s implementation of the intelligent service strategy and completed the evaluations of human–human and human–machine interactivity. To strengthen the effect of the manipulation, all participants were required to spend up to 5 min carefully describing their feelings and views on M company’s intelligent service strategy, with no word limit. As an experimental manipulation check, all participants were asked to assess the intelligent service strategy on a two-item scale, the same as in Study 1. Next, the participants read a description of M company’s innovation climate (see Appendix B). The participants then completed the measurements of organizational innovation climate and perception of career achievement on a seven-point Likert-type scale. At the end of the experiment, the participants filled in demographic information such as gender, age, education, monthly income level, and work experience with using AI in their previous jobs.

Measures

The evaluations of the perception of career achievement, and human–human and human–machine interactivity were the same as those in Study 1. The measurement of organizational innovation climate included six items: “M company advocates new attempts and encourages new mindsets and creativity,” “M company appreciates and recognizes the use of new technologies and methods by employees to provide innovative services to customers,” “M company’s managers support and encourage subordinates to express their new ideas,” “M company can provide sufficient resources such as new equipment and technologies to enable employees to do innovative work,” “M company’s reward system encourages innovation,” and “M company’s employees have enough time to pursue creative ideas.” The items of organizational innovation climate were adapted from Scott and Bruce (1994) and scored on a seven-point Likert-type scale (1 = strongly disagree; 7 = strongly agree).

Study 2: Results and discussion

Reliability and validity analysis

Table 4 presents the descriptive statistics and correlations among the research variables. We conducted a confirmatory factor analysis (CFA) to analyse the constructs’ reliability and validity. We estimated a three-factor model (i.e., human–human interactivity, human–machine interactivity, and perception of career achievement) using item-level indicators. Results revealed the following fit statistics: \({\chi }^{2}\) = 174.49 (p = 0.00), \({\chi }^{2}\)/DF = 1.73, GFI = 0.92, CFI = 0.97, IFI = 0.97, RMSEA = 0.06. The standardized factor loading of each variable ranged from 0.67 to 0.90 and the composite reliability (CR) was greater than 0.8, consistent with α (see Table 5). The average variance extraction (AVE) of each variable was greater than 0.5, thus the convergence validity was supported. Each square root of AVE (SAVE) was greater than the correlation coefficient between the corresponding variable and other variables, manifesting that each variable had good discriminant validity. In general, the measurement model was suitable for further analysis.

Table 4 Study 2: Descriptive statistics and correlations.
Table 5 Study 2: Reliability and validity.

Manipulation check: intelligent service strategy

The results of the independent sample t-test showed that when compared to those in the intelligence collaboration service strategy group (M = 2.14, SD = 0.90), the participants in the intelligence substitution service strategy group (M = 5.89, SD = 1.01) had a stronger perception that the organization was implementing the intelligent service strategy to replace employees (t (230) = 29.92, p < 0.001). By contrast, compared to those in the intelligence substitution service strategy group (M = 2.07, SD = 1.02), the participants in the intelligence collaboration service strategy group (M = 5.96, SD = 0.84) had a stronger perception that the organization was implementing the intelligent service strategy to cooperate with employees (t (230) = 31.86, p < 0.001). The manipulation of intelligent service strategy was successful.

Manipulation check: organizational innovation climate

The results of the independent sample t-test showed that when compared to those in the low organizational innovation climate group (M = 2.93, SD = 1.50), the participants in the high one (M = 6.02, SD = 0.46) had a stronger perception of organizational innovation climate (t (230) = 20.80, p < 0.001). The manipulation of the organizational innovation climate was successful.

Test of hypotheses

The ANOVA results showed that when compared to those in the organizational environment with the intelligence collaboration service strategy (M = 5.32, SD = 1.16), the employees in the organizational environment with the intelligence substitution service strategy (M = 4.84, SD = 1.38) had a lower perception of career achievement (F (1, 230) = 8.31, p < 0.01, partial η2 = 0.04), which verified Hypothesis 1 (see Table 6 for regression analysis).

Table 6 Study 2: Regression results.

The MANOVA results showed that when compared to their counterparts in the environment with the intelligence collaboration service strategy (MI = 5.33, SDI = 0.93; MH = 4.77, SDH = 1.09), the employees in the environment with the intelligence substitution service strategy had lower human–human (MI = 3.63, SDI = 1.58, F (1, 230) = 100.93, p < 0.001, partial η2 = 0.31) and higher human–machine interactivity (MH = 5.69, SDH = 0.88, F (1, 230) = 49.19, p < 0.001, partial η2 = 0.18), supporting Hypothesis 2a and Hypothesis 2b.

The results of regression analysis showed that human–human (b = 0.20, SE = 0.06, p < 0.01) and human–machine interactivity (b = −0.21, SE = 0.09, p < 0.05) had a positive and negative effect on the perception of career achievement, respectively. Hypothesis 3a and Hypothesis 3b were supported.

PROCESS Model 4 and 5000 bootstrapped resamples were used to test the mediation effects (Hayes, 2018). The direct effect of the intelligent service strategy on the perception of career achievement (path: intelligent service strategy → perception of career achievement) was not significant since the confidence interval contained 0 (b = −0.07, SE = 0.19, 95% CI [−0.45, 0.30]). The indirect effects through human–human interactivity (path: intelligent service strategy → human–human interactivity → perception of career achievement; b = 0.35, SE = 0.15, 95% CI [0.07, 0.66]) and human–machine interactivity (path: intelligent service strategy → human–machine interactivity → perception of career achievement; b = 0.20, SE = 0.10, 95% CI [0.02, 0.42]) were both significant. Since the confidence interval of the indirect effects excluded 0, the mediating effect of human–human interactivity and human–machine interactivity were qualified.

A 2 × 2 ANOVA (see Fig. 3) revealed the significant interaction between organizational innovation climate and intelligent service strategy (F (1, 228) = 8.91, p < 0.01). Under the condition of the intelligence substitution service strategy, the perception of career achievement was lower in the low organizational innovation climate group than in the high one (Mlow = 4.09, SDlow = 1.40; Mhigh = 5.60, SDhigh = 0.84; F (1, 228) = 48.45, p < 0.001). Under the condition of the intelligence collaboration service strategy, the perception of career achievement was lower in the low organizational innovation climate group than in the high one (Mlow = 5.04, SDlow = 1.26; Mhigh = 5.65, SDhigh = 0.95; F (1, 228) = 8.63, p < 0.01).

Fig. 3: Study 2: Moderating effect of organizational innovation climate.
figure 3

Figure shows the moderating role of organizational innovation climate in the relationship between interactivity and perception of career achievement.

To assess the moderating effect of organizational innovation climate, we conducted a moderated mediation analysis using a Hayes (2018) PROCESS model (Model 14, 5000 bootstrapped resamples, 95% confidence level; see Table 7). The moderating effect of organizational innovation climate on the path of human-human interactivity → perception of career achievement was significant, as the confidence interval excluded 0 (b = −0.22, SE = 0.11, 95% CI [−0.44, −0.01]), supporting Hypothesis 4a. Participants considered the intelligence collaboration service strategy more human–human interactivity (b = 1.72, SE = 0.17, p < 0.001) than the intelligence substitution service strategy, which increased their perception of career achievement (b = 0.20, SE = 0.06, p < 0.001). Human–human interactivity enhanced the perception of career achievement among participants in an environment with low organizational innovation climate (b = 0.33, SE = 0.08, p < 0.001), but not in an environment with high organizational innovation climate (b = 0.09, SE = 0.08, p = 0.27).

Table 7 Study 2: Regression output of PROCESS model 14 analysis for perception of career achievement.

The moderating effect of organizational innovation climate on the path of human-machine interactivity → perception of career achievement was significant, as the confidence interval excluded 0 (b = 0.40, SE = 0.16, 95% CI [0.09, 0.72]), supporting Hypothesis 4b. Participants considered the intelligence collaboration service strategy less human–machine interactivity (b = −0.94, SE = 0.13, p < 0.001) than the intelligence substitution service strategy, which increased their perception of career achievement (b = −0.26, SE = 0.08, p < 0.001). Human–machine interactivity decreased the perception of career achievement among participants in an environment with low organizational innovation climate (b = −0.47, SE = 0.12, p < 0.001), but not in an environment with high organizational innovation climate (b = −0.05, SE = 0.10, p = 0.64).

Discussion

The results of Study 2 show that organizational innovation climate moderates the impact of intelligent service strategy on the perception of career achievement. Hypothesis 4a and Hypothesis 4b are verified. Organizational innovation climate reduces and increases employees’ perception of career achievement based on human–human and human–machine interactivity, respectively. Consistent with Marshall et al. (2019) and Zhang et al. (2022), our study also uncovers the moderating role of organizational innovation climate on employee psychology and behavior. However, unlike these studies that focus on the impact of organizational innovation climate on employee innovation behavior, our research further expands the scope of organizational innovation climate to AI-related research and identifies a new mechanism by which organizational innovation climate influences employees’ perception of career achievement.

Study 3: Methods

Sample, design, and procedures

To ensure more reliable inputs, we specifically targeted respondents with front-line work experience. Consequently, the experiment utilized a convenience sampling methodology to recruit 290 employees from a provincial branch of the Agricultural Bank of China to participate. Participants who failed the attention test were excluded, resulting in a valid sample of 281 (63.70% female, Mage = 31.69). A 2 (intelligent service strategy: intelligence substitution service strategy vs. intelligence collaboration service strategy) × 2 (organizational innovation climate: high vs. low) between-subjects experimental design was adopted. All participants were randomly assigned to four groups. First, they read descriptions of the bank’s service strategies. To strengthen the manipulation effect, all participants were required to spend up to 5 min carefully describing their feelings and views on the Agricultural Bank of China’s intelligent service strategy, with no word limit. Subsequently, they were asked to complete evaluations of human–human and human–machine interactivity. Second, participants read a description of the bank’s organizational innovation climate and completed measurements assessing organizational innovation climate and perception of career achievement. The procedures and measurements were similar to those in Study 2.

Finally, respondents were asked to assess the intelligent service strategy on a two-item scale as the manipulation check and to answer questions about demographic information, identical to those used in Study 1.

Study 3: Results and discussion

Manipulation check: intelligent service strategy

The results of the independent sample t-test showed that when compared to those in the intelligence collaboration service strategy group (M = 2.48, SD = 1.42), the participants in the intelligence substitution service strategy group (M = 5.77, SD = 1.33) had a stronger perception that the organization was implementing the intelligent service strategy to replace employees (t (279) = 20.03, p < 0.001). By contrast, compared to those in the intelligence substitution service strategy group (M = 3.76, SD = 2.01), the participants in the intelligence collaboration service strategy group (M = 6.15, SD = 0.83) had a stronger perception that the organization was implementing the intelligent service strategy to cooperate with employees (t (279) = 13.17, p < 0.001). The manipulation of intelligent service strategy was successful.

Manipulation check: organizational innovation climate

The results of the independent sample t-test showed that when compared to those in the low organizational innovation climate group (M = 2.97, SD = 1.66), the participants in the high one (M = 5.91, SD = 0.61) had a stronger perception of organizational innovation climate (t (279) = 20.18, p < 0.001). The manipulation of the organizational innovation climate was successful.

Reliability and validity analysis

Table 8 presents the descriptive statistics and correlations among the research variables. Partial least squares structural equation modeling (PLS-SEM) was used to validate the research model. The results of the confirmatory factor analysis are presented in Table 9, revealing that the majority of factor loadings exceeded 0.7. One item of the human–machine interactivity (i.e., MI5) exhibited a factor loading slightly below 0.7 but above 0.6; it was retained because it was important to the relevant factor. The composite reliability (CR) was greater than 0.8, consistent with α. The average variance extraction (AVE) of each variable was greater than 0.5, thus supporting convergence validity. Each square root of AVE (SAVE) surpassed the correlation coefficient between the corresponding variable and other variables. Furthermore, all the heterotrait-monotrait (HTMT) values were below the cut-off value of 0.85 (see Table 10), indicating that each variable had good discriminant validity.

Table 8 Study 3: Descriptive statistics and correlations.
Table 9 Study 3: Reliability and validity.
Table 10 Study 3: Discriminant validity assessment (HTMT).

Structural model evaluation

PLS-SEM analyses were conducted, encompassing the measurement models of intelligent service strategy, human–human and human–machine interactivity, employees’ perception of career achievement, and organizational innovation climate. Control variables, including gender, education, and work experience with using AI, were specified as predictors of employees’ perception of career achievement. First, Fig. 4 illustrates the competing model analysis. The model investigated the relationship between intelligent service strategy (1 = collaboration, 0 = substitution) and employees’ perception of career achievement, with human–human interactivity and human–machine interactivity acting as mediators and organizational innovation climate (1 = high, 0 = low) acting as the moderator. To model the moderating effects, we followed the approach of Al-Gahtani et al. (2007) by multiplying the corresponding indicators of the independent and moderator variables. Intelligent service strategy had a positive (total) effect on the perception of career achievement (b = 0.73, SD = 0.13, p < 0.001), supporting Hypothesis 1. Intelligent service strategy had a positive effect on human–human interactivity (b = 0.78, SD = 0.11, p < 0.001) and a negative effect on human–machine interactivity (b = −0.69, SD = 0.10, p < 0.001). Human–human (b = 0.60, SD = 0.11, p < 0.001) and human–machine interactivity (b = −0.30, SD = 0.10, p < 0.01) had a positive and negative effect on the perception of career achievement, respectively. Thus, Hypotheses 2–3 were also supported. Regarding the moderator variables, organizational innovation climate had a negative interacting effect with human–human interactivity (b = −0.29, SD = 0.14, p < 0.05) and a positive interacting effect with human–machine interactivity (b = 0.26, SD = 0.11, p < 0.05) on perception of career achievement, respectively. These results supported Hypothesis 4. In addition, the indirect effect of intelligent service strategy on the perception of career achievement via human–human interactivity (b = 0.47, SD = 0.10, p < 0.001) and human–machine interactivity (b = 0.20, SD = 0.07, p < 0.01) were also supported by the results. However, the direct effect of intelligent service strategy on the perception of career achievement was insignificant (b = 0.06, SD = 0.10, p = 0.56).

Fig. 4: Study 3: Structural model results.
figure 4

Figure shows the structural model results using the PLS-SEM analysis method.

The results showed that the coefficient of determination (R2) of perception of career achievement was 0.41, indicating the model can explain variance in the perception of career achievement to some extent. The value of Q2 (0.10 for the construct cross-validated redundancy) was above 0, indicating the model can demonstrate sufficient predictive relevance. The results of the structural model with moderator variables are shown in Table 11.

Table 11 Study 3: Results of the structural model.

Multigroup analysis

Multigroup analysis was employed to further assess whether the effect of intelligent service strategy on employees’ perception of career achievement, mediated by human–human and human–machine interactivity, is moderated by organizational innovation climate. The results of the path coefficients between different groups are presented in Table 12. The results showed that perception of career achievement was strongly influenced by human–human interactivity (blow = 0.58, SDlow = 0.09, p < 0.001) and human–machine interactivity (blow = −0.29, SDlow = 0.08, p = 0.001) in an environment with a low organizational innovation climate. The moderated mediating effect through human–human interactivity and human–machine interactivity was stronger for the low organizational innovation climate group, and these differences were significant at p < 0.05. Hypothesis 4 was supported again.

Table 12 Study 3: The moderating effect of organizational innovation climate (multigroup analysis).

Discussion

Study 3, through field experiments, reaffirms that the enterprise’s intelligence substitution service strategy (vs. intelligence collaboration service strategy) reduces employees’ perception of career achievement by decreasing human–human interactivity and increasing human–machine interactivity. Hypotheses 1–3 are retested. Furthermore, Study 3 reconfirms Hypothesis 4, indicating that the organizational innovation climate moderates the impact of human–human and human–machine interactivity on employees’ perception of career achievement. Table 13 summarizes the experimental design and the hypothesis-supported results for all three studies.

Table 13 Study 3: Summary of experimental design and support for the hypotheses.

General discussion

In some ways, AI may not differ much from previous technologies, as technological change persists for organizations (Tang et al., 2022b). The impact of AI on employees’ perception of career achievement, however, may not be entirely positive. To reduce labor costs and achieve standardized management, enterprises use AI to replace human labor (Fountaine et al., 2019), which may rapidly weaken employees’ perception of career achievement. Enterprises aim to use AI to help employees complete their work more efficiently and advance the development of human–machine collaboration (Marinova et al., 2017; Tong et al., 2021), potentially enhancing their perception of career achievement. This study examines the impact of enterprises’ intelligent service strategy on employees’ perception of career achievement from this contradictory perspective. When the intelligence substitution service strategy (vs. intelligence collaboration service strategy) is adopted, employees’ perception of career achievement drops sharply. Two online experiments and one field experiment prove that this result stems from the intelligence substitution service strategy (vs. intelligence collaboration service strategy) reducing the human–human interactivity and enhancing human–machine interactivity. Human–human interactivity raises, and human–machine interactivity weakens, employees’ perception of career achievement. Organizational innovation climate plays an important moderating role in this process. To ensure the managerial relevance and robustness of the findings, this research covers service firms with experience (Study 1) and credence attributes (Study 2 and Study 3).

Theoretical implications

The impact of AI technology on people’s willingness to accept service experiences has been a focus of inquiry in recent years. However, this research has a different perspective by exploring the impact of enterprises’ intelligent service strategy on employees, especially on their perception of career achievement. This research makes the following main contributions. First, this study enriches the literature on the widespread use of AI in enterprises and its impact on employees, with a particular focus on the intrinsic motivation of front-line employees. The widespread application of AI technology in service has emerged as one of the most significant changes in enterprise service management (Liang et al., 2022; Mende et al., 2019). AI is progressively replacing human employees in the service industries for social interaction with consumers (Huang and Rust, 2018). Previous studies have predominantly explored the impact of AI technology on customer service experience or consumers’ cognitive preference for AI from the perspective of recipients (e.g., Garvey et al., 2023; Mende et al., 2019). However, little literature addresses how employees perceive changes in their jobs in an AI service environment. Research on employees has typically centered on functional advantages (e.g., Raisch and Krakowski, 2021; Wilson and Daugherty, 2018) and factors affecting employee job performance (e.g., Qiu et al., 2022; Song et al., 2022; Tang et al., 2022b), overlooking the effects of AI technology on employees’ intrinsic motivation. Given the importance of the perception of career achievement for front-line employees, our study addresses this research gap by drawing on the perspective of the self-perception theory. Constructing a theoretical model of “intelligent service strategy–human–human and human–machine interactivity–perception of career achievement” provides a specific analytical framework to elucidate how intelligent service strategy impacts employees’ intrinsic motivation. Our study establishes the theoretical foundation for future research to understand employees’ perception of career achievement in an intelligent service environment, which is a relatively unexplored area. The widespread application of AI technology across various industries, especially in service sectors, further strengthens the theoretical contribution of the research framework.

Second, our research deepens the understanding of research related to the influence of AI adoption on employees. Prior research on the impact of AI adoption mainly focused on a single level, either focusing on the positive effects of AI adoption, such as improving employees’ working efficiency (e.g., Tong et al., 2021), or the negative effects, such as reducing employees’ job performance and inhibiting their service performance (e.g., Liang et al., 2022; Tang et al., 2022b). However, few empirical studies have delved into the reasons behind this difference. Our research expands AI-related literature by showcasing the potential impact of enterprise intelligent service strategy on employees’ perception of career achievement. On the one hand, we propose that enterprises can adopt two distinct strategies to establish an intelligent service environment: intelligence substitution and collaboration service strategies. Both strategies can change how employees perceive the interactive environment within an enterprise, explaining changes in their perception of career achievement. On the other hand, by providing a framework that explains the negative and positive impacts of intelligent service strategy on employees’ perception of career achievement, our research reveals the distinct paths of intelligence substitution service strategy and intelligence collaboration service strategy on employees’ perception of career achievement, thus enriching the existing research on employees’ intrinsic motivation. Our findings provide novel insights that hold the potential to advance thinking on AI adoption.

Third, our study contributes to the AI literature by uncovering the conditions under which intelligent service strategy either inhibits or promotes employees’ perception of career achievement. Specifically, by examining organizational innovation climate as the boundary condition, the current research proposes that organizational innovation climate plays a crucial role in influencing the relationship between human–human interactivity and human–machine interactivity and perception of career achievement. Organizational innovation climate will not only alleviate the positive impact of human–human interactivity on employees’ perception of career achievement but also mitigate the negative impact of human–machine interactivity on employees’ perception of career achievement. Given the differential impact of intelligence substitution service strategy and intelligence collaboration service strategy on employees’ perception of career achievement, it is important to understand when and how intelligent service strategy results in differential outcomes. In addition, research on the impact of AI adoption on employees is still in its infancy, and the previous research has mainly focused on how employees’ perceptions of human–machine relationships or the application of AI technology affect their work performance or adoption of AI technology (Prentice et al., 2019; Tang et al., 2022a). Most studies focused on the individual level, leaving the relationship between the organizational environment and individual cognition levels largely unexplored. This study introduces organizational innovation climate to examine how intelligent service strategy (organizational level) affects employees’ perception of career achievement (individual level), providing innovative theoretical ideas for mitigating the negative impacts of AI adoption.

Implications for practice

Within the service industry, companies make substantial investments in innovating and standardizing services to attract consumers and improve service quality. Our findings do not dispute the value of intelligent service applications. However, they serve as a cautionary note for companies: While promoting and establishing an intelligent service environment, they cannot ignore the impacts of such an environment on employees. Therefore, the conclusions drawn from this study bear significant implications for the strategic decision-making of service enterprises.

First, it is crucial to recognize that the adoption of AI is a double-edged sword, capable of either enhancing employees’ perception of career achievement or decreasing it. This outcome depends on the type of intelligent service strategy adopted and employees’ perception of interactivity within the organization. These findings provide valuable insights for service enterprises in strategic decision-making. On the one hand, during the process of building an AI service environment, enterprises should prioritize the choice of an appropriate intelligent service strategy. Our findings show that, compared with the intelligence substitution service strategy, the intelligence collaboration service strategy can enhance employees’ perception of career achievement. Therefore, for enterprises aspiring to leverage AI technology for innovation and reforms in the service environment, opting for the intelligence collaboration service strategy is a more advisable choice. On the other hand, service companies must recognize that in a dynamic AI service environment, employees’ perceptions of interactivity can undergo significant changes. Given the differential effects of human–human interactivity and human–machine interactivity on front-line employees’ perception of career achievement, we provide suggestions on guiding enterprises to foster work value and meaning for front-line employees. When introducing and developing AI services, service companies need to pay attention to enhancing human–human interactivity in the enterprise to alleviate employees’ doubts about the meaning and value of their work arising from increased human–machine interactivity. For example, companies can organize regular networking events or team-building activities to boost employee-to-employee interaction.

Second, this research identifies organizational innovation climate as an important boundary condition influencing the relationship between interactivity and perception of career achievement. In particular, organizational innovation climate can alleviate the negative impact of human–machine interactivity on employees’ perception of career achievement. Therefore, fostering a strong organizational innovation climate can help alleviate employees’ sense of loss and anxiety due to increased AI applications and human–machine interactions. During the establishment of an AI service environment, enterprises can continuously bolster innovative ideas and the innovation climate through recognition, reward, and systemic reform. It is necessary to make employees understand that the organization supports and encourages the optimization and improvement of their original work with AI technology. Consequently, to a certain extent, this can alleviate employees’ diminished perception of career achievement resulting from their inability to leverage their original specialities and skills. Meanwhile, managers should provide employees with adequate learning and career development opportunities, resources, and other organizational support. These resources will significantly assist employees in acquiring new skills and developing new competencies. This approach enables employees to adapt their abilities to the evolving workplace and enhance their perception of career achievement by improving their professional knowledge and skills.

Limitations and future research directions

First, this research discusses employees’ perception of interactivity in the AI service environment from the perspectives of employees–customers and employees (customers)–AI. Prior studies have pointed out that customers and employees interact with each other (Guerrero et al., 2018; Nasr et al., 2015). Consequently, the three-way interaction form of consumers–AI–employees can affect and change employees’ behavior and cognitive performance (Song et al., 2022). Deepening the classification of interactivity can enhance the understanding of employees’ behavior and psychology in the AI service field. Second, this research explores the impact of employees’ perception of interactivity in the AI service environment on their perception of career achievement. However, factors affecting employees’ perception of career achievement may come from external information and feedback because customers play a pivotal role in service production and transactions (Prentice et al., 2019). Future research may consider how customers’ evaluations of enterprises’ AI services affect employees’ perception of career achievement. Third, individual differences may have affected the results. Employees’ psychological job demand is also an important factor that affects employee behavior. Research has found that the impact of AI on employees’ psychology and emotions depends on the degree to which employees need and use these resources in their work (Qiu et al., 2022). One key factor in determining this “degree” is psychological job demand (Ariza-Montes et al., 2018). Therefore, another area worthy of further exploration is the moderating effect of psychological job demand.

Conclusion

The integration of AI and services represents the prevailing trend in current technological development and application. The application of AI technology, however, has a dual impact on employees’ perception of career achievement, especially for front-line employees in the service industry. Against this background, we conducted three studies to develop and validate a conceptual model assessing the impact of enterprise intelligent service strategies on front-line employees’ perception of career achievement. Our findings indicate that the intelligence substitution service strategy (vs. intelligence collaboration service strategy) reduces front-line employees’ perception of career achievement, both directly and indirectly, primarily due to decreased human–human interactivity and increased human–machine interactivity. Human–human interactivity positively influences employees’ perception of career achievement, while human–machine interactivity exerts a negative influence. Study 2, employing a 2 (intelligent service strategy: intelligence substitution service strategy vs. intelligence collaboration service strategy) × 2 (organizational innovation climate: high vs. low) between-subjects experimental design, further discovers that the organizational innovation climate alleviates the negative impact of human–machine interactivity and mitigates the positive impact of human–human interactivity has on employees’ perception of career achievement. Study 3, through field data collection, reaffirmes the influence mechanism of enterprises’ intelligent service strategy on employees’ perception of career achievement. Our conceptual model and empirical findings provide novel insights for implementing enterprise intelligent service strategies and inspire further research on employee intrinsic motivation within intelligent service environments, contributing to the ongoing conversation about the new era of service.