1 s2.0 S0148296322010748 Main

Journal of Business Research 157 (2023) 113609
Contents lists available at ScienceDirect
Journal of Business Research

journal homepage: www.elsevier.com/locate/jbusres
The impact of artificial intelligence capabilities on servitization: The

moderating role of absorptive capacity-A dynamic capabilities perspective
Mohamad Abou-Foul a, *, Jose L. Ruiz-Alba b, Pablo J. López-Tenorio c
a
Al-Azhar University, Jamal Abdl, Naser St, Gaza, Palestine
b
University of Westminster, 35, Marylebone Road, London NW1 5LS, UK
c
UNIE Universidad, Facultad de Ciencias Sociales Aplicadas y de la Comunicación C/ Arapiles, 14. 28015. Chamberí. Madrid., Spain
A R T I C L E I N F O A B S T R A C T
Keywords: The advent of artificial intelligence (AI)-based technologies has opened new opportunities for manufacturers to
Artificial intelligence maintain their technological edge and address pressing societal challenges. This research investigates the nature
Servitization of the relationships between AI capabilities, servitization, and the role of absorptive capacity. Building on dy
Social innovation
namic capabilities literature, we developed and empirically tested a model using structural equation modeling
Fuzzy set qualitative comparative analysis
(fsQCA)
(SEM) and further applied a fuzzy-set qualitative comparative analysis (fsQCA). Through the construct of AI
Dynamic capabilities capabilities and its four sub-dimensions, we find supportive evidence from our model estimates employing data
from 185 manufacturing firms in the US and EU. The study findings highlight the positive impact of AI capa
bilities on servitization; this relationship is positively moderated by absorptive capacity. Furthermore, the road to
servitization is through advancing AI capabilities related to internal process and resource optimization coupled
with AI for social innovation services. The study’s theoretical and pragmatic implications are discussed.
1. Introduction et al., 2019; Spring & Araujo, 2017).

Current development in machine learning (ML) and AI have super
A burning question for servitization researchers and practitioners is charged the innovation process and service breakthroughs, which has
how artificial intelligence (AI) can be incorporated to enhance opera far-reaching business and societal consequences. This AI development is
tional efficiency, market offerings, customer experience, and social driven by the explosion in available and accessible data warehouses,
innovation (Haefner et al., 2021). Therefore, AI is increasingly becoming along with sensor connectivity data facilitated by the internet of things
a focal point for manufacturers’ innovation debate, which is in line with (IoT), which, in some cases, leads to digital servitization (Kohtamäki
the main challenges articulated and enriched by B2B marketing theory, et al., 2019). Therefore, AI should be valued as a business capability
in terms of technology advancement, value creation, analytics, and rather than a mere technological advancement (Davenport & Ronanki,
overall ecosystem innovation (Mora Cortez & Johnston, 2017). 2018). Capabilities can be defined as “complex bundles of skills and
In the B2B context, firms have widely viewed servitization as “a accumulated knowledge, exercised through organizational processes,
transformational process whereby a company shifts from a product- that enable firms to coordinate activities and make use of their assets”
centric to a service-centric business model and logic” (Kowalkowski (Day, 1994p.38). Therefore, data are a valuable asset that can help in
et al., 2017a, p. 8). Foundational research in this area has identified the building the enterprise AI ecosystem that enhances service innovation.
servitization business model configuration (Palo et al., 2019), categories Despite the growing body of literature which sporadically examines
of resources and capabilities that must underpin the transition to ser the strategic value of digital transformation on a company service
vices (Baines & Lightfoot, 2014; Dmitrijeva et al., 2020; Kanninen et al., transition (Abou-foul et al., 2021; Gebauer et al., 2021; Mikalef & Gupta,
2017b; Ulaga & Reinartz, 2011). Another stream of servitization liter 2021) there are still debates about the potential of AI capabilities -as an
ature has argued that servitization increasingly constitutes a socio imperative extension to digital transformation and its business-to-
technical dimension that must be addressed in the current business business (B2B) organizational constitutions and configurations (Sjödin
ecosystem buzzing with sustainability incentives, intelligent decarbon et al., 2021).
ization solutions, and powerful government regulations (Huntingford One source of these debates is the fact that the use of broad
* Corresponding author.
E-mail addresses: [email protected] (M. Abou-Foul), [email protected] (J.L. Ruiz-Alba).
https://doi.org/10.1016/j.jbusres.2022.113609
Received 12 October 2021; Received in revised form 21 December 2022; Accepted 23 December 2022
Available online 28 December 2022
0148-2963/© 2022 Elsevier Inc. All rights reserved.
M. Abou-Foul et al. Journal of Business Research 157 (2023) 113609
information technology constructs coupled with the availability of suc patterns and optimal configurations that leverage servitization.
cessive generations of new technologies precluded the advancement of The paper proceeds as follows: the next section reviews the theo
consistent, explicit, readily comparable empirical studies on the in retical underpinnings of servitization, AI capabilities, and absorptive
terdependencies between AI artifacts and servitization, rendering pre capacity, as well as the hypothesized relationships. Section 3 describes
vious research impractical (Desouza et al., 2020; Mikalef & Gupta, the research methodology with construct operationalization. Section 4
2021). Additionally, the literature mostly viewed technological capa presents the study’s analyses and the results of the proposed structural
bilities from an inward perspective arising from the IT division, without model. Section 5 discusses the research findings and dwells on their
considering the role of business customers in strategically harnessing theoretical and pragmatic ramifications and, finally, the study concludes
technological capabilities, resulting in a widening gap in the literature by examining research limitations and future avenues of research.
concerning the conceptual association of AI with other theoretical as
pects. Confounding the debates further is the fact that popular AI ca 2. Theory and hypotheses
pabilities literature focuses on firm-level analysis, which, in some cases,
obscures the real impact of AI on specific business processes and limits 2.1. The dynamic capabilities perspective
our understanding of its theoretical and practical interconnections
(Iansiti & Lakhani, 2020; Kohtamäki et al., 2019). Prior literature also The main premise of dynamic capabilities theory (DCT) is that, for
fails to identify those higher-order dynamic capabilities required to companies to achieve superior performance in volatile industries, they
successfully change and adapt (Teece, 2007) to avoid the infamous must possess higher-order capabilities – namely, adaptive processes and
competency trap (Barnett & Pontikes, 2008). structures that enable them to change their baseline capabilities so that
The convergence of advanced cognitive computing manifested in AI they can sense, seize, and adapt to an ever-evolving competitive land
technologies and servitization shows that the latter represents a good scape (Felin & Powell, 2016). Teece (2007) advanced the threefold
example of the successful implementation of the former (Agarwal et al., classification of firm-level dynamic capabilities, namely sensing, seizing,
2022). The scholarly interest in the impact of AI on servitization is still and reconfiguring, required to reach the optimal enterprise structure,
embryonic and lacks empirical evidence (Wirtz et al., 2018), the central continuous innovation anticipation, and knowledge management. While
premise of AI and machine learning is that it can enhance efficiency and the business landscape is shifting to more data-driven enterprises, AI
better customization for value proposition (Haefner et al., 2021), with cognitive applications on the lower end of the spectrum are altering
the vast majority of servitization research built on the interpretation of customer value propositions, and operational efficiency, but, more
the radical transformation process while continuing to stress innovation importantly, they are helping to tackle real business problems and
and technology implementation that supports the creation of value- pressing society challenges (Dangelico et al., 2016).
added (Garcia Martin et al., 2019). To date, AI has been widely Following this line of argumentation, we maintain that cognitive
viewed as a radical source of reform of patterns of production and technologies are a source of idiosyncratic, higher-order capabilities and
enhancing the learning process that complements decision support ultimately impact the micro-foundation of the dynamic capabilities
systems in place (Kasie et al., 2017; Mikalef & Gupta, 2021). AI impact framework because it helps companies change their baseline capabilities
on service provision is still underdeveloped as advancing a firm’s AI in areas such as product design, customer service, and manufacturing
capabilities requires an incremental investment in complex technology processes. In this regard, we are echoing the view of Hercheui and
and a delicate balance between centripetal forces that push manufac Ranjith (2020) in which they found that the diffusion of AI capabilities
turers’ activities toward integration and centrifugal forces that pull in manufacturing affects the dynamic capabilities of organizations in
value proposition out into the market in a customer-driven fashion, terms of the firm’s ability to sense rapid industry changes, seize op
which creates an evolution in the services ecosystem in development portunities in terms of more personalized customer value proposition
(Holgersson et al., 2022). and reconfigure the use of internal process and resources for speed and
AI in the industrial marketing context is widely viewed as a cost-cutting in the transformational process of digital servitization
knowledge-based system, yet it is not clear whether manufacturers’ ef (Tronvoll et al., 2020). Furthermore, literature has conceptualized firm
forts to acquire, assimilate, and exploit new knowledge would play a capabilities as a hierarchy; therefore, for manufacturing firms to achieve
role in the advancement of servitization and AI, warranting a joint dynamic capabilities, they need to efficiently exploit their current,
investigation of these perspectives and their optimal configuration lower-order capabilities, in terms of systems, assets, and competencies
(Valtakoski, 2017). they have acquired while operating (Slotegraaf, 2007). Previous
To address these literature contentions, first, this study develops the research has highlighted the importance of dynamic capabilities in
AI capabilities construct (as opposed to the IT capabilities construct). As advancing servitization processes (Coreynen et al., 2020; Fischer et al.,
discussed in Section 2, we propose the notion of AI capabilities con 2010; Kanninen et al., 2017a), but it has fallen short of addressing the AI
sisting of four components: (1) AI for customer value proposition, (2) AI capabilities required to achieve servitization in terms of the optimal
for key processes optimization, (3) AI for key resources optimization, configuration for endogenous and exogenous resources needed to
and (4) AI for societal good; and we have provided a measurement scale improve the manufacturing business model and business-process
for this construct. Second, building on dynamic capability theory (DCT), redesign.
this study introduces and develops a theoretical framework delineating The intersection between dynamic capabilities literature, B2B mar
the mechanism by which AI capabilities leverage servitization, taking keting and information system (IS) research, indicates that a potentially
into consideration how AI affects specific organizational processes. transformative technology such as AI is imperative for companies
Furthermore, we have yet to understand the institutionalization of seeking new competitive advantage, that leads to unlocking new busi
absorptive capacity and its interactive effect on the relationship between ness models by sensing opportunities in new technologies and seizing
AI capabilities and servitization. Third, we apply both covariance-based the new revenue streams related to those technologies, which requires
structural equation modeling (CBSEM) and a fuzzy-set qualitative reconfiguration and realignment of firms’ resources, structure, and
comparative analysis (fsQCA) to assess the net and combinatorial effects strategy (Teece, 2023). AI as an extension to information technologies
of the proposed four components of AI capabilities on servitization, can help in exploring and exploiting firms’ dynamic capabilities related
which leads to better development of the theoretical parsimony of AI- to process, supplier, and alliances building, operations, and marketing
servitization configurations. Consequently, this combination of both that impact overall performance (Majhi et al., 2021).
techniques can greatly enhance our understanding of the underlying Following the strategic orientation of dynamic capability literature,
complex reality in this configuration, paving the way to important data one can argue that advancing both servitization and AI capabilities re
that inform practitioners and decision-makers in terms of finding quires a process of change, guided by the micro-foundations of DCT. For
2
instance, trained machine learning models can help in sensing new op Table 1
portunities or threats; seizing new opportunities through the integration Definitions of the constructs.
of intelligence capabilities targeted to innovate business models (Lee Construct Definition
et al., 2019); and finally transforming existing business model configu
Servitization The transformational processes whereby a company
rations by resource realignment and optimization to leverage trans (Kowalkowski et al., shifts from a product-centric to a service-centric business
formative social progress (Toma et al., 2020). Prior IS literature also 2017a) model and logic.
expanded the dialogue of the transformative value of AI as a firm- AI Capabilities A firm’s ability to select, orchestrate, and leverage its AI-
specific complementary (Nishant et al., 2020). (Mikalef & Gupta, specific resources, in order to identify, interpret, make
2021) inferences, and learn from data to achieve predetermined
In addition, DCT is built on the premise of organizational learning organizational and societal goals.
abilities that leverage innovation and that deliver more customized Absorptive Capacity A set of organizational routines and processes by which
servitized offers. Screening external environments for new knowledge (Zahra & George, firms acquire, assimilate, transform, and exploit
and continuous assimilation, transformation, and exploitation of such 2002) knowledge to produce a dynamic organizational
capability.
market knowledge enhance a firm’s responsiveness and operational
agility, making absorptive capacity an important type of dynamic
capability that impacts servitization processes (Malhotra et al., 2005; extracting knowledge and insights from data, using statistical tech
Wheeler, 2002). niques to support decision-making (Provost & Fawcett, 2013).
The widely used AI in the public domain is known as artificial narrow
2.2. Conceptual framework and hypotheses development intelligence (ANI) (Rosa et al., 2016). Wirtz et al. (2018) categorized AI
capabilities that are widely used in business applications, ranging from
Fig. 1 presents the research model, while Table 1 shows the opera AI process automation systems; virtual agents; predictive analytics and
tional definitions of the research model constructs. data visualization; cognitive robotics and autonomous systems; intelli
gent digital assistants (IDA); cognitive security analytics and threat in
2.2.1. AI capabilities and servitization telligence; identity analytics; edge analytics; and finally, machine vision
Mikalef and Gupta, (2021) define artificial intelligence as “the ability and sensing. In the same vein, Sjödin et al. (2020a) classified AI capa
of a system to identify, interpret, make inferences, and learn from data to bilities in a digitally servitized manufacturing context into data pipeline
achieve predetermined organizational and societal goals” (p. 3). capabilities, algorithm development capabilities, AI democratization
Artificial intelligence, therefore, constitutes an interdisciplinary field capabilities customer co-creation capabilities, and data-driven delivery
that opens a new horizon of opportunities that have practical implica operation. In addition, IS literature also classified the value types of AI
tions for both businesses and society as a whole. Despite the great applications in terms of process automation, cognitive insights, and
attention this field of research has drawn from both researchers and cognitive engagement (Collins et al., 2021). However, the prior litera
practitioners, its ultimate uses and applications are still to a certain ture on the operationalization and conceptualization of AI capabilities
extent new and over-promised (Bughin et al., 2017). To cut through the still lacks some dimensions related to sustainability and value proposi
hype, AI is built on the advancement of machine learning (ML) capa tion co-creation (Nishant et al., 2020; Vinuesa et al., 2020). In this
bilities using a vast amount of structured and/or unstructured data to research, we take these classifications further. Building on both AI and
mimic a certain level of human cognitive ability (Smith, 2019). Arthur business model literature (Hercheui & Ranjith, 2020; Johnson et al.,
Samuel, a pioneering figure in ML, conceptualized it as a field of study 2008; Teece, 2010), we argue that AI capabilities can be divided into
that gives computers the ability to learn without being explicitly pro four categories: those which advance the customer value proposition
grammed (Samuel, 1959), giving rise to the variety of self-learning al (sensing and seizing capabilities), those which help in optimizing key
gorithms and knowledge acquisition platforms that make up the business processes, those which help in optimizing key resources
backbone of modern expert systems. It is noteworthy that, while the (transforming capabilities), and finally those which enhance societal
terms ‘AI’ and ‘data’ science are widely used interchangeably in the good (Day & Schoemaker, 2016).
industry, they are distinct and widely overlapped. Data science helps in The main promise of AI and data science is to provide managers with
Fig. 1. Research Model.
3
actionable insights that help them enhance their business value, antic detection, and reduction, and economic optimization, enhancing overall
ipate customer preferences, and more accurately price their market of effectiveness and efficiency. The latter, on the other hand, is more about
ferings. Researchers have found that manufacturing firms using deep optimizing internal resources and manufacturer’s workloads by better
learning personalization algorithms achieved better customer success resource orchestration in terms of resource coordination, leverage, and
and more profitable servitized offerings (Dubé & Misra, 2019), implying deployment; it also includes the optimization of the inter-firm resources,
a positive relationship between AI capabilities targeting customer value which involves suppliers’ networks, distributors, and external data
propositions and digital servitization (Davenport et al., 2020; Rachinger warehouses, which enhance parties’ collaboration, inventory predic
et al., 2019; Ritter & Lund, 2020). Furthermore, prior research has tion, labor productivity, and data governance to achieve higher effi
hinted at the positive impact of algorithmic pricing on service success ciency. Previous servitization literature shed some light on the positive
(Assad et al., 2021). The introduction of AI that predicts future trends impact of specialized AI platforms for resource and process optimization
has had a positive impact on delivering servitized offers such as pre on servitization (Barbieri et al., 2021). Therefore, AI capabilities tar
dictive maintenance and asset management (Wang & Wang, 2017). In geting resource optimization are fundamental for smart servitization
addition, AI is a strong enabler of business model innovation and digital success (Coreynen et al., 2017; Huikkola et al., 2016; Sjödin et al.,
servitization (Sjödin et al., 2021) allowing manufacturing firms to create 2021). Parallel to this, Klumpp (2018) found a positive impact from AI
value from servitization, building on the advancement of digital service targeting logistic and service provision transition. Empirical research
innovation. has also concluded that AI and robotics are fundamental to fostering
While the servitization process is viewed as an outcome of digital process automation, which positively impacts servitization (Blöcher &
service innovation (Raddats et al., 2022), AI capabilities such as using Alt, 2020). Suppatvech et al. (2019) stressed the importance of using AI
and analyzing big data can hugely impact the customization part of applications and IOT to increase resource utilization to support serviti
service design and delivery, especially in value-based pricing and pre zation activities, added to which Zhang et al. (2020) found that using
dictive maintenance services. Furthermore, in looking for growth, Artificial Intelligence/Machine Learning (AI/ML) dedicated to resource
manufacturers opt to apply AI to optimize, personalize and transform optimization, automation and orchestration can positively impact Pro
every customer touchpoint, and use AI to enhance value proposition duct–Service System (PSS), by enhancing resource elasticity and
manifested in advancing service call scheduling and new service rec configuration, helping manufacturers to streamline workflow and
ommendations that require capturing, analyzing, and utilizing customer pipelines to deliver service solutions in minimum time, cost and
data to sense, shape and optimize customers’ experience, leading to maximum reliability.
better marketing capabilities (Ameen et al., 2021; Mora Cortez & While the literature has widely ignored the social impact of serviti
Johnston, 2018). In the B2B context, AI has also been found to allow zation (Doni et al., 2019), those AI capabilities that tackle sustainability
manufacturers to understand the specific granular requirement deemed issues and decarbonization solutions are highly important from a cus
necessary to deliver more servitized offers in a more customer-driven tomer’s perspective (Bag, Gupta et al., 2021, Bag, Pretorius, et al., 2021;
use cases fashion (Kushwaha et al., 2021). Therefore, AI related to Chandy et al., 2021). Customers’ pull and regulatory bodies’ push exert
enhancing customer value propositions has been found to have a posi high pressure on companies to strategically think about providing ser
tive impact on servitization processes in terms of designing relevant, vitized offers that tackle climate change challenges. Current literature
outcome-based contracts, leading to better customer service, and on industrial marketing has found a positive impact of AI on climate-
customer engagement (Kumar et al., 2021). Syam and Sharma (2018) driven service analytics capabilities, as well as service innovation and
found a positive impact of AI capabilities and machine learning on performance (Akter et al., 2021). As a driver of servitization, AI plays a
servitization success, stemming from the breakthroughs achieved by significant role in advancing social innovation capabilities, which help
gathering real-time data from the sales processes and collecting to reduce carbon emissions through smart energy management services,
actionable after-sales insights. safety assistant agents using deep reinforcement learning, and effective
AI’s applications and capabilities related to process optimization are waste management services that apply complex data science analytics
profoundly useful for businesses looking to increase efficiency, improve (Calabrese et al., 2018; Vázquez-Canteli & Nagy, 2019). Previous liter
uptime using predictive asset maintenance, cut costs, improve process ature has also highlighted the importance of AI capabilities to provide
reliability, execute business model renewal, and enhance yield optimi manufacturers with more sustainable smart supply chains (Sanders
zation, which leads to better-servitized market offerings (Baines et al., et al., 2019); the data-driven approach to supply chain management and
2009; Baines & Lightfoot, 2014). IS literature argues that AI for process manufacturing could not be materialized without the advancement in
efficiency can help augment human intelligence using appropriate ML ML and core predictive models that manage risk and increase sustain
models, especially in repetitive routine tasks, leading to more integra able market offerings (Tseng et al., 2021). The impact of AI applications
tion between humans and AI that helps in overcoming some cognitive designed to advance sustainable causes has also been an enabler of
limitations inherited in humans (Enholm et al., 2021). AI applications service innovation and solutions (van Wynsberghe, 2021). AI also can
for process control and optimization can hugely enhance manufacturing help to manufacture enhance resource efficiency, which positively im
scheduling, multi-period planning, and real-time optimization, which pacts the ecological performance of manufacturers (Waltersmann et al.,
can lead to better service delivery in terms of quality and flexibility. 2021).
Furthermore, the use of data rectification by extracting meaningful Taken together, the previous arguments indicate that overarching AI
features from collected data can help AI applications in predictive capabilities most likely have a positive impact on servitization. There
modeling, fault detection, process optimization, and control (Thon et al., fore, we stipulate the following hypotheses:
2021). The use of AI applications in process optimization and control has
a positive impact on digital servitization and value co-creation H1. A firm’s AI capabilities are positively related to servitization.
(Boehmer et al., 2020; Paiola & Gebauer, 2020). H1a. A firm’s AI capabilities for customer value propositions are
Consequently, AI capabilities related to resource optimization can positively related to servitization.
facilitate internal and external resource allocation and supplier sourc H1b. A firm’s AI capabilities for key process optimization are posi
ing, leading to a reduction in service waiting time and other adminis tively related to servitization.
trative issues. It is noteworthy to draw some distinctions between AI H1c. A firm’s AI capabilities for key resource optimization are
applications for process optimization and AI applications for resource positively related to servitization.
optimization in the manufacturing context, in which the former is more H1d. A firm’s AI capabilities for societal good are positively related
related to generating action recommendations in real time to avoid any to servitization.
benchmark deviation, ensuring better performance monitoring, error
4
2.2.2. Moderating role of absorptive capacity Table 2

The concept of absorptive capacity, which emerged from the field of Sample characteristics.
macroeconomics (Adler, 1965), emphasizes the ability of a company to %
commercially exploit, absorb, and assimilate external knowledge and (N ¼ 185)
resources. The notion of absorptive capacity encapsulates the impor Industry (US 2 digits SIC code)
tance of such capabilities to ensure a manufacturer’s growth; this can be (13) Oil and gas extraction 10
achieved by good execution of a unique set of intra-organizational (29) Petroleum refining and related industries 14
routines and incremental adjustment and reconfiguration of the firm (35) Industrial and commercial machinery and computer equipment 24
(36) Electronic and other electrical equipment and components 22
dynamic capabilities (Zahra & George, 2002). It could be argued that (37) Transportation equipment 6
organizational learning capabilities are central to providing solutions to (38) Measuring, analyzing, and controlling instruments 19
customers’ problems (Davies et al., 2007). Therefore, external knowl (49) Electric, gas, and sanitary services 5
edge management is paramount from a capability perspective to ensure Company Size
1–250 employee 12
a smooth and effective service transition which requires micro-vertical
250 + employee 88
integration within different partners in a servitized context (Todorova Respondent position
& Durisin, 2007). Current literature has classified absorptive capacity as Service Manager 27
a dynamic capability (Malhotra et al., 2005; Pavlou & El Sawy, 2006), Head of IT 25
and companies with higher degrees of absorptive capacity can utilize Chief Information/ Technology Officer 21
Chief Operations Officer 11
knowledge-based assets more efficiently, sensing technological changes Other (Lead data scientist, IT Project Manager, Marketing Manager, 16
and work to reconfigure functional capabilities to create better market etc.)
offerings. IT and/or service experience
Consequently, the success of service innovation depends predomi Less than one year 14
1–2 years 21
nantly upon the firm’s level of absorptive capacities. Customers’ data
3–4 years 27
can be internalized and funneled into machine learning models to create More than 4 years 38
more comprehensive consumer profiles, better insights, better-
customized propositions, and better customer experiences (Wilson &
Daugherty, 2018). On the one hand, this creates opportunities for firms undergoing digital transformations and also offer sufficient service so
to shift to service provision by acquiring the cognitive technological lutions (Fang et al., 2008). Secondly, this method enhances the gener
capabilities required to deliver servitized offers such as AI capabilities alizability of the research findings. The self-administrated online survey
(Mikalef et al., 2021). The capacity to absorb external knowledge is was developed and fielded over a period of nine weeks and was sent to
considered an important organizational antecedent in which ambidex 584 prospect companies identified as belonging to a potential target
terity literature unearthed a positive relationship between absorptive audience, and respective key informants were identified based on their
capacity, service transition, and business model adaptation (Kranz et al., possession of sufficient knowledge about the research context and their
2016). Manufacturers with high levels of absorptive capacity can easily appropriate involvement in steering decision-making with a good
sense new opportunities and create value by enhancing servitized offers. overview of the entire firm (Kumar et al., 1993). Professional social
Heterogeneity in absorptive capacity levels -low versus high- plays a media platform groups such as Artificial Intelligence and Business An
vital role in operational efficiency and the embeddedness of new tech alytics (AIBA) on LinkedIn were used to identify prospective re
nologies deemed paramount to the servitization process (Escribano spondents in targeted companies. Further filtering criteria and
et al., 2009). Rothaermel and Alexandre (2009) found that absorptive validation items were included in the survey instrument to further
capacity exerts a positive moderating effect on a company’s technology isolate both the correct key informants and servitized firms. Three weeks
capabilities and service provision performance. In addition, Zhang et al. after the initial online invitation, Dillman’s (2011) follow-up protocol
(2018) found a positive effect of IT capabilities on product and service was used to enhance the response rate by targeting those subjects who
innovation, which is mediated by absorptive capacity. Empirical failed to return a usable questionnaire. A total of 185 fully usable
research has also found that absorptive capacity moderates the rela questionnaires were received and included in the final data set, with a
tionship between process innovation and service innovation (Ahlin response rate of 31.6 %, in line with other response rates associated with
et al., 2014), while also moderating the relationship between strategic organizational research (Tippins & Sohi, 2003).
collaboration —facilitated by AI advancement— and digital trans Following Armstrong and Overton’s (1977) recommendations, a test
formation (Siachou et al., 2021). Therefore, given the discussion above for a potential non-response bias was performed by splitting our sample
and the empirical evidence from prior literature, we postulate the answers for all study constructs —servitization, absorptive capacity, and
following hypothesis: the four dimensions of AI capabilities— into two groups: early and late
respondents. The t. test resulted in no significant statistical difference
H2. A firm’s absorptive capacity moderates the relationship between between respondents and non-respondents. Furthermore, a one-way
AI capabilities and servitization. Specifically, the relationship di ANOVA was performed for all constructs between key informants with
minishes under conditions of low absorptive capacity and becomes IT knowledge and those with other functional knowledge. The results
stronger as absorptive capacity increases. obtained indicate no statistically significant differences between the two
groups, with p < 0.01.
3. Method Furthermore, to assess the degree to which common method vari
ance (CMV) might influence the study results, the researchers employed
3.1. Data collection and sample decomposition ex-ante remedies by Podsakoff et al. (2003), in terms of survey design,
anonymous participation, and the use of reverse items, etc. The study
Consistent with previous research (e.g., Wales et al., 2013), this also used Harman’s single-factor test as an ex-post procedure. We sub
study collects both secondary data by leveraging the OSIRIS database jected the study variables to a principal component factor analysis and
and perceptual data —respondent survey— from a sample consisting of varimax rotation with eigenvalues > 1.0. The test showed that six factors
manufacturing firms in the United States (58 %) and the European Union accounted for 74.38 % of the variance, and the first factor accounted
(42 %). Table 2 shows the demographic characteristics of our research only for 21.39 % of the variance, resulting in no dominant single factor
sample. We purposely chose a wide range of firms in different industries emerging due to CMV. We also tested the variance inflation factor (VIF)
for two reasons. First, doing so accounts for those firms which are for all items, with the highest VIF of 2.19 being well below the
5
recommended threshold value of 4 (O’Brien, 2007). Therefore, it is safe proportional reduction in loss (PRL) for inter-judge agreements was
to conclude that non-response bias, CMV, and multicollinearity are calculated (Rust & Cooil, 1994b). The results show that the proportion of
minimal and not serious concerns in this study. inter-judge agreement for the panel was 0.86 between 11 judges (in
dustry experts), creating a PRL coefficient of 1, thereby ensuring AICAP
construct reliability (Rust & Cooil, 1994, p. 8.
3.2. Measures Taking into consideration the C-OAR-SE method recommendations
when deploying the Likert scale format (see Rossiter, 2002, p. 322). This
Appendix A shows the study’s measurement scales. Some of the newly developed scale with its four main sub-dimensions was oper
scales were adopted from previous literature, while AI capabilities are a ationalized using a 5-point Likert ranging from “1—strongly disagree” to
novel construct with newly developed items. “5—strongly agree.” Finally, and for more scale development rigor and
parsimony, the use of Rossiter’s method for scale development -which
3.2.1. Independent variable relies on content validity- was coupled with the two steps approach
recommended by Gerbing and Anderson, (1988) which entails per
3.2.1.1. The C-OAR-SE method for AICAP scale development. The novel forming Exploratory Factor Analysis (EFA) and Confirmatory Factor
construct of AI capabilities (AICAP), was developed by employing the C- Analysis (CFA). The EFA has been carried out to address the multidi
OAR-SE method (Rossiter, 2002; Rossiter, 2011) which stands for mensionality of the AI capabilities construct, using principal axis
construct definition, object classification, attribute classification, rater factoring and ProMax oblique rotation with eigenvalues > 1.0. The
identification, scale formation, and enumeration. According to this analysis yielded four main dimensions supporting the construct’s theo
method, a construct can be conceptualized according to an object (a retical conceptualization, and all items -except three- were loaded
focal object being rated), attribute (a dimension of judgment), and rater cleanly on their corresponding factors with loadings exceeding the
entity (the judges). In the first step, the study operationalized the AI cutoff value for the newly developed scale of 0.6 without any major
capabilities construct while considering both the study’s research design cross-loading to report (Kline, 2014). Proceeding to the CFA stage, the
and its specific empirical context. In the construct definition stage, we single factor CFA indicates that three items were deemed to be not
have built on the working definition of AI capability (see Table 1 based satisfactory due to high modification indices and were removed from the
on Mikalef and Gupta, 2021), we also broadened the definition to final items set (Bagozzi & Yi, 1988), leading to the unidimensional
encapsulate different works in business model innovation, information model to show acceptable goodness-of-fit indices, see Appendix B.
management, and computer science that are closely related to the Furthermore, all sub-dimensions scales yielded satisfactory levels of
constitution of the AI capabilities conceptualization process (Johnson Cronbach’s alpha (α > 0.70) and a satisfactory level of average variance
et al., 2008; LeCun et al., 2015; Wirtz et al., 2018; Yao et al., 2017; extracted (AVE), which was greater than the threshold level of 0.50,
Zadeh, 1996). The object of the construct in this study is AI in providing more evidence of the construct’s reliability and convergent
manufacturing firms, an abstract formed object according to the C-OAR- validity (Bagozzi & Yi, 1988). Additional validity tests were also per
SE method, while AI capabilities are a component of the AI object. formed using second-order factor analysis for the AICAP construct, and
Second, the attribute of AI capabilities is a second-order eliciting attri the study tested whether the AICAP sub-constructs could be aggregated
bute encapsulating four components which are also eliciting attributes into one main construct. Appendix B shows the confirmatory factor
according to the C-OAR-SE framework. Third, the rater entities were analysis (CFA) results, which provide good support for the sub-
selected and arranged to include a panel of three academics (in indus constructs aggregation into one second-order construct.
trial marketing and AI) and 11 industry experts familiar with the concept
of service innovation and AI and its practical applications in the 3.2.2. Dependent variables
manufacturing context. The first and the second steps were accom The servitization construct (SERV) was measured using the scale
plished by an open-ended interview, during which, in step three, the proposed by Abou-foul et al. (2021). The study conceptualizes and
identified rater entities (expert judges) ratified the main components of operationalizes servitization as a higher-order construct that can be
AICAP (Rossiter, 2002). Accordingly, AI capabilities are conceptualized decomposed into three dimensions: top management service orienta
as a reflective-reflective second-order construct following Jarvis et al.’s tion, resource mobilization, and market offering. Each item was
(2003) suggestion for construct types specifications. The components of measured using a 5-point Likert scale ranging from “1—strongly
the AICAP construct are as follows: (1) AI for customer value proposition disagree” to “5—strongly agree.”.
(AICVP), (2) AI for key process optimization (AIKPO), (3) AI for key
resource optimization (AIKRO), and (4) AI for societal good (AISGD). 3.2.3. Moderating variables
Building on the aforementioned three steps, a total of 28 potential The absorptive capacity (ACAP) scale items were adopted from
stem items (seven for each component) were formulated guided by a Flatten et al. (2011) and operationalized using a 5-point Likert scale
literature review and qualitative study using open-ended interviews ranging from “1—strongly disagree” to “5—strongly agree.” This study
with the selected rater entities, to formulate the items part of each uses a shorter version of the original measurement scale on the aggre
component of AICAP and also to address any issues related to items’ gated level, following the recommendations of Sandy et al. (2016). The
clarity and relevance to component definition (El & Akrout, 2020; Sabri shorter version preserves the construct’s theoretical domain and
& Obermiller, 2012). multidimensionality, and the scale used also encapsulates the four ACAP
The same panel has also been used in the fourth step of scale for dimensions of knowledge acquisition, assimilation, transformation, and
mation, in which the stems of the scale items were generated for pre- exploitation proposed by Zahra and George (2002).
testing using the cognitive interviewing method with the raters panel
(D. Collins, 2003). During this, experts’ rating was used to establish 3.2.4. Controls
AICAP content validity. As recommended by the C-OAR-SE method, The research model was subjected to control variables to account for
content validity index (average I-CVI) was calculated for all AICAP di any extraneous sources of variation in the firm’s servitization processes.
mensions, and the result for all items exceeded the cutoff value of 0.79 Firm age (FRMAG), firm size (FRMSZ), industry (IND), and slack re
(Polit & Beck, 2006). However, eight items were rejected and deleted sources (SLRES) were controlled for following Wales et al. (2013) and
from the final measurement scale for receiving an I-CVI value below Abou-foul et al. (2021). The aforementioned controls were measured
0.70, ensuring AICAP content validity (Polit & Beck, 2006). using objective data obtained from the OSIRIS database. The mean
Finally, the AICAP was enumerated by deriving a total score from the values of firm size and age over three years —from 2017 to 2019— were
scale items (Rossiter, 2002), in establishing construct reliability, the reported and transformed using the natural logarithm of a firm’s number
6
of employees and its age since incorporation. The 2-digit 2019 US 4. Analysis and results
standard industry classification (SIC) was used to control for industry
membership effects; this variable was operationalized using dummy Building on both CBSEM and fsQCA (e.g., Jahanmir et al., 2019;
variables, with industrial and commercial machinery and computer Torres et al., 2018), this study seeks to enhance this study’s pragmatic
equipment representing a reference group. Finally, slack resources were implications and results. The implementation of variable-oriented
controlled and calculated using the current ratio. Research has found techniques and case-oriented techniques in one analysis increases the
that higher levels of slack resources increase efficiency and the utiliza granularity of the research results and the synergy in converging both
tion of key resources and capabilities (Thornhill & Amit, 2003). methods to leverage causal complexity theory (Fiss, 2011; Misangyi
et al., 2016).
3.3. Psychometric properties of the measurement model
4.1. Structural model results
The two-step approach (Anderson and Gerbing, 1988) was followed
to assess the psychometric properties of this study’s measurement Table 4 reports the results of the structural models following con
model. CFA was employed using AMOS 26 to assess the nomological ventional CBSEM; two models have been tested. The first is the model of
validity of constructs, as well as the scale’s reliability and validity. The the direct effects of the four sub-dimensions of AICAP on servitization.
study’s measurement model achieved a good model fit and supported This model’s goodness-of-fit statistics indicate that the study’s structural
the unidimensional nature of the measurement items, with χ2 = 173 on model fits the data, with χ2 = 24.801 (df = 22; p < 0.01), χ2/df = 1.127,
89 df, RMSEA = 0.072, CFI = 0.97, IFI = 0.97, NFI = 0.94, and TLI = NFI = 0.962, CFI = 0.995, TLI = 0.993, RMSEA = 0.027, SRMR = 0.059.
0.95, exceeding the recommended values in the literature (Kline, 2015). The path coefficient from AI for customer value proposition (AICVP) to
Table 3 reports a summary of descriptive statistics and constructs in servitization was significant, with β = 0.32; p < 0.001. The path coef
tercorrelations which indicate good support for the direction of the ficient from AI for key processes optimization (AIKPO) to servitization
study’s hypothesized relationships. Cronbach’s alpha was used to assess was significant, with β = 0.27; p < 0.001, while the path coefficients
scale internal consistency reliability, and all scales achieved a satisfac from AI for key resources optimization (AIKRO) and AI for societal good
tory level of at least 0.70. All values of composite reliability (CR) were (AISGD) to servitization were significant, with β = 0.18; p < 0.001 and β
well above the benchmark level of 0.70, and the average variance = 0.30; p < 0.001, respectively. Consequently, H1a, H1b, H1c, and H1d
extracted (AVE) values were all satisfactory, exceeding the threshold are supported (see Fig. 2), with a coefficient of determination R2 = 0.84
level of 0.50. These results are deemed satisfactory to establish which explains 84 % of the variance in the servitization construct.
convergent validity (Bagozzi & Yi, 1988). The second model (see Fig. 3) was developed to test the moderation
Discriminant validity (see Appendix C) was assessed for the study’s effect of absorptive capacity (ACAP). Following the recommendations of
core constructs using the Heterotrait-Monotrait Ratio (HMTM), and all Hayes et al. (2017) in testing interaction effects, the model achieved
average correlations were below the cutoff value of 0.85, indicating sufficient goodness-of-fit statistics, with χ2 = 20.310 (df = 18; p < 0.01),
sufficient evidence of discriminant validity (Henseler et al., 2015). We χ2/df = 1.112, NFI = 0.970, CFI = 0.996, TLI = 0.994, RMSEA = 0.026,
also tested the difference between χ2 in the constrained CFA model and SRMR = 0.062. The path coefficient from AICAP to servitization was
the unconstrained model; the differences were found to be significant, significant, with β = 0.83; p < 0.001. The moderation effect of ACAP) on
providing more evidence of discriminant validity (Anderson and Gerb AICAP and servitization was significant, with path coefficient β = 0.08;
ing, 1988). Furthermore, the study’s scale items loaded significantly on p < 0.01. Thus, the results give compelling support to H1 and H2.
their prospective construct. All standardized factor loadings are above Meanwhile, the coefficient of determination R2 = 0.86 explains 86 % of
0.70 (except for two items at 0.65–0.64), and are all significant at p < the variance in the servitization construct. Finally, none of the control
0.001 without any major cross-loading, establishing convergent validity variables had any significant effect on servitization (p > 0.05), and no
(Anderson and Gerbing, 1988; Hair, 2010). Due to some inter-construct significant drop in R2 (<1%) was reported when controls were not
correlations being higher than the benchmark value of 0.60, a test of included in the tested models. Fig. 4 graphically depicts the moderated
multicollinearity was performed which resulted in variance inflation relationship.
factors (VIFs) ranging from 1.6 to 2.7, well below the recommended
threshold of 5 (Hair, 2010). From this, we conclude that multi 4.2. Data and calibration
collinearity won’t be a concern in our dataset.
The data calibration phase is the first procedure in fsQCA and
Table 3
Summary of descriptive statistics, latent variables inter-correlations for first-order constructs, and psychometric characteristics.
Mean SD CR AVE 1 2 3 4 5 6 7 8 9 10
1 AICVP 3.6 1 0.88 0.65 (0.82)

2 AIKPO 4 1 0.91 0.63 0.52** (0.88)
3 AIKRO 4.2 0.98 0.85 0.59 0.56** 0.68** (0.78)
4 AISGD 3.9 1 0.88 0.72 0.52** 0.59** 0.70** (0.80)
5 ACAP 3.9 1 0.90 0.72 0.65** 0.76** 0.75** 0.71** (0.90)
6 SERV 3.8 1 0.92 0.55 0.72** 0.74** 0.75** 0.75** 0.63** (0.91)
7 FRMSZ (LN)a 10.7 1 – – 0.06 − 0.02 − 0.13 − 0.04 − 0.00 − 0.06 (n/a)
8 SLRES 1.5 0.7 – – − 0.15* − 0.10 − 0.06 − 0.16* − 0.15* − 0.13 − 0.13 (n/a)
9 FRMAG (LN) a 2.9 3.4 – – − 0.05 − 0.03 − 0.04 − 0.13 − 0.03 − 0.08 − 0.02 0.11 (n/a)
10 IND 34 10 – – − 0.02 − 0.03 − 0.06 − 0.09 − 0.09 − 0.09 − 0.09 − 0.05 0.01 (n/a)
Notes: N = 185.
SD—standard deviation.
Cronbach’s alpha (α) presented along diagonals, CR—composite reliability, AVE—average variance extracted.
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
(n/a) Not applicable.
a
logarithmically transformed to reduce skewness.
7
Table 4
Test results for the structural models.
Parameter a Direct effects Interactive Result
model b model c
Hypothesized paths
AICAP → SERV (H1) – 0.83***(17.24)
AICAP × ACAP → SERV (H2) – 0.08**(3.30)
Control variables
IND → SERV − 0.05ns (1.38) − 0.04ns (1.43)
FIRMSZ → SERV 0.01ns (0.06) − 0.04ns (1.68)
FIRMAG → SERV 0.01 ns (0.19) − 0.03ns (0.95)
SLRES → SERV − 0.04ns (1.24) − 0.01ns (0.33)
Measurement model and first-
order factors d
AICVP → SERV(H1a) 0.32*** (7.5) –
AIKPO → SERV(H1b) 0.27***(6.4) –
AIRKRO → SERV(H1c) 0.18***(3.8) –
AISGD → SERV (H1d) 0.30***(5.6) –
Goodness-of-fit statistics
χ2 =24.801*** =20.310 ***
df =22 = 18
NFI =0.962 =0.970
CFI =0.995 =0.996
TLI =0.993 =0.994
RMSEA =0.027 =0.026
SRMR =0.059 =0.062
R2 = 0.840 =0.866
N 185 185
Fig. 3. The result of the intreractive effect model.
Indicates the full degree of confirmation.
Chi-square (χ2); Degree of Freedom (df); Normed Fit Index (NFI); Goodness of
Fit Index (GFI); Comparative Fit Index (CFI); Tucker-Lewis index (TLI); Root
Mean Square Error Approximation (RMSEA); Standardized Root Mean Squared
Residual (SRMR); R2 coefficient of determinant.
*p < 0.05; **p < 0.01; ***p < 0.001.
a
Estimates are standardized with t-values shown in parentheses.
b
Includes only the direct effect of the decomposition of AI capabilities on
servitization.
c
Includes only the moderating effect of absorptive capacity on the second-
order AI capabilities construct and servitization.
d
All the item loadings for the first-order factors were significant at p < 0.001.
ns
p-value is not significant.
Fig. 4. Moderating effect of absorptive capacity.
In this study, the four causal conditions and the outcome are con
structs measured with multiple items. To obtain fuzzy-set membership
scores from the constructs, we needed to compute one value per
construct to be used as input in fsQCA; to get this value we computed the
mean of all the items that make up each construct.
Once we had transformed the constructs into single variables, we
calibrated all the variables, in the same way, using the direct method of
calibration (Ragin, 2009). In this process, the researcher needs to choose
three breakpoints, or anchors, which define the level of membership in
the fuzzy set for each case. Following Ragin’s (2009) recommendations,
we used fuzzy values of 0.95 for full membership, 0.50 for the crossover
Fig. 2. The Result of direct effects model. point, and 0.05 for full non-membership. To decide which values in our
data set corresponded to the three anchor points, we used percentiles
requires transforming the conventional variables into fuzzy-set scores, i. following the recommendations of Pappas and Woodside (2021). Our
e., the input data for fsQCA analysis. Fuzzy-set scores reflect the degree data did not follow a normal distribution but instead were skewed. Then
of membership in the target set and range from 0 to 1. Thus, the four we computed the 80th, 50th, and 20th percentiles of our measures and
causal conditions —AI for customer value proposition, AI for key pro used these values as the thresholds for full membership, the crossover
cesses optimization, AI for key resources optimization, and AI for soci point, and non-membership, respectively, in fsQCA software (Pappas &
etal good — and the outcome servitization need to be calibrated. Woodside, 2021).
8
In fsQCA, the cases that are exactly on 0.50 are dropped from the Table 6
analysis because it represents the point of maximum ambiguity (Ragin, Sufficient configurations of conditions for servitization.
2009). To overcome this, Fiss (2011) suggested adding a constant of The configurations leading Raw Unique Consistency
0.001 to the causal conditions below full membership scores of 1. SERV coverage coverage
Once all variables were calibrated, we used fsQCA 2.0 free software AICVP * AISGD 0.670 0.102 0.873
to identify which causal conditions were necessary and sufficient for the AICVP * AIKRO 0.639 0.070 0.881
outcome (SERV). AIKPO *AIKRO * AISGD 0.591 0.075 0.899
Solution coverage: 0.815
Solution consistency: 0.834
4.3. Analysis of necessary conditions
Note: AICVP = AI for customer value proposition; AIKPO = AI for key processes
The analysis of necessary conditions examines whether any of the optimization; AIKRO = AI for key resources optimization; AISGD = AI for so
cietal good.
four causal conditions can be regarded as necessary for the outcome.
Empirically, testing conditions for their necessity involves checking
consistency and coverage thresholds. A condition can be considered configuration, Pathway 1, is the most empirically relevant (raw
necessary when its consistency and coverage values are above the 0.90 coverage = 0.670, unique coverage = 0.102, consistency = 0.873) and
and 0.50 thresholds, respectively (Ragin, 2009). shows how a combination of AICVP and AISG is sufficient to lead to high
Table 5 depicts the results of the analysis of necessary conditions and levels of servitization. Pathway 2 (raw coverage = 0. 639, unique
indicates that the consistency of the conditions was below 0.90 in all coverage = 0.070, consistency = 0.881) combines the presence of AICVP
cases. Therefore, we find that none of the conditions in this study can be and AIKRO to lead to high levels of servitization.
considered necessary for servitization. Finally, pathway 3 shows a combination of three causal configura
tions, AIKPO, AIKRO, and AISG (raw coverage = 0. 591, unique
coverage = 0.075, consistency = 0.899), to achieve a high level of ser
4.4. Analysis of sufficient conditions
vitization. The consistencies of every single pathway are above 0.75,
indicating consistently sufficient routes to achieve servitization in all
According to Schneider and Wagemann (2010), sufficient conditions
cases.
analysis includes three steps. The first is creating a truth table that in
cludes the membership scores for all the possible configurations of
causal conditions (Ragin, 2009). In our study, the truth table contains 16 5. Discussion
rows = 2 k, where k corresponds to the number of causal conditions.
Second, the truth table must be simplified based on frequency and The possibility that AI capabilities, service innovation, and knowl
consistency thresholds. To reduce the truth table to relevant configu edge leveraging can lead to digital transformation success and social
rations, we set the cutoff points for the frequency at three observations progress as well as solidify the basis for competitive advantage has
as the minimum number of cases that need to be considered for a so emerged as paramount in recent years. Studies that examine the rela
lution, capturing more than 80 % of cases. The minimum consistency tionship between AI capabilities and servitization are still scarce and
threshold was set at 0.80 (Fiss, 2011; Ragin, 2009). Finally, the last step inconclusive, and businesses still struggle to capture the value of digital
is to obtain the solutions. Here, ‘solution’ refers to a set of conditions or transformation due to bad measures and untargeted technology
configurations of conditions —which could be referred to as pathways— investment.
that constantly lead to high levels of servitization. These would consti Prior research has mainly emphasized the implementation of new
tute sufficient conditions. The fsQCA software was used to produce technologies and skills as a proxy for AI capability, without adherence to
complex, parsimonious, and intermediate solutions. Following Ragin’s the micro-foundation of the business model in place (Mikalef & Gupta,
(2009) recommendations, we report that the last one is the most inter 2021). Therefore, this study underscores the importance of looking
pretable. Table 6 provides the intermediate solution for the analysis of beyond introducing specific technology and embedding AI capabilities
sufficient conditions. in all interrelated components of a business model. Customer value
propositions, resource and process optimization, and societal good were
proposed as sub-dimensions of a firm’s AI capabilities profile. This
4.5. Formula of pathways
study’s CBSEM found a significant positive effect of those sub-
dimensions of AI capabilities on servitization, especially the effect of
The three solutions we obtained were as follows (AICVP and AISGD)
AI capabilities that enhance customer value propositions and societal
or (AICVP and AIKRO) or (AIKPO and AIKRO and AIKGD) => SERV.
good. These two AI capabilities help manufacturers advance their ser
Observing the results (see Table 6), the solution’s overall consistency
vitization processes and give confidence to management to proceed in
for high servitization is 0.815 (>0.740), and the overall solution
the development of new servitized market offers. Our findings are also in
coverage is 0.834 (>0.450). Both indicators are above Ragin’s (2009)
line with other research findings (Akter et al., 2021; Burström et al.,
recommended thresholds. We can see that there are three pathways to
2021). In line with prior literature AI applications and their related ML
reach servitization, confirming equifinality (Fiss, 2011). The first
models are important to achieve resource efficiency and better sus
tainability performance, leading to personalized service provision and
Table 5
operational agility (Paiola & Gebauer, 2020; Sjödin et al., 2021). The
Analysis of necessary conditions.
findings also highlight the positive impact of AI capabilities that target
Conditions tested Consistency Coverage resource mobilization on servitization, providing more evidence about
AICVP 0.790 0.755 the nature of this relationship, which is in line with previous studies that
~AICVP 0.349 0.332 examined such relationships (e.g., Abou-foul et al., 2021). While our
AIKPO 0.826 0.787
fsQCA did not reveal any necessary condition for servitization, this
~AIKPO 0.358 0.341
AIKRO 0.794 0.749 finding is in line with Davenport & Ronanki, (2018) and Park and
~AIKRO 0.369 0.355 Mithas, (2020), suggesting AI-enabled capabilities are not necessary for
AISGD 0.797 0.745 any configuration for service performance; a good explanation of such
~AISGD 0.362 0.351 findings is that servitized market offering can be advanced building on
Outcome variable: SERV. cheap widely available technologies, matching the manufacturer level of
“~” represents the absence of a condition. servitization maturity in terms of market offerings. for instance. base
9
services vs advanced services (Marcon et al., 2022; Raddats et al., 2022). by developing firm-specific methodological individualism (Felin &
However, it became evident that manufacturing firms require the Hesterly, 2007) in the form of unique cognitive technologies and
presence of three out of four AI capability dimensions to consistently resource orchestration to leverage both knowledge internalization in
attain servitization, meaning that those capabilities are sufficient to data infusion and service innovation (Barrett et al., 2015; Neirotti et al.,
achieve servitization and play an enabling role in some service inno 2021). Our results also emphasize the objectives of AI applications
vation contexts. This finding, combined with the results of SEM, is also within the firm business model to achieve servitization. Consistent with
suggesting a symmetric view of causality that underpins servitization a growing body of research, the current research classification of AI
related to market heterogeneity (Parida et al., 2015). capabilities and its integration of the dynamic capabilities framework
Our fsQCA also highlighted that process and resource optimization highlights the role of AI applications in advancing customers’ value
combined with a contribution to societal good must be configured and proposition, which requires sensing capabilities for opportunities
presented to leverage collaboration between parties, servitization, and assessment and identification. It also complements DCT in which AI
customer service provision, echoing the findings of Ferreira et al. (2020) application for resources mobilization can advance a firm’s seizing ca
and Toma et al. (2020). The net effect of AI on customer value propo pabilities and its complementary assets. Our model also pinpoints the
sitions was significant and positive in the SEM analysis, while this transformational capabilities embedded within AI applications targeting
finding echoes Payne et al. (2021) and Sjödin et al. (2021). However, the sustainability issues emphasizing the strategic continual renewal
study offers a new perspective on AI for customer value propositions as through business model innovation and knowledge absorption (Bag,
our fsQCA failed to include it in all sufficient configurations to achieve Gupta et al., 2021, Bag, Pretorius, et al., 2021; Scuotto et al., 2022;
servitization; despite achieving the highest path coefficient in our SEM Sjödin et al., 2021). Our study also unearths a set of factors that shows
analysis. the mechanism in which a firm’s AI capabilities can be advanced to
This finding could be attributed to a lack of customer data integra advance servitization and we also opened the discussion on the optimal
tion used to customize value propositions, which requires a certain de configuration of AI applications to tackle the complex configurations of
gree of co-located data; this raises sensitive issues with prospective service innovation in manufacturing firms (Payne et al., 2021).
customers related to data governance, privacy, and data sovereignty, Third, this research offers and empirically validates a multidimen
thus leading to trade-offs between privacy and ML models’ performance, sional scale to assess AI capabilities (Mikalef & Gupta, 2021). Consistent
which might impact servitization value propositions (Li et al., 2020; with the dynamic capabilities perspective, this paper advances the AI
Rieke et al., 2020). capabilities conceptualization domain more comprehensively, in which
The study findings also offer some insights into the dynamics of findings highlight both the critical higher-order capabilities achieved by
organizational learning. We found some evidence that absorptive ca executing applied intelligence and the integration of AI capacities in
pacity strengthens the positive relationship between AI capabilities and transforming and innovating manufacturing business models that
servitization, giving great insight into such conditional dynamism. So emphasize servitization (Kamp et al., 2017). Finally, the novel scale was
cietal good is shown to be important for advancing servitization. In our tested across several industrial classifications, helping to improve its
analysis, societal good appeared in two main servitization paths, indi validity and providing a desirable level of generalizability.
cating a substantial effect of this condition. However, the fsQCA at first
glance might not support our SEM analysis about the importance of AI 5.2. Managerial implications
capabilities related to customer value propositions. Here, however, two
important configurations included AICVP as a sufficient condition to This paper provides a fine-grained perspective to managers engaging
reach servitization, balancing our regression analysis results. While in developing AI capabilities and servitization. First, manufacturing
prior research has advocated the assumption that the introduction of firms may consider their knowledge-intensive, higher-order capabilities
digitalization would enhance service provision (Abou-foul et al., 2021; such as AI capabilities and absorptive capacity, to better sense customer
Gebauer et al., 2021; Marcon et al., 2019), it might be true that it can preferences, shape internal processes and resources to integrate cus
foster efficiencies while failing to achieve competitive advantages, tomers and deliver new value propositions, and seize opportunities by
because other competing firms can adopt the same technologies, embedding AI capabilities in business models to better fit turbulent
resulting in a so-called “Red Queen’’ competition (Barnett & Pontikes, environments and technological advancement. Furthermore, AI
2008). Therefore, the results of our study provide additional insight into advancement can also help managers in increasing service agility and
the role of data-driven AI capabilities in developing a firm’s organic AI responsiveness, which require companies to fine-tune internal pro
competencies, ecosystems, and co-specialized resources to leverage the cesses, functional capabilities, structures, and, ultimately, strategy
firm’s dynamic capabilities (Mikalef et al., 2021). (Felin & Powell, 2016). AI capabilities can also help companies disrupt
key revenue streams of incumbents in the ecosystem, especially in
5.1. Theoretical implications advancing direct-to-customer functions that mainly service the com
pany’s customer base (Sjödin et al., 2020a).
This paper offers several important theoretical contributions. First, it Second, the results of this study highlight that decision-makers
integrates studies on AI capabilities, dynamic capabilities, and serviti should not focus only on internal optimization of the manufacturing
zation. It develops and empirically tests a theoretical model to address process, but should instead look outward and increase their machine
those relationships in a manufacturing context. The study results help in learning capabilities, big data analytics, deep learning, and robotics to
addressing the controversial interdependencies and the value of AI in deliver services that benefit social causes. Managers must also treat data
vestment for servitization (Garcia Martin et al., 2019; Leone et al., 2021; as a strategic resource powering prescriptive key performance indicator
Libai et al., 2020; Waltersmann et al., 2021). The results demonstrate (KPIs) that define, develop, and refine servitization (Payne et al., 2021).
that higher-order capabilities such as those related to AI have a direct On a similar note, managers should increase their investment in AI
impact on firm servitization, and the strength of this relationship is research and commercialization, especially in customer-facing AI, to
contingent on the company’s absorptive capacity. Furthermore, drawing enhance predictive customer intelligence that directly enhances cus
on the emergent literature on complexity and set theory (Anderson, tomers’ lifetime value. This requires coordination efforts between
1999; Ragin, 2009; Woodside, 2014), this study reveals the mechanism parties to create better integration and better machine learning models,
and the differences in configurations in which servitization can be meaning that the success of AI implementation is dependent on fine-
achieved. Second, this study makes a direct contribution to DCT by tuned ML models that feed on data. Managers also should understand
extending our understanding of the micro-foundation of dynamic ca that acquiring data requires a trusting partnership between players in
pabilities and how they can be translated into a manufacturing context the ecosystem. Parallel to this, manufacturers should also focus on
10
services with clear positive societal impacts which accelerate and parsimonious explanation of servitization. Second, from a methodo
cultivate decarbonization initiatives, and the mandatory consideration logical point of view, we recommend testing the proposed model in
of data governance issues. Third, managers should fully understand the different settings using different datasets to provide more empirical
progressive nature of absorptive capacity and its cross-functional evidence of our model’s predictive power using longitudinal data.
properties. Therefore, they have to integrate and assimilate knowledge Further, we are still in the first stages of understanding AI developments
acquired internally and externally in the development process of organic and their impact on business contexts (Mikalef & Gupta, 2021). There
AI applications. In this context, AI capabilities can be utilized and fore, our AI capability construct is not a one size fits all panacea to
further developed to enhance raw data exploration, ML model archi measure manufacturing firms’ AI initiatives, and its scale is by no means
tecture and prototyping, application platforms, data security, and ana exhaustive. We highly recommend embedding AICAP and its four di
lytics (Lehrer et al., 2018). mensions into a nomological network that has causal antecedents to
Furthermore, building on the study, theoretical model managers better test the concept and reduce any artificial entities that are gener
should focus their efforts on enhancing AI applications that highly drive ated by statistics. Researchers should treat this scale with caution when
servitization; this can be achieved by advancing customer value prop applying it in different settings and address any other dimensions
ositions and data integration techniques, for better market offerings that deemed necessary to any contingent aspects. Third, it has been a
are built on collecting valuable and granular customer insight. Of equal temptation from our side to eliminate some causal biases attributed to
importance, managers should focus their AI investment on enhancing model specification and sample heterogeneity which might confound
internal processes that enhance effectiveness and efficiency. AI for key the analysis; therefore, we opt to support our work by using set theory
resources optimization can help in managing supply-chain trans applications by examining the combinatorial effect of AI capabilities on
portation bottlenecks and enables prescriptive insights into workloads servitization. Therefore, it is of great importance that researchers
and risk management. Managers also should view AI capabilities as a revalidate our assumptions and configurations to further inspect the role
source of competitive advantage in advancing stronger energy policies. of customer value propositions in service innovation to resolve any
Manufacturers also need to prioritize the investment in AI algorithms conflicting empirical findings. All in all, the positive effects of AI capa
that enhance value proposition that serves sustainability causes, espe bilities and servitization in creating business and societal value are still
cially in industries facing a tough time immigrating to a new greener to be established in the literature (Raddats et al., 2019; Vinuesa et al.,
solution amid the current energy crisis. Therefore, managers can 2020). Our study finds that AI capabilities positively impact servitiza
advance their servitized market offerings by providing services that tion, while higher absorptive capacity is linked to a stronger impact of AI
reduce clients’ energy emissions, improve workplace safety, and extend capabilities on servitization. Our study also concludes that, ceteris par
product lifetime and overall environmental performance. Finally, ibus, AI capabilities designed for leveraging societal good, resource
managers should proactively pursue the current and future opportu optimization, and process optimization are sufficient to lead to serviti
nities rises from the assistive role of AI, by developing more powerful zation. Finally, this research is inclined to inspire additional scholarly
cognitive technologies, targeting process and human augmentation, inquiry linking AI capabilities to moderating conditions and nonlinear
service provision and governance, this can be achieved by enhancing the service innovation outcomes.
quality of information and the use of expert systems that provide a
platform for business adaptation and ultimately evolution (Dwivedi CRediT authorship contribution statement
et al., 2021).
Mohamad Abou-Foul: Methodology, Formal analysis, Conceptual
6. Limitations and conclusion ization, Data curation, Investigation, Writing - original draft. Jose L.
Ruiz-Alba: Methodology, Conceptualization, Investigation, Writing -
The findings of this research are instructive in many ways. However, review & editing. Pablo J. López-Tenorio: Writing – review & editing,
they come with some limitations that can be resolved in future research. Methodology, Formal analysis.
First, while using dynamic capabilities as a theoretical lens is justified,
we address its scope and conditions (Winter, 2003). Therefore, it is Declaration of Competing Interest
recommended to stress test our findings using different theoretical len
ses, especially for those companies which operate in ecosystems built on The authors declare that they have no known competing financial
process outsourcing to achieve operational flexibility, which might interests or personal relationships that could have appeared to influence
impede the internal development of AI capabilities. Furthermore, a good the work reported in this paper.
avenue of future research could be to incorporate organizations’ AI
maturity levels when collecting samples. Further empirical work should Data availability
also focus on AI governance, technical diligence, affordability, and
industry-level capabilities (Thuraisingham, 2020) to provide a more Data will be made available on request.
Appendix A. Constructs and measurement items.A
Variables Standardized factor

loading
AI capabilities b
AI customer value proposition
Our company is collecting after-sales insights and uses AI to personalize the customer experience and ensure our customers’ success. 0.82***
Our specialized data science team uses tools to calculate our customer’s optimal warranty cost and duration. 0.78***
Our company is using machine learning models in pricing and quoting optimization. 0.86***
Our company collects and analyses embedded sensor data to provide our customers with predictive maintenance, and operation optimization 0.76***
services.
AI key processes optimization
Our company is using advanced data science in demand forecasting and stocking. 0.77***
(continued on next page)
11
(continued )
Variables Standardized factor
loading
Our company is making a strategic data acquisition to fulfill customers’ orders on time. 0.73***
Our company integrates AI conversational agents’ capabilities such as chatbots in our next-generation CRM. 0.80***
Our company uses advanced robotics and predictive maintenance in our internal operations applications. 0.81***
Our company uses intelligence capabilities such as machine vision and edge analytics in enhancing yield optimization. 0.80***
Our company uses AI data mining capabilities and big data systems to enhance our product innovation process and bill of material (BOM). 0.84***
AI key resources optimization
Our company applies analytics to unified data warehouses to optimize our suppliers’ network. 0.78***
Our company uses AI applications to optimize our labor workforce. 0.80***
Our company uses advanced analytics to optimize our network’s resources, ensure cybersecurity and safeguard our data. 0.72***
Our company uses AI applications to identify our lowest-cost provider. 0.76***
AI societal good
Our company trains AI assistants to enhance workplace safety. 0.87***
Our company uses applied AI such as deep reinforcement learning to cut our operation’s energy consumption, emission, waste, and equity. 0.90***
Our company uses data analytics and benchmarks to provide green solutions to our customers that tackle the most prominent societal challenges 0.77***
such as decarbonization.
Servitization c
Top management service orientation
Our senior leaders are rarely aligned around the strategic importance of servitization transformation. r 0.76***
Our senior leaders are actively promoting a vision of the future that involves servitized offerings. 0.85***
We regularly review with the top team our progress on servitization transformation. 0.86***
Mobilization of resources
Our employees understand the benefits of servitization change. 0.72***
Our firm is investing in the necessary skills and capabilities to provide servitized offerings. 0.74***
Our business cases and key performance indicators are linked to our roadmap. 0.81***
Market offering
Our firm has taken over some of our customer’s business processes. 0.75***
Our firm has taken over the operational functions of our products in customers’ businesses. 0.79***
Our service contracts related to our products are designed to share ‘risk and reward’ with our customers, so our customers pay for the product’s 0.82***
capabilities, outcomes, and results.
Absorptive capacity d
The search for relevant information concerning our industry is everyday business in our company. 0.79***
Our management demands periodic cross-departmental meetings to discuss new developments, problems, and achievements. 0.87***
In our company there is a quick information flow, e.g., if a business unit obtains important information, it communicates this information promptly 0.65***
to all other business units or departments.
Our employees successfully link existing knowledge with new insights. 0.83***
Our company rarely reconsiders technologies and adapts them according to new knowledge. r 0.64***
Our company lacks the ability to work more effectively by adopting new technologies. r 0.80***
Control variables e
Firm size (LN employee) -
Firm age (Since incorporation) -
Industry classification (2 digits US-SIC code) -
Slack resources (Current ratio) -
a
All items were measured using a 5-point Likert scale with 1 = Strongly disagree, 2 = Disagree, 3 = Neither agree nor disagree, 4 = Agree, and 5 =
Strongly agree.
b
New items developed by authors.
c
Adopted from (Abou-foul et al., 2021).
d
Adopted from (Flatten et al., 2011).
e
Objective measures.
r
Item reversed.
***p < 0.001.
Appendix B. Second-order factor results.
Model Normed χ2 TLI NFI CFI RMSEA
Recommended Value a ≤2 >0.90 >0.90 >0.90 <0.08

AI Capabilities
Model 1 (one-factor model) 6.468 0.801 0.808 0.842 0.139
Model 2 (four uncorrelated factors) 5.362 0.937 0.914 0.922 0.170
Model 3 (four correlated factors) 2.512 0.930 0.924 0.956 0.081
Model 4 (one second-order factor) 1.944 0.947 0.932 0.965 0.062
a
Based on (Kline, 2015).
Normed χ2 = χ2/df (relative chi-square); TLI—Tucker Lewis index; NFI—normed fit index; CFI—comparative fit index; RMSEA—root mean square
error aggregate.
12
Appendix C. Heterotrait-Monotrait ratio (HMTM).
(1) (2) (3) (4) (5) (6)
(1) ACAP
(2) AIKPO 0.653
(3) AIKRO 0.700 0.772
(4) SERV 0.799 0.818 0.830
(5) AISGD 0.851 0.688 0.746 0.729
(6) AICVP 0.584 0.596 0.645 0.821 0.626
References Calabrese, A., Forte, G., & Ghiron, N. L. (2018). Fostering sustainability-oriented service
innovation (SOSI) through business model renewal: The SOSI tool. Journal of Cleaner
Production, 201, 783–791. https://doi.org/10.1016/j.jclepro.2018.08.102
Abou-foul, M., Ruiz-Alba, J. L., & Soares, A. (2021). The impact of digitalization and
Chandy, R. K., Johar, G. V., Moorman, C., & Roberts, J. H. (2021). Better Marketing for a
servitization on the financial performance of a firm: An empirical analysis.
Better World. Journal of Marketing, 85(3), 1–9. https://doi.org/10.1177/
Production Planning and Control, 32(12), 975–989. https://doi.org/10.1080/
00222429211003690
09537287.2020.1780508
Collins, D. (2003). Pretesting survey instruments: An overview of cognitive methods.
Adler, J. H. (1965). Absorptive Capacity: The Concept and Its Determinants. Brookings
Quality of Life Research, 12(3), 229–238. https://doi.org/10.1023/A:
Institution.
1023254226592
Agarwal, G. K., Simonsson, J., Magnusson, M., Hald, K. S., & Johanson, A. (2022). Value-
Collins, C., Dennehy, D., Conboy, K., & Mikalef, P. (2021). International Journal of
capture in digital servitization. Journal of Manufacturing Technology Management,
Information Management Artificial intelligence in information systems research : A
ahead-of-print(ahead-of-print). 10.1108/JMTM-05-2021-0168.
systematic literature review and research agenda. International Journal of Information
Ahlin, B., Drnovšek, M., & Hisrich, R. D. (2014). Exploring the moderating effects of
Management, 60(November 2020), Article 102383. https://doi.org/10.1016/j.
absorptive capacity on the relationship between social networks and innovation.
ijinfomgt.2021.102383
Journal of East European Management Studies, 19(2), 213–235. http://www.jstor.org/
Coreynen, W., Matthyssens, P., & Van Bockhaven, W. (2017). Boosting servitization
stable/24330972.
through digitization: Pathways and dynamic resource configurations for
Akter, S., Wamba, S. F., Mariani, M., & Hani, U. (2021). How to Build an AI Climate-
manufacturers. Industrial Marketing Management, 60, 42–53.
Driven Service Analytics Capability for Innovation and Performance in Industrial
Coreynen, W., Matthyssens, P., Vanderstraeten, J., & van Witteloostuijn, A. (2020).
Markets? Industrial Marketing Management, 97, 258–273. https://doi.org/10.1016/j.
Unravelling the internal and external drivers of digital servitization: A dynamic
indmarman.2021.07.014
capabilities and contingency perspective on firm strategy. Industrial Marketing
Ameen, N., Tarhini, A., Reppel, A., & Anand, A. (2021). Customer experiences in the age
Management, 89, 265–277.
of artificial intelligence. Computers in Human Behavior, 114, Article 106548. https://
Dangelico, R. M., Pujari, D., & Pontrandolfo, P. (2016). Green Product Innovation in
doi.org/10.1016/j.chb.2020.106548
Manufacturing Firms: A Sustainability-Oriented Dynamic Capability Perspective. June.
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A
10.1002/bse.1932.
review and recommended two-step approach. Psychological Bulletin, 103(3), 411.
Davenport, T., Guha, A., Grewal, D., & Bressgott, T. (2020). How artificial intelligence
Anderson, P. (1999). Perspective: Complexity theory and organization science.
will change the future of marketing. Journal of the Academy of Marketing Science, 48
Organization Science, 10(3), 216–232.
(1), 24–42. https://doi.org/10.1007/s11747-019-00696-0
Armstrong, J. S., & Overton, T. S. (1977). Estimating Nonresponse Bias in Mail Surveys.
Davenport, T. H., & Ronanki, R. (2018). Artificial intelligence for the real world. Harvard
Journal of Marketing Research, 14(3), 396–402. https://doi.org/10.1177/
Business Review, 96(1), 108–116.
002224377701400320
Davies, A., Brady, T., & Hobday, M. (2007). Organizing for solutions: Systems seller vs.
Assad, S., Clark, R., Ershov, D., & Xu, L. (2021). Algorithmic Pricing and Competition :
systems integrator. Industrial Marketing Management, 36(2), 183–193. https://doi.
Empirical Evidence from the German Retail Gasoline Market *. January, 1–77.
org/10.1016/j.indmarman.2006.04.009
Bag, S., Gupta, S., Kumar, A., & Sivarajah, U. (2021). An integrated artificial intelligence
Day, G. S. (1994). The Capabilities of Market-Driven Organizations. Journal of Marketing,
framework for knowledge creation and B2B marketing rational decision-making for
58(4), 37–52. https://doi.org/10.1177/002224299405800404
improving firm performance. Industrial Marketing Management, 92, 178–189. https://
Day, G. S., & Schoemaker, P. J. H. (2016). Adapting to Fast-Changing Markets and
doi.org/10.1016/j.indmarman.2020.12.001
Technologies. California Management Review, 58(4), 59–77. https://doi.org/
Bag, S., Pretorius, J. H. C., Gupta, S., & Dwivedi, Y. K. (2021). Role of institutional
10.1525/cmr.2016.58.4.59
pressures and resources in the adoption of big data analytics powered artificial
Desouza, K. C., Dawson, G. S., & Chenok, D. (2020). Designing, developing, and
intelligence, sustainable manufacturing practices, and circular economy capabilities.
deploying artificial intelligence systems: Lessons from and for the public sector.
Technological Forecasting and Social Change, 163, Article 120420.
Business Horizons, 63(2), 205–213. https://doi.org/10.1016/j.bushor.2019.11.004
Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal
Dillman, D. A. (2011). Mail and Internet surveys: The tailored design method–2007 Update
of the Academy of Marketing Science, 16(1), 74–94.
with new Internet, visual, and mixed-mode guide. John Wiley & Sons.
Baines, T., Lightfoot, H., Peppard, J., Johnson, M., Tiwari, A., Shehab, E., & Swink, M.
Dmitrijeva, J., Schroeder, A., Ziaee Bigdeli, A., & Baines, T. (2020). Context matters:
(2009). Towards an operations strategy for product-centric servitization.
How internal and external factors impact servitization. Production Planning &
International Journal of Operations & Production Management, 29(5), 494–519.
Control, 31(13), 1077–1097. https://doi.org/10.1080/09537287.2019.1699195
https://doi.org/10.1108/01443570910953603
Doni, F., Corvino, A., & Martini, S. B. (2019). Servitization and sustainability actions.
Baines, T., & Lightfoot, W. H. (2014). Servitization of the manufacturing firm.
Evidence from European manufacturing companies. Journal of Environmental
International Journal of Operations & Production Management, 34(1), 2–35. https://
Management, 234, 367–378.
doi.org/10.1108/IJOPM-02-2012-0086
Dubé, J.-P., & Misra, S. (2019). Personalized pricing and customer welfare. Available at
Barbieri, C., Rapaccini, M., Adrodegari, F., Saccani, N., & Baccarin, G. (2021). The Role of
SSRN 2992257.
AI Platforms for the Servitization of Manufacturing Companies BT - Smart Services
Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., …
Summit (S. West, J. Meierhofer, & C. Ganz (eds.); pp. 95–104). Springer International
Williams, M. D. (2021). Artificial Intelligence (AI): Multidisciplinary perspectives on
Publishing.
emerging challenges, opportunities, and agenda for research, practice and policy.
Barnett, W. P., & Pontikes, E. G. (2008). The Red Queen, Success Bias, and
International Journal of Information Management, 57, Article 101994. https://doi.org/
Organizational Inertia. Management Science, 54(7)(July 2008), 1237–1251. 10.1287/
10.1016/j.ijinfomgt.2019.08.002
mnsc.1070.0808.
El, D., & Akrout, H. (2020). Perceived design affordance of new products : Scale
Barrett, M., Davidson, E., Prabhu, J., & Vargo, S. L. (2015). Service innovation in the
development and validation. Journal of Business Research, 121(September 2019),
digital age. MIS Quarterly, 39(1), 135–154.
127–141. https://doi.org/10.1016/j.jbusres.2020.08.010
Blöcher, K., & Alt, R. (2020). AI and robotics in the European restaurant sector: Assessing
Enholm, I. M., Papagiannidis, E., Mikalef, P., & Krogstie, J. (2021). Artificial Intelligence
potentials for process innovation in a high-contact service industry. Electronic
and Business Value: A Literature Review. Information Systems Frontiers. https://doi.
Markets. https://doi.org/10.1007/s12525-020-00443-2
org/10.1007/s10796-021-10186-w
Boehmer, J. H., Shukla, M., Kapletia, D., & Tiwari, M. K. (2020). The impact of the
Escribano, A., Fosfuri, A., & Tribó, J. A. (2009). Managing external knowledge flows: The
Internet of Things (IoT) on servitization: An exploration of changing supply
moderating role of absorptive capacity. Research Policy, 38(1), 96–105. https://doi.
relationships. Production Planning & Control, 31(2–3), 203–219.
org/10.1016/j.respol.2008.10.022
Bughin, J., Hazan, E., Manyika, J., & Woetzel, J. (2017). Artificial intelligence: the next
Fang, E. (Er), Palmatier, R. W., & Steenkamp, J.-B. E. M. (2008). Effect of Service
digital frontier? McKinsey Global Institute. APO-210501.
Transition Strategies on Firm Value. Journal of Marketing, 72(5), 1–14. 10.1509/
Burström, T., Parida, V., Lahti, T., & Wincent, J. (2021). AI-enabled business-model
jmkg.72.5.001.
innovation and transformation in industrial ecosystems: A framework, model and
Felin, T., & Hesterly, W. S. (2007). The Knowledge-Based View, Nested Heterogeneity,
outline for further research. Journal of Business Research, 127, 85–95. https://doi.
and New Value Creation: Philosophical Considerations on the Locus of Knowledge.
org/10.1016/j.jbusres.2021.01.016
13
The Academy of Management Review, 32(1), 195–218. https://doi.org/10.2307/ Kowalkowski, C., Gebauer, H., Kamp, B., & Parry, G. (2017a). Industrial Marketing
20159288 Management Servitization and deservitization : Overview , concepts , and de fi nitions. 60,
Felin, T., & Powell, T. C. (2016). Designing Organizations for Dynamic Capabilities. 4–10. 10.1016/j.indmarman.2016.12.007.
California Management Review, 58(4), 78–96. https://doi.org/10.1525/ Kranz, J. J., Hanelt, A., & Kolbe, L. M. (2016). Understanding the influence of absorptive
cmr.2016.58.4.78 capacity and ambidexterity on the process of business model change – the case of on-
Ferreira, P., Teixeira, J. G., & Teixeira, L. F. (2020). Understanding the Impact of premise and cloud-computing software. Information Systems Journal, 26(5), 477–517.
Artificial Intelligence on Services. In H. Nóvoa, M. Dr\uagoicea, & N. Kühl (Eds.), https://doi.org/10.1111/isj.12102
Exploring Service Science (pp. 202–213). Springer International Publishing. Kumar, N., Stern, L. W., & Anderson, J. C. (1993). Conducting interorganizational
Fischer, T., Gebauer, H., Gregory, M., Ren, G., & Fleisch, E. (2010). Exploitation or research using key informants. In Academy of Management Journal (Vol. 36, Issue 6,
exploration in service business development? Journal of Service Management, 21(5), pp. 1633–1651). Academy of Management. 10.2307/256824.
591–624. https://doi.org/10.1108/09564231011079066 Kumar, V., Ramachandran, D., & Kumar, B. (2021). Influence of new-age technologies on
Fiss, P. C. (2011). Building better causal theories: A fuzzy set approach to typologies in marketing: A research agenda. Journal of Business Research, 125, 864–877. https://
organization research. The Academy of Management Journal, 54(2), 393–420. https:// doi.org/10.1016/j.jbusres.2020.01.007
doi.org/10.5465/amj.2011.60263120 Kushwaha, A. K., Kumar, P., & Kar, A. K. (2021). What impacts customer experience for
Flatten, T. C., Engelen, A., Zahra, S. A., & Brettel, M. (2011). A measure of absorptive B2B enterprises on using AI-enabled chatbots? Insights from Big data analytics.
capacity : Scale development and validation q. European Management Journal, 29(2), Industrial Marketing Management, 98, 207–221. https://doi.org/10.1016/j.
98–116. https://doi.org/10.1016/j.emj.2010.11.002 indmarman.2021.08.011
Garcia Martin, P. C., Schroeder, A., & Ziaee Bigdeli, A. (2019). The value architecture of LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444.
servitization: Expanding the research scope. Journal of Business Research, 104, https://doi.org/10.1038/nature14539
438–449. https://doi.org/10.1016/j.jbusres.2019.04.010 Lee, J., Suh, T., Roy, D., & Baucus, M. (2019). Emerging technology and business model
Gebauer, H., Paiola, M., Saccani, N., & Rapaccini, M. (2021). Digital servitization: innovation: The case of artificial intelligence. Journal of Open Innovation: Technology,
Crossing the perspectives of digitization and servitization. Industrial Marketing Market, and Complexity, 5(3), 44.
Management, 93, 382–388. https://doi.org/10.1016/j.indmarman.2020.05.011 Lehrer, C., Wieneke, A., vom Brocke, J., Jung, R., & Seidel, S. (2018). How Big Data
Gerbing, D. W., & Anderson, J. C. (1988). An Updated Paradigm for Scale Development Analytics Enables Service Innovation: Materiality, Affordance, and the
Incorporating Unidimensionality and Its Assessment. Journal of Marketing Research, Individualization of Service. Journal of Management Information Systems, 35(2),
25(2), 186–192. https://doi.org/10.2307/3172650 424–460. https://doi.org/10.1080/07421222.2018.1451953
Haefner, N., Wincent, J., Parida, V., & Gassmann, O. (2021). Artificial intelligence and Leone, D., Schiavone, F., Appio, F. P., & Chiao, B. (2021). How does artificial intelligence
innovation management: A review, framework, and research agenda✰. Technological enable and enhance value co-creation in industrial markets? An exploratory case
Forecasting and Social Change, 162, Article 120392. https://doi.org/10.1016/j. study in the healthcare ecosystem. Journal of Business Research, 129, 849–859.
techfore.2020.120392 https://doi.org/10.1016/j.jbusres.2020.11.008
Hair, J. F. (2010). Multivariate data analysis: A global perspective. Pearson Education. Li, T., Sahu, A. K., Talwalkar, A., & Smith, V. (2020). Federated learning: Challenges,
Hayes, A. F., Montoya, A. K., & Rockwood, N. J. (2017). The analysis of mechanisms and methods, and future directions. IEEE Signal Processing Magazine, 37(3), 50–60.
their contingencies: PROCESS versus structural equation modeling. Australasian Libai, B., Bart, Y., Gensler, S., Hofacker, C. F., Kaplan, A., & Benjamin, E. (2020).
Marketing Journal (AMJ), 1–6. https://doi.org/10.1016/j.ausmj.2017.02.001 ScienceDirect Brave New World ? On AI and the Management of Customer
Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing Relationships., 51, 44–56. https://doi.org/10.1016/j.intmar.2020.04.002
discriminant validity in variance-based structural equation modeling. Journal of the Majhi, S. G., Anand, A., Mukherjee, A., & Rana, N. P. (2021). The Optimal Configuration
Academy of Marketing Science, 43(1), 115–135. https://doi.org/10.1007/s11747- of IT-Enabled Dynamic Capabilities in a firm’s Capabilities Portfolio: A Strategic
014-0403-8 Alignment Perspective. Information Systems Frontiers. https://doi.org/10.1007/
Hercheui, M., & Ranjith, R. (2020). Improving organization dynamic capabilities using s10796-021-10145-5
artificial intelligence. Global Journal of Business Research, 14(1), 87–96. Malhotra, A., Gosain, S., & El Sawy, O. A. (2005). Absorptive Capacity Configurations in
Holgersson, M., Baldwin, C. Y., Chesbrough, H., Bogers, M., & M. l. a.. (2022). The Forces Supply Chains: Gearing for Partner-Enabled Market Knowledge Creation. MIS
of Ecosystem Evolution. California Management Review, 64(3), 5–23. https://doi.org/ Quarterly, 29(1), 145–187. https://doi.org/10.2307/25148671
10.1177/00081256221086038 Marcon, É., Marcon, A., Ayala, N. F., Frank, A. G., Story, V., Burton, J., Raddats, C., &
Huikkola, T., Kohtamäki, M., & Rabetino, R. (2016). Resource Realignment in Zolkiewski, J. (2022). Capabilities supporting digital servitization: A multi-actor
Servitization. Research-Technology Management, 59(4), 30–39. https://doi.org/ perspective. Industrial Marketing Management, 103, 97–116.
10.1080/08956308.2016.1185341 Marcon, É., Marcon, A., Le Dain, M.-A., Ayala, N. F., Frank, A. G., & Matthieu, J. (2019).
Huntingford, C., Jeffers, E. S., Bonsall, M. B., Christensen, H. M., & Lees, T. (2019). Barriers for the digitalization of servitization. Procedia CIRP, 83, 254–259. https://
Machine learning and artificial intelligence to aid climate change research and doi.org/10.1016/j.procir.2019.03.129
preparedness Machine learning and arti fi cial intelligence to aid climate change research Mikalef, P., Conboy, K., & Krogstie, J. (2021). Artificial intelligence as an enabler of B2B
and preparedness. marketing: A dynamic capabilities micro-foundations approach. Industrial Marketing
Iansiti, M., & Lakhani, K. (2020). Competing in the Age of AI: How machine intelligence Management, 98(August), 80–92. https://doi.org/10.1016/j.
changes the rules of business. January-February: Harvard Business Review. indmarman.2021.08.003
Jahanmir, S. F., Miranda, G., Gomes, P. J., & Martins, H. (2019). Determinants of users ’ Mikalef, P., & Gupta, M. (2021). Information & Management Artificial intelligence
continuance intention toward digital innovations : Are late adopters different ? capability : Conceptualization, measurement calibration, and empirical study on its
Journal of Business Research, June, 1–9. 10.1016/j.jbusres.2019.11.010. impact on organizational creativity and firm performance. Information &
Johnson, M. W., Christensen, C. M., & Kagermann, H. (2008). Reinventing your business Management, 58(3), Article 103434. https://doi.org/10.1016/j.im.2021.103434
model. Harvard Business Review, 86(12), 57–68. Misangyi, V. F., Greckhamer, T., Furnari, S., Fiss, P. C., Crilly, D., & Aguilera, R. (2016).
Jarvis, C., MacKenzie, S., & Podsakoff, P. (2003). A Critical Review of Construct Embracing Causal Complexity: The Emergence of a Neo-Configurational Perspective.
Indicators and Measurement Model Misspecification in Marketing and Consumer Journal of Management, 43(1), 255–282. https://doi.org/10.1177/
Research. Journal of Consumer Research, 30(2), 199–218. https://doi.org/10.1086/ 0149206316679252
376806 Mora Cortez, R., & Johnston, W. J. (2017). The future of B2B marketing theory: A
Kamp, B., Ochoa, A., & Diaz, J. (2017). Smart servitization within the context of historical and prospective analysis. Industrial Marketing Management, 66, 90–102.
industrial user–supplier relationships: Contingencies according to a machine tool https://doi.org/10.1016/j.indmarman.2017.07.017
manufacturer. International Journal on Interactive Design and Manufacturing (IJIDeM), Mora Cortez, R., & Johnston, W. J. (2018). Needed B2B marketing capabilities: Insights
11(3), 651–663. https://doi.org/10.1007/s12008-016-0345-0 from the USA and emerging Latin America. International Business Review, 27(3),
Kanninen, T., Penttinen, E., Tinnilä, M., & Kaario, K. (2017a). Exploring the dynamic 594–609. https://doi.org/10.1016/j.ibusrev.2017.10.008
capabilities required for servitization: The case process industry. Business Process Nishant, R., Kennedy, M., & Corbett, J. (2020). Artificial intelligence for sustainability:
Management Journal. Challenges, opportunities, and a research agenda. International Journal of Information
Kanninen, T., Penttinen, E., Tinnilä, M., & Kaario, K. (2017b). Exploring the dynamic Management, 53, Article 102104. https://doi.org/10.1016/j.ijinfomgt.2020.102104
capabilities required for servitization. Business Process Management Journal, 23(2), Neirotti, P., Pesce, D., & Battaglia, D. (2021). Algorithms for operational decision-
226–247. https://doi.org/10.1108/BPMJ-03-2015-0036 making: An absorptive capacity perspective on the process of converting data into
Kasie, F. M., Bright, G., & Walker, A. (2017). Decision support systems in manufacturing: relevant knowledge. Technological Forecasting and Social Change, 173, Article
A survey and future trends. Journal of Modelling in Management, 12(3), 432–454. 121088. https://doi.org/10.1016/j.techfore.2021.121088
https://doi.org/10.1108/JM2-02-2016-0015 O’Brien, R. M. (2007). A Caution Regarding Rules of Thumb for Variance Inflation
Kline, P. (2014). An easy guide to factor analysis. Routledge. Factors. Quality & Quantity, 41(5), 673–690. https://doi.org/10.1007/s11135-006-
Kline, R. B. (2015). Principles and practice of structural equation modeling. Guilford. 9018-6
publications. Paiola, M., & Gebauer, H. (2020). Internet of things technologies, digital servitization
Klumpp, M. (2018). Automation and artificial intelligence in business logistics systems: and business model innovation in BtoB manufacturing firms. Industrial Marketing
Human reactions and collaboration requirements. International Journal of Logistics Management, 89, 245–264.
Research and Applications, 21(3), 224–242. https://doi.org/10.1080/ Palo, T., Åkesson, M., & Löfberg, N. (2019). Servitization as business model contestation
13675567.2017.1384451 : A practice approach. Journal of Business Research, September 2017, 1–11. 10.1016/j.
Kohtamäki, M., Parida, V., Oghazi, P., Gebauer, H., & Baines, T. (2019). Digital jbusres.2018.10.037.
servitization business models in ecosystems: A theory of the firm. Journal of Business Pappas, I. O., & Woodside, A. G. (2021). Fuzzy-set Qualitative Comparative Analysis
Research, 104, 380–392. https://doi.org/10.1016/j.jbusres.2019.06.027 (fsQCA): Guidelines for research practice in Information Systems and marketing.
International Journal of Information Management, 58, Article 102310. https://doi.org/
10.1016/j.ijinfomgt.2021.102310
14
Parida, V., Sjödin, D. R., Lenka, S., & Wincent, J. (2015). Developing Global Service Spring, M., & Araujo, L. (2017). Product biographies in servitization and the circular
Innovation Capabilities: How Global Manufacturers Address the Challenges of economy. Industrial Marketing Management, 60, 126–137.
Market Heterogeneity. Research-Technology Management, 58(5), 35–44. https://doi. Suppatvech, C., Godsell, J., & Day, S. (2019). The roles of internet of things technology in
org/10.5437/08956308X5805360 enabling servitized business models: A systematic literature review. Industrial
Park, Y., & Mithas, S. (2020). Organized Complexity of Digital Business Strategy: A Marketing Management, 82, 70–86. https://doi.org/10.1016/j.
Configurational Perspective. Mis Quarterly, 44(1). indmarman.2019.02.016
Pavlou, P. A., & El Sawy, O. A. (2006). From IT Leveraging Competence to Competitive Syam, N., & Sharma, A. (2018). Waiting for a sales renaissance in the fourth industrial
Advantage in Turbulent Environments : The Case of New Product Development. revolution: Machine learning and artificial intelligence in sales research and
Information Systems Research, 17(3), 198–227. https://doi.org/10.1287/isre. practice. Industrial Marketing Management, 69, 135–146. https://doi.org/10.1016/j.
1060.0094 indmarman.2017.12.019
Payne, M., H, E., Dahl, A. J., & Peltier, J. (2021). Digital servitization value co-creation Teece, D. J. (2007). explicating dynamic capabilities: The nature and microfoundations
framework for AI services: a research agenda for digital transformation in financial of (sustainable) enterprise performance. Strategic Management Journal Strat. Mgmt. J,
service ecosystems. Journal of Research in Interactive Marketing, 15(2), 200–222. 28, 1319–1350. https://doi.org/10.1002/smj.640
10.1108/JRIM-12-2020-0252. Teece, D. J. (2010). Business Models, Business Strategy and Innovation. Long Range
Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common Planning, 43(2–3), 172–194. https://doi.org/10.1016/j.lrp.2009.07.003
method biases in behavioral research: A critical review of the literature and Teece, D. J. (2023). The Evolution of the Dynamic Capabilities Framework BT -
recommended remedies. In Journal of Applied Psychology (Vol. 88, Issue 5, pp. Artificiality and Sustainability in Entrepreneurship: Exploring the Unforeseen, and
879–903). American Psychological Association. 10.1037/0021-9010.88.5.879. Paving the Way to a Sustainable Future (R. Adams, D. Grichnik, A. Pundziene, & C.
Polit, D. F., & Beck, C. T. (2006). The content validity index: Are you sure you know Volkmann (eds.); pp. 113–129). Springer International Publishing. 10.1007/978-3-
what’s being reported? Critique and recommendations. Research in Nursing & Health, 031-11371-0_6.
29(5), 489–497. https://doi.org/10.1002/nur.20147 Thon, C., Finke, B., Kwade, A., & Schilde, C. (2021). Artificial Intelligence in Process
Provost, F., & Fawcett, T. (2013). Data science and its relationship to big data and data- Engineering. Advanced Intelligent Systems, 3(6), 2000261. 10.1002/aisy.202000261.
driven decision making. Big Data, 1(1), 51–59. Thornhill, S., & Amit, R. (2003). Learning about Failure: Bankruptcy, Firm Age, and the
Rachinger, M., Rauter, R., Müller, C., Vorraber, W., & Schirgi, E. (2019). Digitalization Resource-Based View. Organization Science, 14(5), 497–509. http://www.jstor.org/
and its influence on business model innovation. Journal of Manufacturing Technology stable/4135145.
Management, 30(8), 1143–1160. https://doi.org/10.1108/JMTM-01-2018-0020 Thuraisingham, B. (2020). Artificial Intelligence and Data Science Governance: Roles
Raddats, C., Kowalkowski, C., Benedettini, O., Burton, J., & Gebauer, H. (2019). and Responsibilities at the C-Level and the Board. 2020 IEEE 21st International
Servitization: A contemporary thematic review of four major research streams. Conference on Information Reuse and Integration for Data Science (IRI), 314–318.
Industrial Marketing Management, 83, 207–223. https://doi.org/10.1016/j. 10.1109/IRI49571.2020.00052.
indmarman.2019.03.015 Tippins, M. J., & Sohi, R. S. (2003). IT competency and firm performance: Is
Raddats, C., Naik, P., & Ziaee Bigdeli, A. (2022). Creating value in servitization through organizational learning a missing link? Strategic Management Journal, 24(8),
digital service innovations. Industrial Marketing Management, 104, 1–13. https://doi. 745–761. https://doi.org/10.1002/smj.337
org/10.1016/j.indmarman.2022.04.002 Todorova, G., & Durisin, B. (2007). Absorptive Capacity: Valuing a Reconceptualization.
Ragin, C. C. (2009). Redesigning social inquiry: Fuzzy sets and beyond. Social Forces, 88 The Academy of Management Review, 32(3), 774–786. https://doi.org/10.2307/
(4), 1936–1938. 20159334
Rieke, N., Hancox, J., Li, W., Milletarì, F., Roth, H. R., Albarqouni, S., Bakas, S., Toma, N., Cornebise, J., Hutter, F., Mohamed, S., Picciariello, A., Connelly, B., Belgrave,
Galtier, M. N., Landman, B. A., Maier-hein, K., Ourselin, S., Sheller, M., & D. C. M., Ezer, D., & Haert, F. C. Van Der. (2020). AI for social good: unlocking the
Summers, R. M. (2020). The future of digital health with federated learning. Npj opportunity for positive impact. 10.1038/s41467-020-15871-z.
Digital Medicine, 1–7. https://doi.org/10.1038/s41746-020-00323-1 Torres, P., Augusto, M., & Wallace, E. (2018). Improving consumers’ willingness to pay using
Ritter, T., & Lund, C. (2020). Digitization capability and the digitalization of business social media activities. January. 10.1108/JSM-07-2017-0257.
models in business- to-business firms : Past, present, and future. Industrial Marketing Tronvoll, B., Sklyar, A., Sörhammar, D., & Kowalkowski, C. (2020). Transformational
Management, 86(February 2019), 180–190. https://doi.org/10.1016/j. shifts through digital servitization. Industrial Marketing Management, 89, 293–305.
indmarman.2019.11.019 https://doi.org/10.1016/j.indmarman.2020.02.005
Rosa, M., Feyereisl, J., & Collective, T. G. (2016). A framework for searching for general Tseng, M.-L., Tran, T. P. T., Ha, H. M., Bui, T.-D., & Lim, M. K. (2021). Sustainable
artificial intelligence. ArXiv Preprint ArXiv:1611.00685. industrial and operation engineering trends and challenges Toward Industry 4.0: A
Rossiter, J. (2011). Marketing measurement revolution: The C-OAR-SE method and why data driven analysis. Journal of Industrial and Production. Engineering, 38(8),
it must replace psychometrics. European Journal of Marketing, 45, 1561–1588. 581–598. https://doi.org/10.1080/21681015.2021.1950227
Rossiter, J. R. (2002). The C-OAR-SE procedure for scale development in marketing. Ulaga, W., & Reinartz, W. J. (2011). Hybrid Offerings: How Manufacturing Firms
International Journal of Research in Marketing, 19(4), 305–335. Combine Goods and Services Successfully. Journal of Marketing, 75(6), 5–23. https://
Rothaermel, F. T., & Alexandre, M. T. (2009). Ambidexterity in technology sourcing: The doi.org/10.1509/jm.09.0395
moderating role of absorptive capacity. Organization Science, 20(4), 759–780. Valtakoski, A. (2017). Explaining servitization failure and deservitization: A knowledge-
Rust, R. T., & Cooil, B. (1994). Reliability measures for qualitative data: Theory and based perspective. Industrial Marketing Management, 60, 138–150. https://doi.org/
implications. Journal of Marketing Research, 31, 1–14. https://doi.org/10.2307/ 10.1016/j.indmarman.2016.04.009
3151942 van Wynsberghe, A. (2021). Sustainable AI: AI for sustainability and the sustainability of
Sabri, O., & Obermiller, C. (2012). Consumer perception of taboo in ads. Journal of AI. AI and Ethics, 1(3), 213–218.
Business Research, 65(6), 869–873. https://doi.org/10.1016/j.jbusres.2011.01.009 Vázquez-Canteli, J. R., & Nagy, Z. (2019). Reinforcement learning for demand response:
Samuel, A. L. (1959). Some studies in machine learning using the game of checkers. IBM A review of algorithms and modeling techniques. Applied Energy, 235, 1072–1089.
Journal of Research and Development, 3(3), 210–229. Vinuesa, R., Azizpour, H., Leite, I., Balaam, M., Dignum, V., Domisch, S., Felländer, A.,
Sanders, N. R., Boone, T., Ganeshan, R., & Wood, J. D. (2019). Sustainable Supply Chains Langhans, S. D., Tegmark, M., & Fuso Nerini, F. (2020). The role of artificial
in the Age of AI and Digitization: Research Challenges and Opportunities. Journal of intelligence in achieving the Sustainable Development Goals. Nature
Business Logistics, 40(3), 229–240. https://doi.org/10.1111/jbl.12224 Communications, 11(1), 233. https://doi.org/10.1038/s41467-019-14108-y
Sandy, C. J., Gosling, S. D., Schwartz, S. H., Koelkebeck, T., Sandy, C. J., Gosling, S. D., Wales, W. J., Patel, P. C., Parida, V., & Kreiser, P. M. (2013). Nonlinear Effects of
Schwartz, S. H., & Koelkebeck, T. (2016). The Development and Validation of Brief Entrepreneurial Orientation on Small Firm Performance: The Moderating Role of
and Ultrabrief Measures of Values The Development and Validation of Brief and Resource Orchestration Capabilities. Strategic Entrepreneurship Journal, 7(2), 93–121.
Ultrabrief Measures of Values. Journal of Personality Assessment, 1–11. https://doi. Waltersmann, L., Kiemel, S., Stuhlsatz, J., Sauer, A., & Miehe, R. (2021). Artificial
org/10.1080/00223891.2016.1231115 Intelligence Applications for Increasing Resource Efficiency in Manufacturing
Schneider, C. Q., & Wagemann, C. (2010). Standards of Good Practice in Qualitative Companies—A Comprehensive Review. In Sustainability (Vol. 13, Issue 12). 10.3390/
Comparative Analysis (QCA) and Fuzzy-Sets. Comparative Sociology, 9(3), 397–418. su13126689.
https://doi.org/10.1163/156913210X12493538729793 Wang, K., & Wang, Y. (2017). How AI affects the future predictive maintenance: A primer
Scuotto, V., Magni, D., Palladino, R., & Nicotra, M. (2022). Triggering disruptive of deep learning. International Workshop of Advanced Manufacturing and Automation,
technology absorptive capacity by CIOs. Explorative research on a micro-foundation 1–9.
lens. Technological Forecasting and Social Change, 174, Article 121234. https://doi. Wheeler, B. C. (2002). NEBIC: A dynamic capabilities theory for assessing net-
org/10.1016/j.techfore.2021.121234 enablement. Information Systems Research, 13(2), 125–146.
Siachou, E., Vrontis, D., & Trichina, E. (2021). Can traditional organizations be digitally Wilson, H. J., & Daugherty, P. R. (2018). Collaborative intelligence: Humans and AI are
transformed by themselves? The moderating role of absorptive capacity and strategic joining forces. Harvard Business Review, 96(4), 114–123.
interdependence. Journal of Business Research, 124, 408–421. Winter, S. G. (2003). Understanding dynamic capabilities. Strategic Management Journal,
Sjödin, D., Parida, V., Kohtamäki, M., & Wincent, J. (2020a). An agile co-creation process 24(10), 991–995.
for digital servitization : A micro-service innovation approach. Journal of Business Wirtz, B. W., Weyerer, J. C., & Geyer, C. (2018). Artificial Intelligence and the Public
Research, 112(June 2019), 478–491. https://doi.org/10.1016/j.jbusres.2020.01.009 Sector — Applications and Challenges Artificial Intelligence and the Public Sector —
Sjödin, D., Parida, V., Palmié, M., & Wincent, J. (2021). How AI capabilities enable Applications and Challenges. International Journal of Public Administration, 00(00),
business model innovation: Scaling AI through co-evolutionary processes and 1–20. https://doi.org/10.1080/01900692.2018.1498103
feedback loops. Journal of Business Research, 134, 574–587. https://doi.org/ Woodside, A. G. (2014). Embrace • perform • model: Complexity theory, contrarian case
10.1016/j.jbusres.2021.05.009 analysis, and multiple realities ☆. Journal of Business Research, 67(12), 2495–2503.
Slotegraaf, R. J. (2007). Embeddedness of Organizational Capabilities *. 38(3), 451–488. https://doi.org/10.1016/j.jbusres.2014.07.006
Smith, B. C. (2019). The Promise of Artificial Intelligence Reckoning and Judgment. MIT Yao, X., Zhou, J., Zhang, J., & Boër, C. R. (2017). From Intelligent Manufacturing to
Press. Smart Manufacturing for Industry 4.0 Driven by Next Generation Artificial
15
Intelligence and Further On. 2017 5th International Conference on Enterprise Systems José L. Ruiz-Alba. Research & Knowledge Exchange Leader, School of Management &
(ES), 311–318. 10.1109/ES.2017.58. Marketing. Professor of Marketing (University of Westminster London, UK). Professional
Zadeh, L. A. (1996). Fuzzy sets. In Fuzzy sets, fuzzy logic, and fuzzy systems: selected papers experience in service firms and in more than 15 Universities. Member of the Editorial
by Lotfi A Zadeh (pp. 394–432). World Scientific. Advisory Board of the Journal of Business & Industrial Marketing; Associate Editor Interna
Zahra, S. A., & George, G. (2002). Absorptive Capacity: A Review, Reconceptualization, tional Journal of Internet Marketing and Advertisingand of the European Research on Man
and Extension. The Academy of Management Review, 27(2), 185–203. https://doi.org/ agement and Business Economics. Best paper award (Servitization Conference). Co-chair
10.2307/4134351 CBIM2018 and CBIM2021 International Conference. Publications in the Journal of Business
Zhang, M., Zhao, X., Lyles, M., Zhang, M., & Lyles, M. (2018). Effects of absorptive Research, International Journal of Hospitality Management, Decision Support Systems, Journal
capacity, trust and information systems on product innovation. 10.1108/IJOPM-11- of Environmental Management, Production Planning & Control, Service Industries Journal, and
2015-0687. Journal of Business and Industrial Marketing amongst others. He is the lead of the Service
Zhang, Z., Xu, D., Ostrosi, E., & Cheng, H. (2020). Optimization of the Product–Service Research and Customer Experience Group (University of Westminster).
System Configuration Based on a Multilayer Network. In Sustainability (Vol. 12, Issue
2). 10.3390/su12020746.
Pablo J. López-Tenorio. Associate lecturer in Marketing at UNIE Universidad in Madrid.
In the professional field, Dr. López-Tenorio is the creator of pablotenorio.com, a website
Mohamad Abou-foul. (PhD, MSc, BSc. Hons) is an assistant professor at Al Azhar uni specialized in consulting services and online training in Data-Driven Marketing. Previ
versity in Palestine, his research examines the role of servitization in driving competitive ously, he has developed his professional activity for more than 20 years holding positions
advantage, and his research interests focus on the digital economy, collaborative con of responsibility in Marketing Research and Analytics areas at Altadis-Imperial Tobacco
sumption platforms, applied AI, IoT, and digitalbusiness transformation. He is published in and Repsol.In the academic field, Dr. López-Tenorio have held management positions in
leading academic journals such as the Journal of Production Planning&Control, and Infor the academic area of Marketing at ESIC Business & Marketing School and UNIR (Uni
mation Technology & People. He received his BSc in information technology and business versidad Internacional de la Rioja). Dr. López-Tenorio’s research interest includes the
from the Northumbria university of Newcastle, School of Computing, Engineering and analysis of advertising effectiveness and brand image applied to sport management. He has
Information Science, his MSc in business management from Swansea University, UK and published several books on measuring return on marketing investment and has had articles
holds a Ph.D. degree in business management from the University of West London, UK. published in Journal of Marketing Management and Event Management.
16

1 s2.0 S0148296322010748 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S0148296322010748 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S0148296322010748 Main

Uploaded by

Copyright:

Available Formats

Journal of Business Research 157 (2023) 113609

Contents lists available at ScienceDirect

Journal of Business Research

The impact of artificial intelligence capabilities on servitization: The

1. Introduction et al., 2019; Spring & Araujo, 2017).

Fig. 1. Research Model.

2.2.2. Moderating role of absorptive capacity Table 2

1 AICVP 3.6 1 0.88 0.65 (0.82)

Fig. 4. Moderating effect of absorptive capacity.

Appendix A. Constructs and measurement items.A

Variables Standardized factor

Appendix B. Second-order factor results.

Model Normed χ2 TLI NFI CFI RMSEA

Recommended Value a ≤2 >0.90 >0.90 >0.90 <0.08

Appendix C. Heterotrait-Monotrait ratio (HMTM).

(1) (2) (3) (4) (5) (6)

You might also like