Technology has Agency too! Disentangling Technology's Causal and Structural Agency in a Service System

Actor-Network Theory and Agency

ANT is particularly suitable for studying technology's role and potential agency in service systems. The ANT perspective arose from the social construction of technology (Bijker and Law 1992). Scholars have encouraged the application of ANT to extend understanding of technology in value cocreation (Kaartemo and Helkkula 2018). Vargo and Lusch (2016) acknowledge that ANT is “theoretically consistent and robust” with S-D logic's conceptualization of value cocreation, as it recognizes a non-hierarchical networked structure in which value is created (p.7). Human agency is not the standard measure of agency for ANT. Non-humans, including technology, are recognized as actors on par with social actors (Latour 2005), allowing us, ontologically and methodologically, to follow technology in a service system.

Applying ANT can extend our understanding of how technology attains agency within a service system from both a conceptual framing and “methodological bracing” (Sayes 2014, p.136). Specifically, ANT enables researchers to follow a non-human actor, such as technology, regardless of its resource status or agency. ANT aligns with the S-D logic perspective by acknowledging that all human actors participate in network creation, including consumers and firm actors (Vargo and Lusch 2016; Wieland, Hartmann, and Vargo 2017). A central distinction of ANT, however, is that human and non-human actors participate in the network (Latour 2005; Law 1993), and “no actor is bigger than another” (Callon and Latour 1981, p.281). A non-human actor has agency “to the extent that they affect the actions of other actors” (Latour 2005; Martin and Schouten 2014, p.857). Importantly, ANT recognizes that any actor (human or non-human) becomes an intermediary when relied upon in the ecosystem. This classification of an intermediary expands upon S-D logic's concept of “resource-integrator,” the social actors that engage in service exchange to cocreate value (Vargo and Lusch 2016). Every act involves the entire ecosystem of actors to construct and enact the phenomenon of interest. S-D logic views the roles, norms, and meanings within the network as endogenously molded by the social actors. ANT extends this conceptualization to realize technology's agency and, as a result, influences the institutional norms. In summary, technologies are part of the system that constitutes action (Hatch 2018) and should be studied as an integral part of a system.

ANT also recognizes the dynamic movement of networks as a result of everchanging actor entry and connections, where destabilization can be more common than stabilization (Law 1993). Because networks are always in flux, the impact of an individual actor can be traced (Latour 2005) by focusing on network connections as they are created, changed, or severed and exploring both network stabilization and destabilization (Callon 1993; Law 1993).

Causal and structural agency

We classify the attainment of agency with the movement of a resource from operand to operant status and leverage ANT to recognize different agency forms based on an actor's ability to change a structure or influence others in a system (Latour 2005; Vargo 2018). Technology with no connection still holds status as an actor, or actant, but maintains as an operand resource and does not have agency. Notably, an operand resource may have proposed value, but this does not assume the resource influences others’ practices (i.e., causal agency), nor might it change the structure of a service system (i.e., structural agency). Any actor, human or technology, must instead influence value cocreation and/or value destabilization to be classified as an operant resource.

As an actor builds connections and becomes an intermediary between others or an enabler of action, the technology is classified as an operant resource due to its causal agency (i.e., when an actor has a causal effect; Sayes 2014). When an actor has causal agency, the actor can “authorize, allow, afford, encourage, permit, suggest, influence, block, render possible, forbid, and so on” (Latour 2004, 2005 quoted in Sayes 2014, p.141). In other words, causal agency can be classified if it can create some sort of difference to others within the system. Specifically related to a service system, causal agency is attained when an actor offers value and can influence others through exchange as an operant resource. Importantly, when an actor attains causal agency, meaning and value transfer are possible (Sayes 2014; Zhao 2022).

Nevertheless, an operant resource's agency may move beyond causal capabilities (Sayes 2014), as it can impact the structure and thus is classified as holding structural agency. Structural agency is attained when an actor participates in value or service exchange, resulting in changing the arrangement and order within the system. At this stage, the actant retains operant resource status and gains mediator classification. As stated by Sayes:

Nonhumans that enter into the human collective are endowed with a certain set of competencies by the network that they have lined up behind them. At the same time, they demand a certain set of competencies by the actors they line up, in turn. Nonhumans, in this rendition, are both changed by their circulation and change the collective through their circulation. (2005, p.138)

As a result, these actors, even technology, will place demands on others in the structure. Resultantly, not only meaning and value but also practices and role change within the system are evident (Latour 2005; Sayes 2014; Zhao 2022). The structural agency capability is aligned with sociological perspectives that focus on system impacts that do not inherently require conscious and deliberate decision-making, often inscribed in cognition-driven perspectives (Zhao 2022). See Table 1 for a review of agency-related definitions.

OPEN IN VIEWER

Table 1.

Definitions of Agency.

Agency form	Definition
Agency	Only available to humans and privileges their ability to catalyze assets for resource integration. Actor possesses cognitive capabilities to support need recognition and formulate value propositions available for value exchange (Zhao 2022).
Causal agency	Available to human or non-human actors that have a causal effect on other actors. Actor offers value and can influence others through direct exchange as an operant resource and intermediary; meaning and value transfer are possible (Sayes 2014; Zhao 2022).
Structural agency	Available to human or non-human actors who participate in value or service exchange that results in a change in the structure as an operant resource and mediator within the system; meaning, value, practices, and role change within the system is evident (Latour 2005; Sayes 2014; Zhao 2022).

Conceptualizing technology's agency translation

Leveraging Callon's four moments of translation as a starting point, we map these moments across steps of value cocreation in S-D logic. We also introduce two additional aspects of a technology's movement to become an agential actor in a service system, specifically connection failure and value destabilization. Our proposed model of Technology's Agency in Value Cocreation and Destabilization (see Figure 1) helps inform the nuance of the contextual value determinations of the VR-tour entrance into the residential real estate service system and provides a map of understanding for diverse emerging technology in other service systems.

OPEN IN VIEWER

Figure 1.

Technology's Agency in Value Cocreation and Destabilization.

To investigate how new actors (human or non-human) can influence a network and attain agency, ANT research offers four moments of translation wherein actors achieve roles in a network and act to make connections to “assemble, stabilize, or destabilize a network” (Callon 1984; Martin and Schouten 2014, p.857). The first moment is problematization, in which actors realize a problem or recognize a goal (Callon 1984). The next moment is interessement, where new actors are introduced and given roles in a network in an attempt to impose influence on the network (Callon 1984). Next, enrollment “persuades representatives to engage in a concrete alliance,” or connection (Callon 1984; Giesler 2012, p.56). Lastly, mobilization signifies reliance on new connections, leading to network stabilization (Callon 1984; Giesler 2012).

Goal recognition

In the seemingly stable ecosystem, common roles and practices are adopted to pursue the multiple actors’ goals. The actors’ goals, or problematization, within the service system, provide a pathway of entry for the new technology, or intressement. This initial stage of problematization can also be considered goal recognition within the service system. The concept of resourcesness within S-D logic literature recognizes how an operand or operant resource becomes available for actors to recognize contextual value within a service system (Koskela-Huotari and Vargo 2016). In problematization, a new entrant (or resource) “becomes” conceptualized within a service system as a result of a problem that needs to be solved; however, the resourcesness process of attaining value has not yet occurred. To be clear, recognizing a goal to enable resourceness is an action available for actors with agency, likely a human. The birth of a technology resource is possible through another actor's agency. Notably, the becoming of a resource is held as a formulation for a potential operand resource, and no technology agency is present within the service system.

Value proposition

In interessement, new actors are introduced, assume roles, and attempt to take action (Callon 1984; Giesler 2012); value propositions arise (Akaka and Vargo 2014). The actor attempts to play a role in the network, yet its perceived “resourcesness” or value determination has not occurred. While the actor exists, it remains an operand resource and does not hold agency. It does not influence others and is not reshaping the system.

Value determination

Next, the new actor begins interacting with other actors in the service system. If the technology connects with others through enrollment, it begins shifting as a resource, eligible for value determination . There is an opportunity for the technology to fulfill its role and prescribed actions, opening the door to becoming an operant resource and attaining agency. These interactions are considered value-creating actions and can establish connections, becoming an intermediary (Latour 2005) within the system. At this stage, there is potential for the new technology to transition from a non-agential operand resource into an agential operant resource. As an intermediary, the technology “transports meaning or force without transformation” to the ecosystem (Latour 2005, p.39). Its station is in place, and its actions are causal based on other actors’ expectations and embedded meanings.

Notably, value determination only sometimes matches the value proposition (Akaka and Vargo 2014). Although value-creating actions that begin as value propositions are similarly determined by the other actors (expected value-creating actions), value determinations may be greater than or utterly different from the original value proposition (e.g., unexpected value-creating actions). In other words, the new technology actor can transform the value within the service system; these are early indicators of potential structural agency. Nevertheless, structural outcomes may initially be absent.

Connection failure

We also recognize connection failure as an important aspect of this process when connections are not made, and the value proposition is rejected. This is akin to service failure and is unlikely to lead to positive value perceptions or cocreation. Agency is unlikely acquired at this stage. It is not atypical for individuals within service ecosystems to reject value propositions as they work against potential change (Vargo, Wieland, and Akaka 2015, 2020), where power is at play (Williams, Davey, and Johnstone 2021). There are cases when these actions could reshape expectations within a service environment as a reaction to the new actor (e.g., value co-destruction; Echeverri and Skålén 2011; Plé and Cáceres 2010; Zainuddin, Dent, and Tam 2017). Although this is not encapsulated within the present empirical study, this could be recognized as structural agency.

Value cocreation

Value cocreation arises from all actors’ contribution to the exchange and attainment of their aligned goals, or “mutual betterment” (Grönroos and Voima 2013; Karpen, Bove, and Lukas 2012, p.22). Through this mobilization process, the new actor becomes relied upon and impacts the wider network and others’ roles. Actors’ goals are achieved by adopting the new entrant where all actors in the network accept and embrace the role of the new technology. The “durability of the bonds” (Callon 1984, p.32) allows actors to attain their goals and reinforce the value-creating actions, or value potential, resulting in value cocreation (Figueiredo and Scaraboto 2016). Outcomes from interactions offer empirical shadows of value that are cocreated, and the system viability increases (Barrett et al. 2015). Because of the cocreative process within the system, the technology may shift to mediator status as a result of its unanticipated ability to “transform, translate, distort, and modify the meaning or the elements they are supposed to carry” (Latour 2005, p.39), indicating the structural agency attained.

Both adoption and diffusion stabilize the system and result in value cocreation (Giesler 2012). Value cocreation through diffusion is evident when other actors intentionally surround the new actor with other actors. By creating additional connections, diffusion of the technology influences value cocreation through network growth that fosters overall network stabilization for the focal actor. The technology moves beyond being a stand-in actor or intermediary. Acting as a spokesperson, the other actor shifts part of their resource constellation to the new actor. Past literature recognizes technology as an operant resource in technology diffusion (Vargo, Akaka, and Wieland 2020). However, to clarify this further, the new technology (operant resource) can secure structural agency through this diffusion process as the other actors modify their behavior and related service systems due to connecting with new technology.

Value destabilization

Simultaneously, actors’ roles and practices evolve as a result of the transformed value cocreated, resulting in long-term positive and negative consequences as detailed in the macromarketing literature (Baker et al. 2015; Domegan et al. 2020; Layton 2019). These unintended values (Caridà, Edvardsson, and Colurcio 2019) create new expectations and roles for the existing actors in the network, resulting in network disruption, or value destabilization . We define value destabilization as a change in value desired and perceived within a network when actors renegotiate roles and practices, often resulting from a new entrant. Outcomes of value destabilization may include changing activities within their roles, renegotiating reliance on connections, establishing new connections, and/or severing connections. Value destabilization moves beyond creative destruction (Schumpeter 1942) within the system that aims toward economic productivity but creates a wider impact on the expected actor roles and practices. As suggested in the conceptual development of service systems, micro-level interactions through value-creating action can reshape higher structures. This occurs as a result of modifying symbolic meanings and valuation. Symbolic meanings and valuation are negotiated through the “socio-technical processes of human actors” (Vargo, Wieland, and Akaka 2015), where these activities move across “blurred boundaries” (p.69). Specifically, operant resources, including technology, can “re-interpret symbols in service systems to develop new meanings, and ultimately new ways of creating value” (Akaka and Vargo 2014, p.372).

Value-creating actions can create new engagement rules that impact actors’ roles and practices in the service system and overlapping systems. We propose and classify this process as value destabilization, as systems are not stable but constantly shifting. Nevertheless, it is important to note that what emerges as value cocreation can be an outcome of mutual benefit, but it may not benefit all actors within the service system. Because of the reshaping, power dynamics, and ultimately, the roles and practices are refashioned – resulting in potentially expanded job responsibilities or even lost jobs. Resultingly, the technology clearly holds structural agential status due to its impact on the system(s). This agency provides a greater bearing to higher-level structures that are not limited to the service system but may hold macro implications as well (Sayes 2014; Siddike, Hidaka, and Kohda 2021).

Residential Real Estate Research Setting

While understudied in our literature, real estate is arguably one of the most significant consumer consumption decisions, financially and socially. In 2022, the housing market accounted for 16.2% of GDP in the United States (Wade 2023), comprising a substantial proportion of economic exchange. In the residential real estate service system, novel marketing technology has historically been engaged for its potential efficiencies of information exchange and enhanced performance (i.e., decreased days on the market, increased purchase price) (Allen et al. 2015; Benefield et al. 2019; Yu et al. 2020). The home, however, is embedded in cultural and social roles, and tracing technology in the real estate context is as much a social and cultural issue as it is a technical one (Huston and Warren 2013; Thompson 2015). On the surface, the exchange of the home moves from the seller to the buyer in the service system. However, the exchange resides more commonly in a larger network of human and non-human actors. The exchange is supported by real estate agents, where a seller's agent guides the de-identification and marketization of the home. Actors in the home exchange network can include buyers, sellers, real estate agents (i.e., buyer- and seller-agents), service providers (i.e., photographers), the physical home, and the novel virtual twin of the home (i.e., VR-tour).

The system of residential real estate surrounds the exchange of a tangible offering, specifically, a home. Two key consumer actors include a buyer and a seller. Yet, consumers do not have the expertise nor the time to invest in this expertise. Inherently, they are transient, secondary actors in this market. Buyers and sellers alike are constrained by necessary requirements such as purchasing a home within their budget and finding a home that meets their requirements of bedrooms, square footage, and location. For sellers, there is a need to hire an agent who understands the market and has the knowledge and skills necessary to sell their home for a specific price to not ‘go in the red.’ While rational processing is vital in their home exchange, the experience can be fraught with emotion. Many times, the process begins as a fun and enjoyable experience. However, the process quickly turns sour and “overwhelming.” Sellers find “bittersweet emotions” in transitioning to their next abode. Additionally, exhaustion can take over, leaving a need for additional reliance on their agent, and the focused objective shifts to quickly complete the home exchange.

The desire to adopt technology and incur additional marketing expenses on these listings is for transaction efficiency (Ullah, Sepasgozar, and Wang 2018), higher commissions, and decreased days on the market (Benefield et al. 2019). However, it is not atypical for agents to minimize their marketing expenses (Clauretie and Daneshvary 2008; Gwin, Ong, and Gwin 2002) and push back on emergent technologies. Sellers’ agents commonly use various promotional strategies to help market a real estate property, including photographs and virtual-tours. In practice, virtual-tours are often video or photos ‘stitched’ together as a video without the ability for buyer control. In other cases, the virtual-tours being produced are property-dedicated websites that include photographs, videos, and links to local amenities. However, it is not atypical for agents to minimize their marketing expenses (Clauretie and Daneshvary 2008; Gwin, Ong, and Gwin 2002) and continue to push back on emergent technologies.

Since the early 2000s, consumers have shifted their information gathering from real estate agents to the internet; agents are no longer the individual resource to acquire ‘exclusive’ details on the listings (Sawyer, Wigand, and Crowston 2005). Structural reshaping included new organizations that recognized and created value propositions to meet the convenience goals of actors. Furthermore, brokerages shifted to meet new expectations, including hosting sites to list property details and adopting technology to automate and digitize transactions (Kerch 2000). While most actors found the novel technologies introduced at that time valuable for efficiency and improved communication, real estate agents feared their commissions would be eroded. As agents pushed back, connection failure was common. Fear of technology as an agential actor is evident in Century 21's 2017 advertisement: “There's no robot for insight or hustle or a handshake … Good luck, robots” (Frankel 2017). Nevertheless, real estate agents have not been eliminated from the service system. In fact, average commissions increased despite technology's move to democratize information – much to the surprise of practitioners (Frankel 2017). As agents began to connect with technologies, their value in the system increased, resulting in positive outcomes.

In today's digital environment, photographs are a common marketing application, and there is a benefit to providing not only more photographs (Allen et al. 2015; Benefield et al. 2011), but attention must be made to the quality of the photographs (Luchtenberg, Seiler, and Sun 2019). Videos, or ‘virtual-tours’, were an innovative medium used approximately 15 years ago and became more common due to their benefits to the agent (Benefield et al. 2019). Virtual-tours can increase consumers’ engagement with a property online (Ullah, Sepasgozar, and Wang 2018). Furthermore, research exploring videos during this time showed a positive relationship between their usage on sales (Allen et al. 2015; Benefield et al. 2019; Yu et al. 2020). Recent research finds that usage of VR-tours can relate to increased sales, especially for homes that are difficult to show (Anderson, Freybote, and Manis 2022).

Data Collection

Data were collected using process-based research techniques to acquire a comprehensive data set to support the interpretive analysis (Giesler and Thompson 2016). Pivotal to this data collection process is to curate “a rich, varied, and multi-layered” data set (Giesler and Thompson 2016, p.499), including data from multiple structural layers.

The first author also engaged in participant observation, attended professional events, visited VR-tours, and used VR-tour hardware to create content. The first author's participant observation included using the technology to market the sale of their home in addition to leveraging a 3D camera to create VR-tour content and using the technology regularly to tour homes through data collection stages. The participant observation and informal interviews with two agents, one service provider, and one technology provider ³ served as a reference point and gave insight into the development of the formal interview protocols.

We then conducted 36 in-depth semi-structured interviews using theoretical sampling to identify informants who had first-hand experiences with VR-tours during a home exchange. In-depth interviews lasted 30 to 60 min each, and informants (buyers, sellers, agents, service providers, and technology providers; see Supplemental Appendix A for details) were asked to reflect deeply on their experiences (Arsel 2017). We used purposive, theoretical sampling (Eisenhardt 1989) to identify informants who connected with VR-tours during a home exchange. Our purpose was to collect data from those who had first-hand experience with the technology. This is particularly important in an early adoption context where random sampling would not yield appropriate data for analysis. The subject pool was expanded using a snowball technique to identify additional informants exposed to this emergent technology (Spradley 1980). Many of these actors participated in multiple roles in the home exchange network (e.g., buyer and seller) and encompassed various geographic areas across the United States of America. As such, planned questions varied based on the informant's role(s) but always began with a grand tour question about their experience, then narrowed to individual experiences with different actors in the system. See Supplemental Appendix B for an example of the semi-structured protocol.

During the interview process, the first author was invited to and attended a Young Realtors Association event. There, the first author had the opportunity to informally interview approximately twenty agents and service providers to better understand their general interest in innovation in the residential real estate industry and discuss prior experience with VR-tours. Additional informal interviews with two agents outside of this event allowed us to understand why agents may not choose to engage in VR-tours for their marketing needs.

Archival materials were also collected from agents, photographers, and technology providers, as well as a review of the extant literature in this context, was layered into the data set, allowing a view into effects at the service system level. Micro- and meso-level data gave richer insight into modified expectations and changed practices over time (i.e., indicators of structural change). This data included marketing materials, case studies, and news periodicals, resulting in an additional 93-pages of double-spaced text. See Table 2 for the data collection synopsis.

OPEN IN VIEWER

Table 2.

Data Collection Synopsis.

Purpose	Method	Sample description
Foundational Understanding	Informal Interviews	2 agents, 1 service provider, 1 technology provider
	Participant Observation	Touring VR-tours Creating VR-tour content Marketing the sale of home
Primary Data Collection	In-Depth Interviews	13 consumers 15 agents/service providers 10 technology providers (449 pages)
	Participant Observation	Young Realtors Association event (∼20 informal interviews)
	Archival Documents	34 press/newspaper articles, case studies, and marketing materials (93 pages)
Check for Understanding	Informal Interviews	2 agents not using VR-tours
	Member Check	2 checks with agents 2 checks with service providers 2 checks with technology providers

Analysis

Our analysis of how technology achieves agency involves multiple levels of analysis. Using a process theorization approach, we first coded the interview data to follow the individual micro-level narratives (Giesler and Thompson 2016) as connections were made and severed by actors within the meso-level service system (Latour 2005). This analysis allows us to investigate network building and shaping. During abductive analysis, we coded our data for instances where connections occurred and categorized them across the four translation moments (Callon 1984). This revealed the emergence of value potential and whether value creation was achieved or if connections were abandoned. See Table 3 for additional representative quotes.

OPEN IN VIEWER

Table 3.

Additional Representative Quotes.

Value creation moment (Translation Moment)	Agency form	Representative quotes
Goal recognition (Problematization)	No agency	We went ahead and looked for a house and bought before we could sell because … we were really trying to look for a good area and we were just when we saw the right house, we wanted to snag it. (Janet, buyer/seller) At the time when we had a lot of relocation clients who were not able to make as many trips here … it was, let's come to Austin for the weekend. We don’t find anything. We’re gonna, look online and then try to make an offer from afar. (Emily, agent) A lot of companies have a reputation to uphold. It's a pride thing. They want to have that new offering because it's a much more highly competitive market. (Jacob, service provider) It's incredibly inefficient for both an agent and a home buyer to go look at 12 properties in person,” she said. “It can eat up nights and weekends for all parties. (Lauren Taylor quoted in a press release)
Value proposition (Intressement)	Potential agency	I also tried to show them, okay, so you’ve had 250 or 300 hits on your VR-tour. That's 250 people that didn’t make you leave your house for two hours while they showed it. So, all the looky-loos out there just wanting to pick through and tour are not really interested in buying. Yeah, they just did this instead. So that didn’t cost you time. (Ben, service provider, sharing his sales pitch) We thought it was very innovative, and my husband's all high tech, you know. And we do Facebook marketing and Facebook ads and we don’t buy print. I mean print is so dead for real estate. (Heather, agent/service provider) Realising a rising demand among agents in Sweden for new technologies to differentiate themselves and win business, the pair decided to change their strategy. They came across [VR-tour] in 2015 and recognized its game-changing technology would allow them to begin offering 3D immersive experiences to their clients. (case study materials) We have two kids in school and so it's not always convenient to take off and buy a plane ticket and go down there. So, this time we found someone who did the virtual reality house hunting, which made it much easier. (Laura, buyer/seller)
Connection failure (No Enrollment) No Connection	Unlikely agency	In terms of utilizing VR, I do get a lot of ‘yeah, that's not going to work for me.’ Granted the average age of the [real estate agent] person telling me that is 43 over probably. Nine times out of 10 if I ask them, ‘Well, what's been your experience with it?’ ‘Well, I don’t use it. I just don’t see the value.’ So, educating somebody on why they don’t have any education on something they’ve never done, that conversation gets old. (Jacob, service provider) I need to keep a high touch relationship with my customers. I want to meet them on site. I want to be there the entire time. I want to handhold them through the process. (Christi, technology provider, sharing responses from agents) And so, on the downside, it can also make it where I don’t necessarily need to go see the house … So, it kinda can go as a pro and con as far as saving customer time. But you don’t necessarily get the customer into your house, which could be a deal-breaker too. (Janet, buyer/seller) I just wonder if it takes away from the buyer-agent's job … ‘Because we don’t need a buyer's agent, we’ll just go with Redfin once we look at the house through virtual reality.’ (Emma, buyer)
Value determination (Enrollment) Expected Unexpected	Causal agency Potential structural agency	I try to be a first adopter of technology, to at least try it when it comes out … When I started using it, it was several years ago when I was still selling full time. I had a property that I really wanted to showcase, and I wanted to win the listing. So, I knew that [VR-tours were] out there. I decided I’m going to try it, and I brought in a photographer. I told the seller about it, and the seller ended up listing with me. (Philip, broker/agent) A buyer is going to be able to look at it and said, yeah, this is valuable, hugely valuable to me, especially if either a), they don’t want to see it in person until they’ve really vetted it, or b) they’re just not able to, and they don’t want to just hop on a plane for a hope and a dream and a wish rather than seeing what they might be able to see. (Emily, agent) It's almost amazing to see, you know, the photos or the [VR-tour] and then go to the house. The [VR-tour] is quite honestly very realistic to what you actually see when you get there. Whereas the photos I think can be a little misleading because sometimes you can make the space look enormous and you go in there and you’re like, oh, it's really not as nice as it does in the photos. (Emily, agent) You can’t really experience it if you’re just hopping from pano to pano, you don’t get an emotional connection to the home, positive or negative. (Mitchell, technology provider)
Value cocreation (Mobilization) Adoption Diffusion	Structural agency	We were like, this is the house, like this is the one that we want. So not even a couple of hours later we had made an offer on this house because we just, we loved it so much through the [VR-tour]. (Mary, buyer) So normally what we would do … we would go see those two [homes] in person. But we felt so confident about our choice, seeing it on the video and walking around the neighborhood and doing the research online, that we actually bought it without seeing it. (Laura, buyer/seller) The ones that do it, they stay with it, and it converts faster by our metrics. (Jacob, service provider) We’re doing commercial spaces as well now … We’re doing restaurants … We’ve done a couple of card shops, novelty card shops, you know, knickknack shops, things like that, gift stores. And what we’re able to do with that as [VR-tour] now works with Google maps, so if you upload the [VR-tour], it’ll accept … them into Google maps and make them clickable. So like if you’re clicking on Google earth past this restaurant and you want to see what this restaurant looks like inside, you can click right into the restaurant and go tour the whole restaurant and then click back out and keep clicking down the street. (Ben, broker/agent/service provider) So when I am a homeowner and I put my place up on Airbnb or HomeAway or whatever, um, is what I have described actually the state of the place? Because if I’m dishonest, then some of that negativity comes back on Vrbo or Airbnb. Some of the questions are like, Hey, can you use [VR-tour] scans? There's a little bit of an audit, like truth detector, Hey, this place was described as a five bedroom, but the [VR-tour] of the place only detects three … so it's interesting cause that's [VR-tour] as a little bit of a ground truth is kind of an interesting thing that's bubbled in a couple areas. So like from an appraisal standpoint, from a, ‘hey, my house is really a 10,000 square feet and if we did a complete scan of it will go, oh, you know, he may be exaggerating maybe a tiny bit.’ So those kinds of things have come up, which are interesting because that goes a little bit beyond just space promotion. As in I see pretty things and I can walk through and I can experience this whole place. We do get tension, especially in the residential real estate world … If we take the approach of, ‘Hey we are ground truth’ and we’ll give a place a good light, but if the place is trash, we’re gonna show that it's trash, right? There's no editing. (Christi, technology provider)
Value destabilization (n/a)	Structural agency	Does it look like the picture or the [video] tour, or is it just a great photographer? … I can have the worst property, and the greatest photographer or videographer, and I can make a place look great. And it's not that great when they see it in person, and they’re seeing what it is. That's what photography can be. It can be that we can take the worst property and make it look great. (Diane, agent, questioning other media in light of VR-tours’ capabilities) Anyone who has ever bought a home can attest to the hours spent looking at homes online only to discover that they look drastically different from the marketing photos. (press article) I’m a little spoiled. Like when I went apartment shopping recently, finding an apartment size and trying to figure out, you know, like the layout and this and that and you have these crappy photos that are super either way too wide. So you see things like, that's not a sports stadium, that's a two foot by two foot closet . Or you have these teeny narrow things where all you can see is a single patch of carpet and that's supposed to be the master bedroom. It's actually really difficult to understand what it is you’re looking at without [VR-tour]. (Edward, technology provider)

Archival data was also coded to follow the four moments of translation (Callon 1984). Using guidelines established by Strauss and Corbin (1998), the data was also open-coded for additional relevant categories (Lincoln and Guba 1985) alongside this process. This analysis allowed us to investigate how agency emerged and valuation changed across the pre-identified stages of agency translation conceptualized in the framework (i.e., Figure 1). Throughout the process, the authors met frequently to resolve any coding discrepancies and reach a consensus on the categorizations. We also conducted member checks (Lincoln and Guba 1985) with two technology providers, two service providers, and two real estate agents, and they verified that our analysis is representative of their experience in the home exchange network.

Goal Recognition and Problematization

When motivations arise, actors seek ways to attain these goals by interacting with others (Findsrud, Tronvoll, and Edvardsson 2018). The first moment of translation is problematization, where actors are motivated by a goal to achieve, or a problem to solve (Callon 1984; Giesler and Thompson 2016). In our research context, two salient goals for actors in the network include convenience and differentiation. This drives actors to connect with other actors to help meet their goal(s). This could include connecting with new technology, like the VR-tour, whose goal is usage and diffusion (as embedded by the developers).

In a typical exchange, consumer actors will engage with additional actors, specifically buyer- and seller-agents, to attain their goals and make their exchange experience more efficient (Findsrud, Tronvoll, and Edvardsson 2018; Koskela-Huotari and Vargo 2016). All informants recognize a problem to solve beyond the basic need for home exchange. Aaron reflects on the common utilitarian goal in home exchange:

It's an inefficient use of time driving all over town, looking at different houses … I’m wasting my time driving to a property that, the second I walked in, I knew, okay, I don’t want this, and I just drove 30 min to get here … (Aaron, service provider, reflecting on his experience as a buyer)

It is common for prospective buyers to visit a home and quickly discover something they dislike, an inconvenience for everyone in the network: “When they walk in, and they see the house. And it's like, ‘oh no, there's not a bedroom downstairs, I’m gone.’ That's huge, probably hours of a seller's time …” (Emily, agent). The need for efficiency in the home exchange as well as firm-actors’ need to build or sustain their businesses (i.e., convenience and differentiation goals, respectively) are two salient goals that can drive connections with our focal actor, the VR-tour. In pursuing traditional goals (problematization), it becomes challenging for some actors in a service system, such as home exchange, to welcome innovative technology. While the technology has no agency at this point, actors’ goals of efficiency and differentiation create an opening for the technology.

Value Proposition and Interessement

Following goal recognition, actors begin to identify potential actors who might participate in the network. Through interessement, new actors are introduced, assume roles, and attempt to take action (Callon 1984). Technology begins to interact with other actors and offers a new way for all actors to participate in goal attainment by proposing expected value in the exchange. VR-tours make connections with agents seeking convenience and differentiation as the VR-tour is introduced, allowing actors to virtually “walk through” the home without being physically present. Nonetheless, some actors are reluctant to connect to the new actor, leading to connection failure.

Our informants identify VR-tours as a new, innovative technology used in digital marketing strategies that allows a home to connect to more buyers, offering convenience and differentiation. As such, the VR-tour's role is defined to help solve actors’ objective of a convenient home exchange, in addition to agents’ objective of differentiation.

A Virtual Open House Saves Time and Energy: They pre-qualify buyers, so sellers don’t have to open up their houses to everyone (marketing materials)

At this point, the technology has potential agency by offering a value proposition of convenience, driving a functional, utilitarian value for the consumer and most actors in the system. As a result, some agents connect with the new technology to offer value potential in their system. Meanwhile, other agents have concerns about the introduction and do not connect with the technology.

Possible Connection Failure: Connection Not Made

The “adoption of new technology is often fraught with pitfalls” and does not always lead to value creation (Hilken et al. 2017, p.900). New technologies do not always make connections to other actors in the network. One cause is a lack of familiarity within the system:

The market doesn’t understand what it is … In the early days, it was very new, and a part of the problem was a lot of people had pinned the phrase virtual tour as tied to basically a slideshow of images … [but it] is truly a virtual tour. You are able to actually navigate and interact with where you want to go, what you want to see when you stop … So, for me the term virtual tour is not used properly … (Aaron, service provider)

The most substantial reason that connections failed was due to concerns of the real estate agents. Agents hesitant to connect with the technology are concerned they will lose other emotional connections.

‘I don’t want to do [a VR-tour], because then my open house attendance will go down. I want people coming through.’ It's a bit of a reliance on an old-line method and just an emotional connection to that open house. Even though we all know open houses don’t sell homes … All that open houses do is help the agents find more customers, or they meet more people and they get a deal. They get a lead out of it, and someone wants to list their home with them. (Mitchell, technology provider)

Here, the technology is unable to make connections because other actors see it as a threat (Law 1993). Agents are concerned that the VR-tour will “make actors do things” (Latour 2005, p.160), severing emotional connections and reducing enrollment with other actors. This would inhibit agents’ positional power (Williams, Davey, and Johnstone 2021) and their ability to make their own network connections (i.e., consumers at open houses), ultimately destabilizing their current value cocreation system. Furthermore, agents not using the technology reported that VR-tour technology is not necessary and that their current marketing practices are good enough. Furthermore, adopting VR-tours would require additional efforts to support this technology, such as role expansion or hiring – something additional they have to monitor or manage. In other words, agents are uncertain and have concerns VR-tours may gain structural agency and cause connections within the system to be reformulated.

Value Determination and Enrollment

Through enrollment, the technology establishes its role in the home exchange network and begins to take actions that are verified by other actors (Callon 1984). Now the system is beginning to change as connections are established in search of value. As value is determined, it can emerge in expected forms (e.g., convenience and differentiation) and unexpected forms. At this point, the new technology achieves causal agency and potential structural agency. Nevertheless, the data reveals shadows of connection failure when connections are severed.

Value Cocreation and Mobilization

Resulting from value determinations, value cocreation is possible due to mobilization via adoption and diffusion. As more connections are made, “the more [the network] exists” and stabilizes (Latour 2005, p.217). Here, the technology attains structural agency. As noted above, once value determination is successful, actors may determine to continue using the technology on an ongoing basis. The technology became an active player in value cocreation through their adoption. As a reminder, ANT recognizes that this reliance and the new connections that facilitate mobilization indicate that the technology moves beyond simply being a facilitator or intermediary status into a mediator and gains structural agency. As such, the VR-tour engages in value cocreation. Agents, such as Philip, Charlie, and Ben, use VR-tour technology in the majority of their listings, as the technology facilitated beneficial results.

Value Destabilization

Our data allowed us to follow the system reshaping, or value destabilization, that emerged from these unexpected value determinations. Similar to the reliance on technology from expected value determinations, unexpected value determination provides additional stabilization and cocreation; however, it also delivers surprising opportunities. In this case, the data reveals that the entrant can transform valuations as it shifts and reshapes the service system. As VR-tours gain additional entry and support other actors’ expected goals, the new technology's structural agency causes changes in the network by renegotiating other connections in the system. These further assign new expectations and roles for the existing actors in the system, resulting in network disruption, or value destabilization.

Technology's Agential Role in Systems

By investigating how technology can attain agency within a service system, our research addresses calls for systems research (Wooliscroft 2021) using a powerful, qualitative methodology (Domegan et al. 2020). Recognizing that operant resources can have agency has substantial implications for the entire service system. Agency is historically conceptualized as a unidimensional concept encapsulated by the cognitive capabilities of humans (Zhao 2022), so a lack of research on technology's agency is understandable. However, an inability to acknowledge technology's agency leads to an incomplete understanding of how technology contributes to value creation and shapes marketing systems and society (Anderson and Thyroff 2022; Pohlmann and Kaartemo 2017; Sharma et al. 2020; Vargo 2018). This study has important implications for future research that explores technology within a marketing system to better recognize the “contemporary impacts” on marketing systems (Layton 2017, p.334) and how technology impacts value exchange and cultural values. While technology is a resource used in exchange, we find that its agency impacts consumer decision-making (i.e., causal agency) and influences future connections, expectations, and roles (i.e., structural agency).

Applying ANT to S-D logic was particularly useful as it allowed us to lay a foundation for understanding how technology, a non-human actor, acquires agency and influences a marketing system. Our evaluation of the technology recognizes the focal actor's function of a separate conduit, with an ability to hold causal and structural agency, enabling diffusion. Scaffolding from S-D logic literature that recognizes technology as an operant resource (Akaka and Vargo 2014; Vargo, Wieland, and Akaka 2015, 2020) and early conceptual recognition of AI technology's potential agency (e.g., cognitive assistants, Siddike, Hidaka, and Kohda 2021), we offer additional clarity to the deliberation of technology's agency (Sharma et al. 2020; Vargo 2018) and how it impacts changing values (Layton 2017). We illuminate the S-D logic perspective that agency determination is not constrained to conscious effort and decision-making (e.g., human agency) but is indicative of its impact on the actors and overarching structure.

By looking at the connections made and broken, we offer a more nuanced view of technology's agency in a service system. Following the micro-level VR-tour interactions provided insights into the specific differing value determinations versus value propositions (Akaka and Vargo 2014) that allow the technology to acquire agency. The data reveals that technology can hold causal and structural agency, impacting the marketing system and its cocreation efforts. As such, our research extends systems literature by detailing the role of non-human actors in a service system, called for in recent literature (Gummerus et al. 2021; Layton 2017; Sharma et al. 2020). We argue that not all actors that participate within a service system should be assumed to hold agency, nor should they be assumed to be an operant resource. Furthermore, human actors do not inherently have causal or structural agency within a specific service system. As noted by Kaartemo and Nyström (2021), nontraditional [human] actors can shape a market and agency is not always deliberate; peripheral actors may not create connections that lead to agency status. Our conceptualization takes Kaartemo and Nyström's (2021) work a step further. This study recognizes that technology can be more than just a market offering or catalyst by highlighting when and how technology's agency enables structural change. ANT's ontological interpretation and “methodological bracing” (Sayes 2014) helped us uncover not only that structure is affected but also that the VR-tour technology has the power to shift roles, practices, and valuation within the service system, with potential long-term implications.

Although this study concentrated on a single technology (i.e., VR-tours), insights may have transferability to other technologies, such as integrating smart devices into consumer homes. For instance, recognizing the entry for technology as utilitarian in nature (e.g., convenience) is still applicable to smart home speakers (e.g., Alexa, Google Home); however, the wider array of value forms that emerge leads to this technology's causal agency for consumers’ usage for information or entertainment. While this technology's causal agency is apparent by evaluating consumers’ dependency on their smart home speaker to receive alerts, listen to music, and even tell jokes. At a micro-level, value is transferred between the firm and client and is facilitated by the speaker. Zooming out, we can see that continued reliance leads to greater structural changes, at multiple meso-level systems, including the familial system of modified practices within the home, including rearranging their home, changing how shopping lists are transferred to other members of the home, and even communication across the home (often-times requiring additional speakers to be added). Expectations change within home builders now are pushed to ensure that homes are Wi-Fi and smart home enabled. Relatedly, new appliances are now expected to connect with not only a speaker but also a platform to allow for seamless integration and value cocreation within (and outside) the home. Similar mapping could be made for technologies such as ChatGPT and Mixed Reality. It may also help unpack when technologies have failed connections are not able to engage in value cocreation and its agency potential is stunted – not able to engage in value exchange at a larger scale due to lacking enrollment (e.g., Google Glass, Metaverse). Applying these concepts to map technology success turned failures could also shed new insights into the boundaries and flows of how a technology's agency changes over time (e.g., Blackberry, AoL).

Value Cocreation: Expected and Unexpected Value

This study contributes to our knowledge of value cocreation and transitions within a service system (e.g., Akaka, Vargo, and Schau 2015; Layton 2019; Vargo and Lusch 2016) and, most pertinently, how innovative technology influences value cocreation (Flint 2006; Sharma et al. 2020; Vargo 2018) and value destabilization. We focus on “value-formation systems” for goal attainment (Caridà, Edvardsson, and Colurcio 2019, p.75) and extend theorization of the resource integration process by evaluating the impact of a technology's expected and unexpected value determinations. Research has shown how consumer practices or the circulation of an object containing value potential can lead to value cocreation within a system (Akaka and Schau 2019; Figueiredo and Scaraboto 2016). Our analysis also reveals that although expected value potential can enable entry into a value cocreation system, unexpected value determinations can generate new expectations of others. The VR-tour was commonly offered to consumers at no cost by the real estate agent. The innovative technology is packaged as a gift of convenience to the consumers. This is not communicated as a suggested reciprocal exchange (Mauss 1925). The anticipation is that the technology offers and works with other actors to cocreate value shrouded in utilitarian benefits. Nevertheless, we find that value emerges and is adjoined with this non-human actor, which holds hedonic value and appeal. Specifically, enrolling VR-tours into their home exchange network can lead to co-constructive activity transforming into unexpected value expectations. Yet, these transformed values create benefits and tensions in the market. Moreover, the unequal distribution of these emergent value forms may enable power redistributions across the home exchange network, particularly impacting consumer actors. These insights affirm Layton's (2015) contention that evolving marketing systems will likely produce unpredictable outcomes.

Destabilization: Micro-Value Influences on System Valuation

The data provides insights into how technology evolution systems can mold “beliefs, behaviors, norms, values and social practices” (Layton 2015, p.303). Our review of a complex yet flat system reveals the influence of co-constructed experiences on the embedded and transformed levels of valuation through new technology diffusion. Specifically, the technology can participate in value cocreation at a micro-level while simultaneously influencing others in the network, generating value destabilization that affects the market-level evolution. As micro-level change occurs gradually through mediating actors, diffusion occurs, and roles are re-interpreted at the meso- and macro-levels, leading toward adaptive changes and self-organizing marketing systems (Layton 2015, 2019).

Our proposed conceptualization of value destabilization, which identifies the dynamic renegotiation of actors’ roles and power in a service system, is a likely outcome of new technology as it participates in value cocreation. Value cocreation is a dynamic process, and scholars acknowledge that interaction in service encounters can lead to value cocreation, value co-destruction (Keeling et al. 2020), or even market failure (Redmond 2018). Our results provide additional insight into the dynamic, alternative value cocreation process beyond the common focus of co-destruction of value (Echeverri and Skålén 2011; Plé and Cáceres 2010; Zainuddin, Dent, and Tam 2017). We suggest that when a new technology becomes relied upon, value destabilization can be expected. A risk in technology enrollment is the potential deskilling of other actors (Wirtz and Zeithaml 2018). While deskilling is a risk, we find that roles are renegotiated among actors and require new skills to sustain actors’ roles in the system. For instance, as ChatGPT and other generative-AI technology user growth climbs and becomes more relied upon (Hu 2023), academic systems shift dynamically. Some educators are pushing against the technology, and other educators and students are simultaneously learning with the tool. We anticipate that educator roles and practices will change unpredictably, impacting norms and societal values toward education.

Actors’ lack of adoption may also risk being perceived as ‘unethical,’ which creates a lack of trust and loyalty. Trust in others plays an essential role in marketing systems, and our findings support the work of Wooliscroft and Ganglmair-Wooliscroft (2018). When technology gains agency within the system, it garners this trust to facilitate exchange and value cocreation. Nevertheless, although VR-tours gain agency and trust, especially as they fulfill an unexpected ethical value form, it is important to recall the risks and, thus, responsibilities beholden on this new actor. For instance, as generative AI increases its capabilities, VR has the opportunity to be manipulated further, which may hold risk to the novel ethical value. Similar to what Vince communicated in his mismatched expectations (e.g., negative disconfirmation), this may lead to severed connections and an unraveling of the technology's agency in the meso-system over time. When a newly introduced technology is used unethically, it may lead to a severed connection and prohibit a technology from achieving agency (causal or structural). Furthermore, privacy is a common concern with new technology. With persistent storage of an authentic representation of the property, there are property rights and privacy policy implications. Consumers’ awareness of firms’ privacy practices can influence consumers’ value perceptions and decision-making processes (Hilken et al. 2017). Swaminathan et al. (2020) note the changing role of brands in recognizing their responsibility as data collectors. If an individual VR-tour is mobilized, gaining connections with others beyond the originally intended connections (e.g., using the VR-tour for home décor direct marketing), will this still offer value potential for consumers? Or, more importantly, can the prospect of reselling the content and data derivatives do harm? Similar implications and questions are likely with technology such as drones or smart speakers. Optimally, consumers need to be given control over their information in the marketing system (Watson and Wu 2022).