Affect-Based Well-Being in Caring Practices with Companion Robots

Healthcare and Care in Service Research

Ongoing discussions in service research related to healthcare customers and related value co-creation processes (Danaher et al. 2024) aim to address questions about how to enhance patient engagement, redesign healthcare practices for greater collaboration, and develop supportive policies. Although a prominent stream of service literature often merges the notion of care provision with that of healthcare services (Berry, Yadav, and Hole 2024), clarifying the distinctions between the two is essential. Healthcare pertains specifically to medical interventions and clinical settings (Danaher et al. 2024); care encompasses broader activities, including empathetic and social support (Guillemot et al. 2022; Kipnis et al. 2022). The provision of assistance improves the quality of life of people who require help, whether due to age, disability, or vulnerability (Verleye and Holvoet 2024), by supporting tasks linked to hygiene, feeding, and mobility (Rosenbaum, Seger-Guttmann, and Giraldo 2017).

At the same time, defining care only according to tasks risks reducing it to a checklist of actions—what is done, rather than how it is done. Providers might see acts of care as “duties,” but patients place more value on the way those tasks are performed. Having one’s face washed gently instead of roughly conveys care and shapes consumers’ sense of being cared for (Kipnis et al. 2022). Such considerations are especially relevant for understanding the unique circumstances of consumers who experience vulnerabilities that influence their service experiences (Kabadayi et al. 2020). In turn, we propose that moving beyond a task-based view and embracing a stronger focus on practice theory and caring practices (Verleye and Holvoet 2024) can better define service provision that contributes to well-being beyond medical interventions (Skålén 2024).

Sociomaterial Caring Practices

Practice theory offers a conceptual bridge between individual agency and social structures, in that it positions practices as fundamental units of social life (Bourdieu 1977). Practice scholars investigate how practices are organized, maintained, and transformed, with particular attention to their interrelations and dynamics (Russo Spena, Mele, and Nuutinen 2017). In turn, practices constitute collective accomplishments, for which actors share understandings and engage in coordinated activities (Spaargaren, Weenink, and Lamers 2016). Practices are also deeply embedded in specific contexts and shaped by a constellation of social, cultural, material, and temporal conditions (Braidotti 2013). Illustrative insights in turn conceptualize care as a set of interconnected practices (Boykin 2001; Mol, Moser, and Pols 2010), framed not as series of discrete tasks but rather as a relational and situated process through which actors manage vulnerabilities while fostering resilience and capabilities (DeFalco 2020). This process demands the mobilization of knowledge, collaboration, tools, technologies, and discursive practices (Gherardi and Rodeschini 2016), which comprise three interdependent elements: materials (including objects and technologies), competencies (skills and knowledge), and meanings (symbolic significance) (Barad 2003; Orlikowski 2007). In turn, growing research emphasizes the role of materiality and illustrates how objects, technologies, and spaces become co-constituted through processes of agencement (Gherardi 2016).

Having introduced this practice-based perspective, we also specify two ways in which it diverges from a task-oriented perspective. First, it emphasizes the process inherent to lived experiences and real-time caregiver–recipient interactions, as when a caregiver adjusts the amount of pressure applied while washing a care recipient’s face (Twigg 2000). Therefore, care results not solely from task outcomes (clean face) but also from the experience it fosters. The “how” matters as much as the “what” because care is the ongoing activity of caring, not a static object or trait (Gherardi and Rodeschini 2016). This shift, from noun to verb, underscores the focus on processes over outcomes. Thoughtful caregiving preserves dignity and identity, which is vital for people experiencing vulnerability (Cruise and Lashewicz 2022); it also imposes substantial cognitive, physical, and resource demands (Rodriquez 2011). Second, a practice perspective treats materials as integral rather than incidental, in line with a sociomaterial view (Gherardi 2016; Orlikowski 2007). Especially as smart technologies and robots become embedded in care, this insight is increasingly important (Mele and Russo Spena 2025).

Social Robots and Human Dynamics

Service research that explores service providers’ adoption and management of robotic technology to enhance their service quality and customer satisfaction has identified some key considerations for understanding human–robot dynamics (Huang and Rust 2023), such as social cues, language comprehension, and nonverbal communication (van Doorn et al. 2017). It also cites the importance of trust, perceived usefulness, and ease of use for determining users’ acceptance of (Shanks et al. 2024) and engagement (Azer and Alexander 2024) with robots. Related research outlines pertinent roles for robots, as providers, facilitators, or companions, depending on their designed levels of autonomy and intelligence (Wirtz et al. 2018). Some studies predict that robots eventually will replace human actors (Xiao and Kumar 2021); others emphasize the promise of human–robot collaborations (Shanks et al. 2024), especially for social robots that engage with humans in socially meaningful ways (Azer and Alexander 2024; Čaić, Odekerken-Schröder, and Mahr 2018) and effectively enhance their interactions (Odekerken-Schröder et al. 2020). Yet such role-based framings risk reducing human–robot dynamics to simplified agency models, rather than acknowledging the emergent nature of these relations in practice. In care contexts especially, rigid categorizations seem less helpful than exploring the outcomes that may arise from human–robot agencement in robotcare.

Robotcare and Agencement

Sociology studies define care robots as those “designed and/or used for care” (DeFalco 2020, p. 32), which can encompass a wide range of services, from physical assistance to complex forms of support. Furthermore, robotcare spans broad social and cultural contexts, and it could have diverse, negative, or positive impacts on caregivers and recipients. From an ethical standpoint, concerns about the use of robots for care provision appear particularly acute in relation to care recipients with cognitive impairments; robotic caregivers arguably could undermine human dignity, especially when users are at risk of being infantilized (Sætra 2020) or deceived (Sharkey 2014). A person with dementia might engage with a robotic animal and come to perceive it as a living being, which raises ethical issues related to deception. Another argument raises the concern that the use of companion robots could displace meaningful human relationships or jeopardize users’ safety (Sætra 2020). In contrast, other scholars propose that robotcare can support and enhance users’ dignity by expanding their capabilities and fostering autonomy (Parviainen and Pirhonen 2017; Sharkey 2014). In this view, even deception can be ethically defensible if it contributes to emotional well-being and social connectedness (DeFalco and Dolezal 2023; Khaksar et al. 2016).

Notably, such ethical debates often depict robots as autonomous entities, even though they lack agency, decision-making, or moral responsibility (Parviainen and Pirhonen 2017). Rather than applying models of conventional interactions, we therefore propose an alternative perspective, rooted in a practice approach, that emphasizes human–robot entanglement (Mele and Russo Spena 2025) and predicts that caregiving outcomes emerge through co-shaped dynamics, rather than the actions of the robots alone. In line with practice-based studies that adopt a relational onto-epistemology, in which human and nonhuman entities are co-constitutive and continuously reshape one another in dynamic interdependence, we embrace the idea of intra-action (Barad 2003). In this view, agency only emerges through relations (Gherardi 2016). The French term agencement (Deleuze and Guattari 1987) captures such mutual co-constitution (Barad 2014). Rather than treating individual actors, technologies, and environments as separate, agencement emphasizes processes of becoming-with (Gherardi, 2016). Applied in service (Mele and Russo Spena 2025) and business (Ramaswamy and Ozcan 2018) research, agencement describes humans’, technologies’, and contexts’ value co-creation. We believe it also offers a valuable framework for understanding how care recipients engage with companion robots.

Companion Robots and Well-Being

Robots interact with humans by leveraging their cognitive, analytical, and social capabilities (Huang and Rust 2023; Wirtz et al. 2018). When used to establish companionship and intimacy, robots can facilitate social engagement, alleviate loneliness, and enhance people’s quality of life (Khaksar et al. 2016; Odekerken-Schröder et al. 2020). For the purposes of this study, we define companionship as “the combination of attachment, commitment, and intimacy” (Lee et al. 2017, p. 924) between parties, whereas intimacy refers to openness to sharing thoughts and feelings. In human–robot interactions, intimacy is co-constructed. Technological mimicry induces feelings, even if artificial or inauthentic, that foster attachment (Belk 2022). Such intimacy requires verbal (speech, personalized responses) and nonverbal (gestures, facial expressions) signals, which then determine the robot’s social presence and perceived connections (Chiang and Chou 2023).

Despite such possibilities, integrating robotic technologies into caregiving settings faces various challenges, including financial limitations, ethical tensions, institutional resistance, and concerns about diminished human interaction (Wirtz et al. 2018). As robots become more advanced though, they increasingly provoke nuanced emotional responses—such as joy, affection, and surprise—shaped by users’ behavior and the cultural context in which they are embedded (Filieri et al. 2022). Service scholars acknowledge these affective dimensions; they also consistently emphasize the irreplaceability of human touch. Robots’ inability to experience or convey genuine empathy limits their capacity to deliver the relational depth that caregiving seemingly demands and that may be essential for enhancing well-being (Kipnis et al. 2022).

Compounding this limitation, service research tends to conceptualize well-being according to utilitarian or hedonic frameworks, privileging subjective measures like satisfaction or happiness. Such models often overlook the affective, embodied, and relational qualities that characterize caregiving interactions (Čaić, Odekerken-Schröder, and Mahr 2018; Mele et al. 2022). Although some recent studies have begun to recognize the importance of intimacy and emotional closeness in improving well-being (Odekerken-Schröder et al. 2020), a more integrative approach remains lacking. This tension is especially salient in dementia care, where tactile interactions with robots can elicit strong affective responses. Embodied experiences may be central to well-being, yet they remain largely unexplored. The absence of research on how affect shapes outcomes in robotic care represents a significant omission, given that sensory states are fundamental to how recipients experience care.

Affect and Well-Being

Service research often uses emotion and affect interchangeably or inconsistently (Bagozzi, Mahesh, and Nyer 1999). For the purposes of this study, we focus explicitly on affect, which extends beyond emotions and can be conceived of “as an umbrella for a set of more specific mental processes, including emotions, moods, and (possibly) attitudes. Thus, affect might be considered a general category . . . [of] feeling processes rather than a particular psychological process” (Bagozzi, Mahesh, and Nyer 1999, p. 184). That is, affect is a dynamic process, not a fixed attribute, encompassing broad physical states (e.g., arousal, pleasure, and displeasure), and it is always directed toward something (Reckwitz 2012). It relates to bodily sensations and their influence on actions, interactions, and interpretations. These responses are both personal and collective, shaped by the materiality of the social context and expressed through shifting engagement with the surroundings (Gregg and Seigworth 2010).

The concept of affective technotouch in turn refers to how robots communicate or elicit feelings, which then shape experiences of care, vulnerability, and attachment, particularly when they enable tactile engagement (DeFalco and Dolezal 2023; Paterson 2023). Designed to stimulate embodied affect, robotcare allows users to both engage in touch and be touched, evoking multiple sensory and affective reactions (DeFalco 2020). For example, the pet robot Paro mimics real animal behaviors and can foster affectionate bonds, because users perceive a nurturing relationship with the robot (Ge and Schleimer 2023). This connection promotes belonging, comfort, and social attachment (DeFalco and Dolezal 2023; Paterson 2023). Considering the lack of research into the role of companion robots for the care of people with cognitive impairments, we seek to examine the implications of affect for well-being in care settings. By integrating an affect-oriented lens with practice theory, we conceive of a new way to examine the interplay of companion robots and patients with cognitive impairments during caring practices.

Caring Practices before Hiro's Introduction

In both nursing homes, caring practices adhered to structured routines, and caregivers dedicated significant effort to essential tasks such as monitoring hygiene, sleep patterns, and physical activity. Such tasks, combined with the need to manage behavioral symptoms (e.g., agitation, depression, and sense of isolation) among the residents, placed considerable strain on caregiving staff, who underscored the importance of effective communication and empathetic listening to forge sensory connections and trust with the residents. Yet their day-to-day duties, required to address immediate needs and manage episodes of agitation and nervousness, made meaningful relations hard to achieve. Both homes were committed to the principles of recipient-centered care, including facilitating leisure activities, but such principles demand substantial time and effort:

In our daily routines, we prioritize ensuring that residents remain calm, and their basic needs are met, including nourishment and cleanliness. Unfortunately, the additional workload, particularly given the residents' conditions, limits our capacity to engage in extended, enjoyable conversations or to entertain residents to the extent we desire. Despite these challenges, we make it a point to find opportunities for relation, such as brief chats while distributing snacks or during walks. (Caregiver No. 3)

Care scenarios also tend to be characterized by monotony while also requiring concerted efforts to mitigate residents’ agitation, depression, and isolation. The cumulative effect of routine care obligations and the struggle to alleviate related symptoms often meant that care recipients did not receive direct, engaging interventions from caregivers and instead spent long periods inactive—sitting quietly, exhibiting desolation, sadness, or nervousness (Online Figure A1), expressions that mirrored their unmet psychological needs and highlighted the imperative for more dynamic engagement strategies. Therapeutic interventions, such as doll or music therapy, had attained varying success, largely dependent on individual residents’ abilities and condition. Yet the passivity of these interventions diminished their appeal, with little impact on participants' sense of engagement or belonging to a community:

To keep the residents entertained, we initially utilized dolls; however, the contacts were short-lived, resulting in boredom. Additionally, the residents found the dolls cumbersome and unappealing due to their inert nature. (Nursing Home Manager No. 1)

To address these challenges, some proposed solutions experimented with an integrated approach, involving multidisciplinary teams such as entertainers and volunteers to create personalized care plans. However, recreational activities often followed a weekly schedule, leading to a repetitive and monotonous routine for both residents and caregivers:

Unfortunately, we can only offer recreational activities to residents a few times a week due to time and budget constraints. Our music therapy sessions happen just once a week, which isn't enough. (Psychologist No. 1).

Considering the pressing need for enhanced engagement, both nursing homes decided to introduce a companion robot into their care contexts.

New (Caring) Practices with Hiro

The Hiro sessions were the only practice in the care regime of each nursing home that changed as part of the action research. Caregivers experienced their own dynamics with Hiro, but for the purposes of this research, we focus on the human–robot relationship with care recipients. In both homes, caregivers presented Hiro to care recipients with empathetic, peer-like invitations that requested a role-reversal dynamic, such that usual recipients of care would give it. Caregivers introduced Hiro with phrases such as, “This is Hiro, would you help by looking after him?” or “I don’t have enough time to comfort my little one. Could you show some affection to Hiro?” Such invitations sparked a sense of purpose among residents, as evidenced by one’s immediate eagerness: “May I embrace the child?. . . His name is the same as my son's, Robi!” The resulting agencement among Hiro, care recipients, and caregivers unfolded according to the situated actions and touches and stimulated by Hiro’s richly modulated ability to vocalize. Its vocalizations depend on how it is managed by each resident, who can touch, hear, see, cuddle, speak to, carry, walk with, name, and dress Hiro.

The care recipients’ sensory engagement with Hiro (e.g., gentle caress vs. firm embrace) promoted a soothing, affirming embodied connection, in line with the robot’s intended design. That is, Hiro represents a shortening of Haptic Interface Robot, where haptic refers to a sense of touch. The touch-mediated intra-actions it supports are not unidirectional: The resident touches Hiro, and Hiro responds to such touch, such as by laughing. Auditory components complement these tactile elements. The sound of Hiro's movements, responses, and ambient noises became familiar over time, contributing to a soundscape that care recipients associated with companionship and attention. Cuddling and speaking with Hiro brought out their nurturing behaviors. In performing these situated actions, care recipients started to personalize Hiro with names and activities, like dressing it or drawing on its face, such that the humans and technology co-created each other in their entanglements. As one caregiver explained:

We avoid suggesting ideas about the robot identity to the residents, leaving them free to imagine and construct the identity they wish to assign to the robot, a doll, a child. In the same way, we help the residents to personalize the robot. We also allow them to continue relating with the robot, monitoring its movements, even when they stand up to rock it. We avoid taking the robot away without a reason. Often, we tell them: “thank you for taking care of it,” “now I'll take it to rest,” or “I'll leave you to have lunch, I'll take care of it, thank you for helping me.” (Caregiver No.2)

Each of these actions reflects a unique form of connection between care recipients and the robot, with caregivers, and with one another, which transformed caring practice regimes in both homes. The caregivers reported finding themselves at the center of a reinvigorated care context, with roles expanded beyond routine tasks, such that they introduced Hiro to care recipients and also engaged with them differently. Hiro's integration into daily activities allowed caregivers more time for deeper engagement and meaningful conversations:

The robot has facilitated relations between us and the residents. Our rapport has significantly flourished. By assuming the responsibility of caring for Hiro, care recipients remain more active and attentive, making it simpler for us to attend to and engage with them. (Caregiver No. 2)

The caregivers, care recipients, Hiro, and forms of talking and sensing together gave rise to a complex agencement in which all parties come into being through their relations and mutual influence. This agencement reveals a nuanced choreography of care, in which human actions and sensory experiences intertwine with robotic responses. The entanglement extends beyond functional connections; it invokes a more resonant connection. Caregivers facilitate the encounter in which Hiro and the care recipients discover modalities for relating and communicating through touch, hearing, and sight. Meanwhile, they gain opportunities to establish deeper affective relations with care recipients, through Hiro. Sensory engagements—touching, hearing, cuddling, and speaking—serve as conduits for affect, engendering a caring practice that supports task execution while also creating an affective, shared experience of becoming-with for care recipients and caregivers in a caring atmosphere. Formally, we propose:

Psychological Dimension

The way care recipients, caregivers, Hiro, and their verbal and nonverbal communication connect within the agencement implies a psychological dimension of affect-based well-being. It fosters sensory intimacy by creating relational dynamics, together with a sense of pleasure and comfort, even when traditional cognitive functions and verbal communication are impaired. Hiro directly influences how care recipients experience and express their affective states and relational presence, underscoring its transformative potential. By engaging with Hiro, care recipients rediscover a range of beneficial embodied sensations, from the simple joy of touch to more complex feelings of intimacy. Take Antoinette’s story:

Antoinette has always been a very active woman. She managed a pizzeria with her husband, worked in a senior living facility, and once retired, she took care of her grandchildren. Due to dementia, she is introverted and shy, but her speech is communicative despite her simple and brief sentences. Her procedural memory in her daily life practices is decent; often, just showing her the correct gestures is enough to replicate them correctly. However, she displays significant busyness: she rhythmically repeats certain gestures, such as rubbing her hands and movements that resemble “cleaning” surfaces. Antoinette begins using her gestures to care for Hiro. She caresses it and speaks to it like a baby. Her previously mentioned busyness becomes purposeful and aimed at caring actions towards Hiro. Her facial expressions become more expressive, with smiles emerging more frequently. (Researcher’s diary comments)

Antoinette’s changed gestures suggest a sense of pleasure and comfort from embodied engagement with Hiro, illustrating how sensory experience can enhance mood and happiness. Two other stories reveal how agencement fostered behaviors and thought patterns conducive to psychological well-being, including diminished apathy and enhanced mental health:

In addition, Hiro’s numerous and appropriate, responsive vocalizations richly mimic a baby. Those sounds tap into care recipients’ profound human experiences of caretaking and parenthood. Many treat the robot as if it were a real baby. For people with dementia, it represents a powerful bridge to their past, revitalizing memories and feelings that affirm their identity and enrich their present life. Participating in care for Hiro somewhat restores the care recipients’ social roles as (grand)parents, such that Hiro gives them a sense of purpose that combats feelings of apathy and boosts affective resilience. In this sense, the robot represents a source of comfort and security, reminiscent of the bond between a relative and a child:

The positive impact on Antoinette's self-esteem, stemming from her engagement with Hiro, contributes to self-acceptance. By valuing her contributions and reconnecting with a cherished aspect of her identity, Antoinette likely experiences an enhanced sense of self-worth and acceptance. This acceptance is a foundational element of psychological well-being, as it entails recognizing and appreciating oneself, including one's strengths and limitations. Such engagement is crucial for individuals, especially in care settings, as it fosters a sense of identity and belonging. (Psychologist No. 2)

Caring implies identity as a value, suggesting the importance of these observations about Antoinette and identity in general. In dynamic agencement processes, Antoinette reconnects “with a cherished aspect of her identity,” and the inferred improvements in her self-worth and acceptance would foster identity further, while the inferred gains in self-esteem and self-acceptance would promote the value of dignity, which care also aspires to achieve.

In summary, Hiro’s capacity to evoke intimacy and affect reduces distress, provides nonverbal reassurance, and contributes to the psychological well-being of care recipients (Online Figure A2a). For people with dementia, psychological well-being is not merely the absence of mental illness but a more complex and dynamic state of positive embodied engagement and sensory experience. Accordingly, we propose:

Physical Dimension

Regarding the physical dimension of affect-based well-being, agencement in robotcare contributes to improved physical outcomes by activating sensory responsiveness. Hiro provides stimulation and improves care recipients’ daily routines, which are crucial to facilitating embodied actions and reactions. Through bodily engagement and auditory feedback, Hiro enables them to activate various, often dormant senses. Such sensory responsiveness also contributes to care recipients’ affective contentment and cognitive alertness and prevents some physical decline associated with inactivity, as demonstrated in the case of Mario's participation:

Mario's apathy and cognitive decline are evident in his significant struggles with personal hygiene tasks. Additionally, when Mario becomes agitated, his impulses become uncontrollable, with gestures that hurt him. Through sessions with Hiro, Mario has shown remarkable moments of clarity and focus. This enhances collaboration in hygiene and motor exercises, previously met with greater resistance. For example, Mario was observed gently guiding Hiro through a series of motions (washing one's face and clapping hands) and mimicrying caregivers. This activity captivated his attention towards finalizing the task and elicited a more collaborative response. (Researcher’s diary)

Engagement directly contributes to care recipients’ sense of presence and vivacity, reinforcing their connection to their bodies and the context around them. Their bodies constitute vital communication channels and connections to the world, so physical activities remain critical for sensory experiencing, in which sense:

Generally, Hiro encourages small yet significant daily movements among all care receivers. For instance, several individuals find motivation to stand up and engage in ambulatory activities, such as walking around while singing lullabies to Hiro, treating it with the same care and affection one would show to a real infant. Those who use wheelchairs also participate by gently rocking Hiro, simulating the soothing motion one might use with a child. This engagement offers them a chance to engage in body movements. (Caregiver No. 3)

Caring for Hiro facilitated physical exercise by activating small, daily gestures and by channeling physical movement toward controlled activities. By focusing their energy and attention on caregiving, they engaged in more purposeful and controlled movements, like rocking or cuddling. The relationship enriched the residents’ bodily activity range and reduced the risks associated with behaviors such as wandering and falling:

Aimless wandering is a typical behavior among Alzheimer's residents, significantly increasing the risk of falling. The introduction of Hiro has marked a positive change in minimizing the risk of falls. By focusing their movements on Hiro, residents find a more meaningful outlet for their impulses, showcasing Hiro's effectiveness in enhancing safety and providing therapeutic engagement (Caregiver No. 5)

By stimulating physical movement, facilitating sensory experiences, and encouraging purposeful connection, Hiro helped enhance care recipients’ physical well-being. The positive changes observed in their mobility and engagement in physical tasks reflect the significant impacts of such integrated physical experiences. For care recipients with dementia, physical well-being is a multifaceted state that transcends mere physical fitness or the absence of disease, to include bodily activity (Online Figure A2b). We propose:

Social Dimension

Agencement in robotcare enables the social dimension of well-being, particularly for people with dementia, and enhances their ability to connect and engage within social networks. Even acknowledging the challenges posed by dementia (e.g., impaired speech and cognitive clarity), adding Hiro into care settings facilitated positive relational engagement through social connections and a heightened sense of community. Care recipients experienced a sense of bonding and active participation in their social contexts; in particular, noticeable improvements in care recipients’ moods promoted social exchanges:

Franco has always been a sociable person. His sociability has gradually diminished due to disorganized speech and cognitive decline. Within the department, his actions are sometimes misunderstood by other residents, and he too does not always correctly interpret the attitudes of others, tending to respond in an occasionally aggressive manner. We have included Franco in the Hiro project to try to promote better sociability and relationships with the rest of the group. When Franco feels happier and more content, he is more likely to engage in social activities, communicate effectively, and build stronger relationships with those around him. These improved relations contribute to a reinforcing cycle, where enhanced socialization leads to better moods, which in turn promote further social engagement. (Researchers’ diary)

Hiro actively engages care recipients in a unique dialogue that often serves as a catalyst for broader attachments. It is not uncommon to observe a resident seeking out the companionship of other care recipients, in spontaneous forms of relational engagement that encourage a shared social space for laughter and contacts. The robot nurtured a sense of relatedness among project participants, when then also emerged through nonverbal engagements. The affective resonance promoted through Hiro’s intervention strengthened relational bonds among residents:

Pierina entered the nursing home last January due to moderate cognitive impairment and depressive syndrome. She has a daughter who currently assists her from home and previously worked as a shopkeeper. She possesses a peaceful soul and, to an outside observer, appears vulnerable and defenceless, almost as if she were frightened by her surroundings and intimidated by the people around her. She appears shy and very attentive to personal care. From the first administrations of Hiro, Pierina showed a surprisingly positive reaction, appearing playful, amused, and at times even moved. The presence of Hiro, whom the care recipients named “Dani” (like her daughter) encouraged her to interact with the staff and even to play with other residents: on more than one occasion, Pierina and Annamaria were effectively playmates; the two elderly women shared some sessions and played together using two Hiro robots. Hiro stands as a social orchestrator, creating a context where socialization is not just encouraged but starts to naturally flourish. (Psychologist No. 1)

Due to its regular presence in the care context, the robot also offered a unique form of continuity for care recipients, that is, a steady reference point around which their daily activities can revolve. Human therapists may change, but Hiro remained a sort of anchor during daily routines, contributing to building a more trusting and comforting relationship between care recipients and caregivers:

Rosa entered Nursing Home 1 in August 2016 due to mild cognitive impairment. Formerly a seamstress and a mother of three, she exudes a sweet and caring demeanour, being sociable, calm, and kind-hearted, though occasionally intimidated by more aggressive care receivers. Whenever Rosa interacts with Hiro, she actively seeks out contact with the caregivers, eager to share anecdotes about her children, reminisce about past experiences, and even provide updates on their current lives. This desire for connection and storytelling, shared with both caregivers—whom she knows or has just met - and fellow care receivers, becomes a repetitive action that proves challenging for dementia residents to establish. For instance, Rosa's relations with caregivers often include repeated mentions of her fond memories, such as her trips to the beach at Bathhouse 33 during her youth. Her enthusiasm for these recollections not only sparks conversations but also fosters social connections among those present. (Researchers’ diary)

Thus, agencement in robotcare opens alternative avenues for social well-being. Care recipients engage in a form of socialization that does not rely solely on verbal communication but that instead features sensory bonding, which is equally powerful for maintaining a sense of relatedness and belonging (Online Figure A2c). Accordingly, we propose:

Dynamic Interplay of Affect-Based Well-Being Dimensions

The three dimensions of well-being do not stand alone but influence one another throughout the new caring practices. They also can create a virtuous cycle, manifest as a positive (self-reinforcing, not self-dampening) feedback loop of improvements. Agencement in robotcare serves as a catalyst.

Specifically, engagement with Hiro through tactile connections or guided activities stimulates sensory intimacy that bolsters a sense of pleasure and comfort (psychological well-being). The resulting affective regulation and cognitive stimulation also encourage care recipients’ sensory responsiveness, reducing their inactivity (physical well-being). Such sensory engagements alleviate feelings of depression and diminish residents' aggressive behaviors, underscoring a mind–body link, or symbiosis between psychological and physical dimensions. The sense of belonging fostered by Hiro also enhances sensory bonding and the social well-being of care recipients. We refer to social bonds as part of the social dimension in Proposition 4, but Antoinette’s psychologist also mentioned the relationship with her sense of identity, an aspect we might associate with the psychological dimension. Considering that care recipients with improved affective states appear more likely to participate in social activities and communicate more effectively, we note symbiosis between psychological and social dimensions, in a reciprocal relationship that showcases a clear pattern: Enhanced social connections and a reinforced sense of belonging contribute to psychological health, creating a virtuous cycle of well-being improvements. This virtuous cycle is further fuelled by the symbiotic relationship of the social and physical dimensions of well-being. Enhanced mobility achieved through Hiro enables residents to participate more actively in social relations and community activities, such as group exercises or chatting with other residents, especially in ways that reflect their experiences of (grand)parenthood. The stronger social bonds created among participants motivate them to increase their activity levels, highlighting the importance of the physical state for maintaining and enhancing social connections.

With Hiro, care receivers have undergone a profound change in their affective and psychological state, leading them to have a more serene attitude and decreased levels of agitation. This calmness fosters a favourable atmosphere, creating social relationships among us. As a result, the overall health status of the care receivers is improved and, for example, they accept physical activity with more ease and less drama. (Caregiver No. 10)

We use this dynamic interplay among psychological, physical, and social dimensions to conceptualize both the whole and the parts and thus to conceive of well-being as a holistic concept that cannot be simply or reductively compartmentalized but that does include discernible, workable parts. Each dimension influences and is influenced by the others. In our study, affect acts within an agencement to foster this interplay through intimacy, responsiveness, and bonding (as depicted in Online Figure A2a–2c). Formally, we propose:

That is, affect-based well-being, emerging within and expanding through agencement in robotcare, constitutes a dynamic cycle in which sensory experiences (sensory intimacy, sensory responsiveness, and sensory bonding) enhance psychological comfort, which stimulates physical activity. Increased mobility fosters social participation, and stronger social bonds reinforce affective balance, creating a self-sustaining loop of well-being improvements. This agencement rekindles identity, enhances self-worth, and promotes dignity, illustrating the transformative potential of embodied interactions in dementia care.