According to the dictionary definition, the word assistance comes from the Latin word assistere which means to stand by or take a stand near, and can mean both the act of assisting someone and that which is supplied during this act. Assistance is driven by a need, consists of aid that is provided by a provider to the recipient, with the aim of achieving a goal, and can be viewed as assistance generally when it brings about a positive change. Digital assistance is modeled after the work of human assistants, who are subordinate to their principals or experts and work closely with them, with a capacity for situational awareness and readiness that is necessary to preempt and support the needs of the principal. This is achieved via communication (i.e., delegation), observation (self-initiative) and action (executing tasks) in a joint work environment with formalized processes and conventions. In working together, assistants and principals use tools and resources that are jointly accessible and contribute to a shared understanding of a situation (Dhiman et al., 2022).

Following this model, in the computing domain an assistant typically consists of an application or an agent which provides interactive support to the user in a given activity context, either based on user delegation or agent initiative. The support can be cognitive or physical in nature, the former is the classic case for Human-Computer Interaction (HCI) [e.g., conversational agents or productivity assistants (Tur et al., 2010)], whereas the latter comprises cases of human-robot interaction (Guerin et al., 2015). The application is accessible to the user via an interface, observes the user and the state of the activity (in terms of the tools and the state of the object of the activity), and supports the user by (a) acting on the tool or the object, and/or (b) by furnishing information in particular modalities at the interface (via speech or visual modality), on the tool (haptic feedback or augmented tools), on the object (in case of augmented reality), or anchored to the work environment (in case of mixed reality). Prototypical use cases include supporting workers in manufacturing and assembly (Mark et al., 2021), logistics (Schwerdtfeger and Klinker, 2008) and maintenance (Dhiman and Rocker, 2019). Figure 1 showcases some of these use cases. As such, the design space of assistance is large, and applications are increasingly incorporating machine intelligence to enhance their abilities to observe the work environment, make decisions and communicate with the user, rooted in fields of research such as context recognition, image processing, natural language interaction, mixed initiative, planning and support (Dhiman et al., 2022). While the word “assistance system” is widely used to describe such applications that assist workers in the industrial domain, in this article we use the terms “assistant” or “assistance application” synonymously to aim for parity with the terms ‘virtual assistant' or ‘intelligent assistant' that are becoming increasingly ubiquitous.

But what constitutes good assistance? The question has conventionally been answered by focusing on three main aspects: the user's attitude (and acceptance), usability and efficiency, and characteristics of the assistant (such as its anthropomorphic nature, its look and feel, personalization, predictability, maintainability, sociability etc.) (Dhiman et al., 2022). Following this line of research, in the industrial domain, the primary goal of assistant development has been the enhancement of employee productivity and efficiency. Hence, measures such as completion time, error rate, and instruction recall are most commonly evaluated, compared against different mediums of instruction (e.g. paper-based instruction) or interaction (e.g. with gamification vs. without) (Mark et al., 2021).

However, in recent years, the research focus of HCI has shifted from usability to user experience, and even further onto user PWB, as advanced by recent movements in human computer interaction such as positive technology (Riva et al., 2012) and positive computing (Calvo and Peters, 2014). According to these movements, although concepts such as usability remain important, a sole focus on productivity and efficiency may be too narrow to comprehend how user experience influences users' PWB. In the industrial context, studies from the field show that, although assistance applications exhibit potential in increasing productivity, especially by reducing error rates, their “human-centricity” is questionable, especially as these systems in their current form have brought about no change in the assignment, organization, diversity or creativity of work, or engendered any opportunities for employee growth and exchange of knowledge (Klippert et al., 2018). Similarly, the authors in Aringer-Walch et al. (2018) report on an “immanent loss of competence and autonomy”, and Warnhoff and De Paiva Lareiro (2019) provide an account of how interacting with such systems can result in a devaluation of experiential and process knowledge by prompting operators to follow concrete work instructions. Further, Apt et al. (2018) warns that assistants which control the execution of tasks create work situations in which they transform human workers into function executing agents, and the resulting cognitive under use negatively impacts motivation, mental health, and long-term workability. Technological advancement has brought about positive changes, but at the same time, put employees in a precarious situation by increasing job demands and regressing workplace motivation and well-being (Parker and Grote, 2022). Hence, it is crucial to comprehend the driving factors behind employee PWB and incorporate them into design.

The achievement and maintenance of human PWB can be considered to be one of the central themes of human existence. It was a topic of extensive philosophical inquiry and debate in ancient times, and the discussion around what constitutes PWB, and how to achieve it, still continues to attract scientific interest and scrutiny.

According to Ryan and Deci (2001), our understanding of PWB can be traced back to two philosophical strands–the hedonic view, and the eudaimonic view. The word hedonia is related to the hedonist view of human happiness propagated by ancient Greek philosophers such as Aristippus, who argued that happiness be seen in terms of a maximization of pleasure. In well-being psychology, the hedonic side of PWB is captured by the concept of subjective well-being (SWB) (Diener et al., 1999) which measures three components - life satisfaction, the presence of positive mood, and the absence of negative mood. In contrast, the eudaimonic view of PWB is often attributed to Aristotelian ethics (Aristoteles, 1962), where eudaimonia is used to characterize a life lived in accordance with the “highest human good”, that is, a life that is purposeful and conforms to high moral standards. Well-being psychology borrows this term and uses it in a divergent manner (Fowers, 2016). In it, eudaimonia refers to a meaningful life characterized by personal growth, autonomy and a sense of vitality and thriving (Ryan et al., 2008; Ryan and Martela, 2016). According to Huta and Waterman (2014), eudaimonic well-being (EWB) converges around four key processes: excellence (striving for high standards and quality), growth (gaining knowledge, insight, and skill), meaning (experiencing significance and value), and authenticity (clarifying one's values).

Research shows that hedonic and eudaimonic well-being are somewhat related, but are still distinct components of PWB. Whereas the former refers to the affective experience, the other refers to cognitive evaluations of one's activities associated with what is worth doing (Waterman et al., 2008). Hence, not all pleasure inducing activities generate a sense of attaining valued objectives in life. For instance, when pleasure is pursued directly by engaging in activities that are instantly gratifying (e.g., consuming delicious food, watching a movie or going to the spa), the resulting change in positive affect only lasts for a short term, whereas pursuing activities that are aimed at achieving eudaimonic goals (e.g., helping someone, putting in extra effort in tasks, having meaningful discussions, taking time to introspect about one's values, organizing and cleaning one's surroundings) result in long term changes in positive affect (Huta and Ryan, 2010). Accordingly, several well-being researchers have argued for the superiority of the eudaimonic view. According to them, it is only when one is self-determined to pursue the ideals of meaning, growth, and excellence, that one achieves long-lasting PWB (Huta and Waterman, 2014).

Consequently, eudaimonia has been studied in psychology as different stages of human behavior, namely, orientations (priorities, motives or goals that are behind human activity), behaviors (specific activities that are pursued to fulfill these motives), experiences (state-level subjective feelings, emotions, appraisals that are indicative of how well an activity supports eudaimonia), and functioning (long-term trait abilities, achievements, habits, etc. that contribute to a general sense of PWB) (Huta and Waterman, 2014). Further, eudaimonic and hedonic orientations can also be differentiated on the basis of their focus of concern: me vs. we—eudaimonic goals reflect changes one wants to bring about in the self as well as beyond (self), whereas hedonic orientations are primarily focused on the self; now vs. future—eudaimonic orientations are about the future, whereas hedonic orientations relate to the present time perspective; tangible vs. broad implications—eudaimonic orientation focuses on how the concrete and tangible relates to the big picture, whereas hedonic orientation is only about the concrete and tangible (Pearce et al., 2021). Altogether, these dimensions of eudamonic well-being can be grouped morphologically (Figure 2).

How are the components of and stages in Figure 2 linked? According to Ryan and Martela (2016), the process of EWB be explained via SDT: when one pursues intrinsic individual and social goals rather than extrinsic goals (orientation), regulates one's behavior autonomously (activity or behavior) and leads a reflective life (evaluating one's experiences and how they relate to the other two antecedent components), one can be expected to satisfy the needs for competence, autonomy and relatedness, which results in positive experiences at the state level, and over time, stable states of “vitality and thriving” at the trait level (functioning). This process is captured by the “Eudaimonic Activity Model” (EAM) proposed by Martela and Sheldon (2019). The EAM emphasizes that eudaimonia is both “doing-well” and “feeling-well” (Figure 3).

The two conceptualizations of hedonic and eudaimonic well-being can also be found in studies of employee PWB, where the concept of job satisfaction has been termed a hedonic measure since it captures employees' pleasure or enjoyment derived from having or doing a job, i.e. subjective reports of positive affect (related to hedonia, i.e., feeling good, happy, and satisfied) (Rothausen et al., 2012; der Kinderen and Khapova, 2021). The eudaimonic perspective, on the other hand, examines the presence of factors that support outcomes related to one's skill development, growth, sense of contribution, personal relationships and the construction of life in a general sense (Rothausen et al., 2012; Budd and Spencer, 2015).

A number of such factors have been identified and are supported by extensive research. For example, the job characteristics model originally proposed five characteristics of satisfying work: autonomy, skill variety, task identity, task significance and task feedback (Hackman and Oldham, 1976). This was extended by Humphrey et al. (2007) further to also include social (e.g., social support, feedback from others) and work context characteristics. Research in the past few decades that highlights the role of the fulfillment of the psychological needs of competence, autonomy and relatedness, as proposed by self-determination theory (SDT) in supporting employee PWB (Broeck et al., 2010). More recently, der Kinderen and Khapova (2021) carried out a systematic literature review and identified the following drivers of EWB at the workplace: task/job significance, skill enhancement, learning new things, crafting feedback, opportunities to give/receive social support, utilization of strengths and skills, perceived control and autonomy and psychological safety. From an outcome perspective, they noted that fostering eudaimonia in work contexts positively affects several job and career-related outcomes, amongst them work-engagement and personal initiative (Hahn et al., 2012; Aiello and Tesi, 2017), willingness to invest time and energies beyond formal tasks (Turban and Yan, 2016; Sattar et al., 2017), contributing to organizational efficiency and performance (Turban and Yan, 2016), job satisfaction (Jones et al., 2015), and reduction of work stress and anxiety (Merrick et al., 2017).

Technological efforts to enhance employee well-being at the workplace mostly view the term “well-being” in terms of physical well-being, with efforts directed at developing digital applications to persuade users to maintain physical fitness and physical health in sedentary or industrial work environments (Orji and Moffatt, 2018; Huang et al., 2019; Damen et al., 2020; Heikkilä et al., 2021). The problem of supporting workplace PWB via technology, however, remains largely unexplored. Although the hedonic and eudaimonic conceptualizations of PWB at work have been studied extensively in the domains of organizational psychology, they are only beginning to be explored by the HCI community. Recent studies have shown that, contrary to the hedonic aspect, technology-induced eudaimonic experiences are associated with long-term importance, need-fulfillment, positive affect, and meaningfulness (Mekler and Hornbæk, 2016; Laschke et al., 2020). Research shows that eudaimonia can be fostered by modifying non-technical aspects of work through interventions that consist of eudaimonic components, e.g., strengths based interventions (Oades et al., 2017) or job-crafting (Wrzesniewski and Dutton, 2001; Demerouti, 2014). As the ongoing discussion in well-being psychology further clarifies the notion of eudaimonia and its distinction from hedonia, it also opens up room for ideation on how to operationalize HCI techniques to practically apply the lessons of EWB in assistance applications at work. To this end, we carried out a literature review to synthesize design streams, strategies, methods and frameworks, with the aim of identifying techniques that could aid designers in the future.

Our strategy for literature identification was threefold. First, we chose Web of Science (WoS) as our primary database because of its broad coverage of qualified articles. In addition to this, we also selected databases from major publishers: Elsevier (ScienceDirect), Tayler & Francis, and Sage. These three were added to increase coverage. Second, the search string was designed by combining two sets of terms related to design framework, method, and tool (Set A in Table 1) and eudaimonic well-being (Set B in Table 1). Design scholars use disparate terms to describe something developed to help designers. Two of the authors, one from the design research and another from the HCI domain, collected typical terms for Set A from each research domain. For Set B, the synonyms of eudaimonic well-being were extracted from the previous review (Vittersø, 2016). Other often-used synonyms of well-being such as “physical health”, “quality of life”, and “welfare” were excluded to avoid papers that regard well-being as a physical health condition and socio-economic indicator. Third, types of articles were limited in original research papers written in English. According to the search strategy, literature was collected between 1st May and 10th May 2023. Our primary search using WoS resulted in identifying 1,214 records. The search string was restricted to Title, Keywords, and Abstract to find more relevant articles. Additionally, our complementary searches using the other databases allowed us to add 56 records. The total number of candidates became 1,262 after the removal of duplicates.

The screening process was separated to three steps. At first, the title and abstract of each record were checked to judge its relevance to our review. This step sifted out 1,198 records considered irrelevant. Then, the remaining 64 articles were retrieved and assessed for eligibility. We conducted a full-text review for each record and applied the following inclusion criteria:

As a result, 13 eligible articles were left. Finally, further articles were searched through cross-reference snowballing and Google Scholar to be more comprehensive. We decided to include 8 of 11 additional candidates. In total, we found 21 eligible articles for analysis and synthesis.

The qualitative analysis consisted of three steps. First, one of the authors read and initially coded the 21 articles. The coding scheme had two categories according to our research questions: what facets of well-being each article postulates and, how each design method supports designers. In the “what” category, we used the morphological classification of eudaimonia (Figure 2) as an analytical lens. To be more precise, three dimensions in Figure 2 gave sub-categories and codes to the “what” category: design objects (which stage it takes into account), outcomes (which element it is intended to achieve or increase), orientations (which orientation it guides users and/or designers to). The “how” category included two sub-categories, the domain and type of method, but detailed codes were incrementally built through the analysis. The latest coding scheme was determined by two of the authors through discussions, and finally the articles were fully coded.

In the second step, a chronological analysis of the literature was conducted to identify the research streams that have different concerns and aims of design. The result was visually mapped, incorporating the “milestone” articles that were not in our literature base but additionally identified as the theoretical sources of each research stream. Through this analysis, we revealed four research streams that had a different tendency in the “what” category.

The third step aimed to organize design methods proposed in the literature. To do so, we created a matrix referring to the four research streams and the types of method in the “how” category. Each design method proposed in the literature was mapped in the matrix. It provides an overview of the available toolkits for eudaimonic assistant design and potential areas for the further development of design methods. We discuss the results in the next section.

Figure 5 shows the chronological map of the literature, and Table 2 reports the result of coding for each article. This map provides us an overview of the research streams. The design approaches considering eudaimonic well-being were presented in early 2010s. There are two different theoretical roots that influenced the articles: self-determination theory in psychology (Ryan and Deci, 2000, 2017; Ryan et al., 2008) and the capability approach in economics (Sen, 1999; Nussbaum, 2003). In the downstream, recent advancements could be synthesized into four streams with different aims. In the following sections, we explain each research stream, distinguishing its different points of view and treatment of eudaimonic well-being in design.

The design approaches identified in the literature were mapped as Table 3. Through the analysis, we identified five types of approaches: methods/techniques, design spaces, design tools, frameworks and heuristics. The following sections summarize each type of existing design methods.

From a theoretical perspective, the field of design is rich with suggestions and frameworks for designing for PWB. As Tables 2, 3 show, design for PWB has been understood through various conceptual lenses, and among those frameworks adhering to the eudaimonic view, only certain aspects of eudaimonic categories and core elements are considered. Our classification of the various methods, tools and frameworks aims to complete the picture and while doing so, also highlights some inconsistencies. First, there is a considerable divergence in the level of abstraction of the design activity these tools and methods intend to support. In some cases, they intend to inspire (e.g., Desmet and Pohlmeyer, 2013; Dirin et al., 2022) while in others they offer procedural guidance (e.g. Szalma, 2014; Perrino and Burmester, 2020); whereas some offer methods and patterns to arrive at concrete design prototypes [e.g. (Bhattacharya et al., 2018; Klapperich et al., 2019; Hagan and Özenç, 2020; Villalobos-Zúñiga and Cherubini, 2020; Bentvelzen et al., 2022)], others present heuristics to evaluate the design effort (Peters, 2022). Second, we also see a difference in the sources from which these frameworks and methods draw their content from. While (Szalma, 2014) and (Peters et al., 2018) are based on theory (e.g., self-determination theory) and attempt to operationalize it, others propose mining situated human practice and experience to finalize design ideas for prospective users (e.g., Klapperich et al., 2019), and some reviews summarize concrete design features (e.g., Villalobos-Zúñiga and Cherubini, 2020; Bentvelzen et al., 2022). Finally, there is also differentiating focus on the temporal aspect of human behavior and experience. Some frameworks focus on envisioning a user's future-self (e.g., Lu and Roto, 2015), others focus more on shaping the interaction in the present (e.g., Peters et al., 2018) and some rely on comprehending and reflecting past user experiences and pivotal moments [e.g., (Bhattacharya et al., 2018; Klapperich et al., 2019; Perrino and Burmester, 2020; Bentvelzen et al., 2022)].

We presume that these distinctions are not completely arbitrary, but are contingent on the existence and significance of existing user experiences/activities, as well as the design goals. A theory-based perspective is helpful because it is based on justified, objective past knowledge, and can serve as a first recourse in novel situations or when behavior change is the aim. As Figure 3 suggests, eudaimonic well-being can be viewed as a process, beginning from specific motivations, goals to performing specific behaviors that lead to particular experiences and feelings. This can be clarified using a goal hierarchy (Höchli et al., 2018). Values are superordinate goals, they are identity based, entail an extended temporal perspective, and are linked to a broad scope of context. These superordinate goals are realized through intermediate goals that correspond to “a general course of action bound to a certain behavioral context” (Höchli et al., 2018). At the lowest level are subordinate goals that precisely define what and how to do something.

Analogously, one can initiate the design activity in a future oriented, top-down manner, (e.g., Hagan and Özenç, 2020; Kheirandish et al., 2020) and bridge the gap between superordinate and intermediate goals by suggesting how the structure of an activity (or behavior) corresponds to a superordinate goal (or value) in a manner that facilitates internalization and endorsement by the self. Moreover, design patterns and strategies (e.g., Hagan and Özenç, 2020) can here serve as concrete suggestions that could help users meet subordinate goals, prompt pivotal moments (Bhattacharya et al., 2018) or help resolve dilemmas resulting from conflicting intermediate and superordinate goals (Ozkaramanli et al., 2017). The second step in Figure 3 refers to mechanisms whose presence during the behavior is necessary to enable PWB. Here, it is implied that intermediate goals that are endorsed by the participant, and, in the pursuit of which (via subordinate goals) the participant experiences events that are supportive of autonomy, competence and relatedness, contribute to well-being. Viewed in the top-down manner, a designer can rely on suggested patterns (Villalobos-Zúñiga and Cherubini, 2020), and in a prospective sense, check that the design provides adequate support for autonomy, competence and relatedness in the process, the material aspects of interaction (e.g., Peters, 2022) as well as opportunities for reflection (Bentvelzen et al., 2022).

The second case relates to (re)creating experiences that assume an existing activity context (or behavior) that may already be a well-established source of well-being within which users can express their imagined and lived experience, that is, their envisaged, perceived, and reconstructed experience (future-present-past) (Doherty and Doherty, 2018; Perrino and Burmester, 2020). Here, in a bottom-up manner, one can mine for the presence of particular behaviors (intermediate goals) and materials supporting both need satisfaction and superordinate goals that could form the basis of future products and applications (e.g., Klapperich et al., 2019), or ask users to articulate their imagined, future experience (Perrino and Burmester, 2020). At the same time, users' remembrances of the past are often colored by cognitive biases and, based on the design goal, may or may not be applicable as predictors of future choices in a given situation (Doherty and Doherty, 2018).

In the context of our research focus, there are no specific design efforts or principles dictating the eudaimonic design of assistants or assistance applications. Rather, we think that a systematic mapping of existing approaches could provide various pathways to design.

As we mentioned previously, to enhance employee PWB, assistants ought to positively enrich an employee's possibilities of internalizing and achieving eudaimonic goals, and incorporate mechanisms that support them, among others, task/job significance (meaning), skill enhancement (excellence), learning new things (growth), crafting feedback (authenticity), opportunities to give/receive social support (meaning), utilization of strengths and skills (excellence), perceived control and autonomy (authenticity) and psychological safety. The choice of a particular design method, or framework, corresponds to the temporal perspective and level of goal hierarchy that the design process targets. To aid the choice of design methods, we have synthesized these facets in Figure 6. The nodes of the diagram represent the core elements of eudaimonia (orientation, behavior, and experience), and the actors (self and others). The edges denote the relationships between entities that can be supported via assistance. The role of the assistant can begin with helping the employee understand how a particular eudaimonic orientation (excellence, growth, meaning and authenticity) involves the self, how it contributes to the future-oriented self, whether it also involves others, and if yes, whom (Edges 1 and 2). For example, when it comes to work, excellence can be considered in terms of meeting standards of quality via one's skill, which involves the employee, but could also include it support from other, more experienced colleagues. Next, the assistance focuses on Edge 3 to support the employee to understand how a particular eudaimonic orientation can be achieved in an activity by setting intermediate goals (e.g. how is quality defined in terms of concrete standards of work tasks and how should the employee's skills be supported). Finally, the concrete activity of doing the task, the entities involved in a particular activity and the support one receives from the assistant (how), its structure and the events that originate in the form of interaction with the assistant and others are taken into consideration (Edges 4 & 5) in the form of measurable, subordinate goals and design of feedback. Proceeding further, the emphasis lies on supporting the employee to reflect and remember the experiences (Edge 6, here excellence, that is, the actual quality of work achieved) and provide ways for others (Edge 7, e.g. colleagues) to help the employee reflect and make connections between the felt experience and components of the activity and one's behavior (Edge 8) to prompt changes to be initiated, e.g. re-defining goals, clarifying problems or customizing assistant features.

Table 4 consolidates the different themes each edge represents, and proposes design tools, patterns and heuristics that could be of use. Some methods or approaches, such as psychological needs, are arguably applicable in all situations, others could be used in a particular phase.

The present article makes several contributions to the research communities of design and HCI. Although scholars have pointed out the importance of fostering workplace psychological well-being, no reviews until now have considered bringing together design methods, tools, patterns or frameworks to do so. Our first contribution is highlighting the implications of advancing the design of workplace assistance toward PWB and offering a consolidated set of options to aid developers and designers. Researchers could use the guiding questions and list of applicable design tools to inform their research efforts. The methods proposed here are not specific to designing assistance or to be used only in the workplace context, but could be extended to other design domains such as mobile apps, games, and human-robot interaction.

Second, it is the first paper that has gathered the diverse definitions and viewpoints through which design for PWB (and, EWB) has been proposed and/or achieved, revealing four research streams: design for positive experiences and emotions, design for autonomous motivation and engagement, design for positive activities and behavior change, and design for shared values and capabilities. This classification contributes to design research by identifying different design objects, outcomes, and orientations within the same overarching field of design. It would be helpful to theorize design for PWB, and useful for scholars to clarify their position within the research streams.

While we have striven to carry out an extensive review with utmost care, we recognize that our review has some limitations. First, we took a deliberate but informed decision to focus on psychological well-being. While this provided us the opportunity to deeply study the design approaches in this area, we had to exclude reviews on physical wellness such as those by Orji and Moffatt (2018), Huang et al. (2019) and (Damen et al., 2020). Second, our review primarily focuses on promoting psychological well-being through the design of interactive artifacts. There are other reviews that summarize well established therapies, trainings and interventions that may be delivered digitally for enhancing workplace mental well-being (e.g Bartlett et al., 2019; Daniels et al., 2021), but which do not take up design as their object. A different review could address the possibility of identifying and organizing such methods to draw potential candidates for design. Finally, although activities form the core of mental well-being, our review does not engage with other theoretical frameworks such as activity theory (Kaptelinin and Nardi, 2009) or how domains and activities in a workplace context could be analyzed to find opportunities for enhancing psychological well-being. This could be the focus of future work, and we discuss some prospects next.

Along with the proposed design methods and approaches, there are several other areas which could be investigated further.

Other sources of design inspiration: An additional source of design inspiration, and one, which in our view, harmonizes well with the design process, especially in the eudaimonic sense, is that of psychological interventions, or activities that are deliberately designed to foster PWB (Lyubomirsky and Layous, 2013). We see these interventions as a part of a continuum of PWB related concepts, starting from the four core, but rather abstract themes of eudaimonia, followed by workplace drivers that have been demonstrated to enhance employee well-being, to concrete prescriptive interventions that could be implemented. The intervention approach offers actionable suggestions, since technology adoption at work occurs in the backdrop of existing work practices, which are already sources of workplace joy or frustration. Interventions are designed to bring about change in people's behaviors or emotional states in everyday life, and since assistants are intimately tied to work practices, they can serve as the technological mechanism to introduce them in the workplace. Furthermore, interventions offer us concrete, prescriptive methods rather than abstract concepts such as psychological need fulfillment. The framework proposed by Wiese et al. (2020) already includes positive activities and behavior change interventions, but there are also sources of interventions especially tailored for the work domain, for instance the job demands and resources (JDR) based job-crafting interventions (Demerouti, 2014) and positive-psychology interventions for resilience (Seligman et al., 2005). These could be mined for potential design patterns and strategies.

Existing domain knowledge and organizational constraints: In industrial contexts, it is the system and process characteristics that delineate the problem space within which workers act. Furthermore, depending on the level of manual involvement, there may be differentiated levels of worker experience and know-how necessary to complete the tasks. How do we extract existing know-how that workers already posses and input it into the design process? And how can expert knowledge be used to design assistants to observe, evaluate and provide feedback to novice workers? Here, we propose that tools for activity analysis (Kaptelinin and Nardi, 2009) and in the organizational sector, participatory domain modeling and collaborative learning workshops such as Event Storming (Brandolini, 2017) and machine learning techniques are crucial to better understand and capture the context of work and recognize opportunities to provide assistance. Still, the room for improving worker well-being through technological interaction may be constrained by the overall structure of the worker's activities and their relationship to broader organizational contexts, a view which is echoed in the analysis of eudaimonia in work contexts (der Kinderen and Khapova, 2021). Work design as an area of research is already well established, and may be used as an investigative lens to distinguish those aspects of work that can be enhanced with, or risk regressing, via digital assistance and digitalization (Parker and Grote, 2022).

Evaluating eudaimonic experiences using digital applications at work: Efficiency and productivity is relatively straightforward to measure—as completion time, error rate, or recall rate in an experimental setting. Cooper (Fisher, 2014) provides an overview of instruments for measuring hedonic and eudaimonic well-being at work. Other measures include flow and interest (Bakker, 2008), work motivation (Gagné et al., 2015), scales for meaning of and in work (Both-Nwabuwe et al., 2017), psychological need satisfaction (Reis et al., 2000; Peters et al., 2018), meaning experience and elevation (Huta and Ryan, 2010), authentic pride (Tracy and Robins, 2007) and goal commitment (Seijts and Latham, 2000). At present, most of the discussion about user experience centers around technologies designed for leisure, and the work domain has largely been neglected (Bargas-Avila and Hornbæk, 2012). Research shows that positive experiences brought about by need fulfillment in work contexts differ from those in leisure contexts (Tuch et al., 2016). Investigating the hedonic and eudaimonic aspects of user experience at work would advance our knowledge in this domain.

Our hope is that the overview of design approaches and the synthesized toolkit presented here enables designers and developers to design assistants that not are only targeted at improving employee productivity, but also their psychological well-being.