Introduction
Digital transformation in industrial environments is often framed as a technical endeavor – one that emphasizes data integration, system performance, and predictive capability. Yet in practice, such transformations are deeply entangled with organizational culture, communication structures, and the distribution of knowledge and expertise. This article explores what happens when digital system design is approached not only as a technical task but as a sociotechnical, participatory process.
The project discussed took place within one of Northern Europe’s largest forest industry companies. The organization approached us with a clear objective: to improve forecasting and productivity through the development of a new data platform that could provide real-time visibility into factory operations. At the outset, the leadership team described their challenge as “driving production while looking in the rearview mirror.” The solution they sought was, on the surface, technological: a system capable of transforming operational data a broad view of live factory operations that production line operators could take action on in real time
However, as designers and ethnographers working at the intersection of data, technology, and design, we recognized early on that the real challenge was not only about data pipelines or interface design. It was about ensuring that the new system would be grounded in the practical realities of factory work, and intelligible to those whose daily decisions would ultimately determine its success. To do so, we proposed an approach centered on embodied and collective intelligence – the knowledge that resides in hands-on practice, and the insight that emerges through collaboration across roles and contexts.
This article traces the trajectory of that project and the methods that shaped it. We demonstrate how ethnographic and participatory approaches helped reframe the problem, challenge narrow hierarchies of knowledge, and support the development of a more inclusive, context-aware solution. Beyond the technical system itself, we show how this process initiated a broader organizational shift – from siloed thinking to shared understanding, and from top-down control to distributed intelligence.
Certainly, ethnographic researchers have a robust history of arguing for the importance of participatory methods in creating viable digital solutions, in the pages of EPIC Proceedings (e.g. Sih, Carey, and Lin 2018; Lee 2017; Hasdell 2016) and beyond (e.g. Srinivasan 2017; Suchman 2006). This literature has emphasized situated, user- and stakeholder-informed approaches in shaping technologies and services that are culturally appropriate and contextually relevant, but our contribution builds on that literature by emphasizing how two core practices were integral to the success of those methods: facilitating constructive conflict and fostering inclusion through intentional structures.
Facilitating Constructive Conflict
Constructive conflict was the catalyst that made embodied and collective intelligences visible – and organizational transformation possible. In a context shaped by hierarchy and tradition, it was not alignment that sparked change, but tension: between top-down assumptions and the lived, sensory realities of factory work. Ethnographic and participatory methods provided the means to surface that tension – revealing overlooked insights, hands-on practices, and contradictions that had gone unchallenged. By facilitating discomfort in a safe, structured way, we created space for workers’ embodied knowledge to enter the conversation, and for collective intelligence to begin forming across organizational boundaries. Trust grew not by avoiding friction, but by navigating it with empathy, openness, and careful mediation. Through this process, participation evolved from an abstract ideal to a practical necessity – reshaping not just the system, but the organization’s understanding of where intelligence resides and how it can be activated.
Fostering Inclusion through Intentional Structures
Organizations cannot benefit from embodied and collective intelligences without actively fostering inclusion through intentional structures. In our case, this meant designing not just tools, but environments – both social and physical – where factory operators felt safe to speak, contribute, and shape outcomes. Trust was not assumed; it was built gradually through listening, transparency, and iteration. As psychological safety increased, hidden insights surfaced: overlooked challenges, material constraints, and unspoken frustrations that shaped how the system needed to work in practice. What began as participation evolved into shared authorship – transforming both the design process and the organization’s understanding of where intelligence resides and how it can be used.
Collective and Embodied Intelligences
As we set out to explore human-centric approaches to digital system design in an industrial context, we were motivated by a fundamental question: whose knowledge gets counted when decisions are made?
Despite the fact that the notion of intelligence as a fixed and individual characteristic has been contested (e.g., Rinaldi and Karmiloff-Smith 2017), organizational structures often continue to favor certain types of knowledge – such as technical expertise, leadership skills, or academic credentials – at the expense of other types of knowledge. This creates a hierarchy of whose voices are considered valid, and by extension, whose needs shape the design of tools and systems.
In hierarchical settings such as factories, where knowledge tends to flow top-down, this imbalance is especially stark (e.g. Robert & Schmenner 1978). Decision-making is often concentrated among those furthest removed from the day-to-day work, while workers on the factory floor – despite their proximity to processes and systems – are excluded from development conversations. This disconnect undermines not only usability and adoption but also the potential for innovation.
We believe that to design more inclusive, context-aware, and usable systems, organizations must draw from a broader range of intelligences than those traditionally privileged in top-down processes. Specifically, our methodological approach is grounded in two intersecting ideas: embodied intelligence and collective intelligence.
Embodied intelligence refers to the intuitive, sensory, and experiential knowledge developed through physical interaction with materials, environments, and tools. This form of intelligence is often difficult to articulate, yet it is fundamental to how people understand and navigate complex systems (Roberts 2020). For example, in industrial work, operators might detect a fault in a machine by the way it vibrates, smells, or sounds – insights that may not appear in data logs but are critical to smooth operation. This tacit, “felt” knowledge is often undervalued or excluded from formal design and decision-making processes, despite being essential for how work actually gets done.
Collective intelligence refers to the enhanced capacity of a group to solve problems, adapt, and make decisions through collaboration. As Pamela Hamilton (2024) suggests, groups can outperform even their most skilled individuals when they operate with inclusive practices, psychological safety, and constructive dialogue. In organizational contexts, collective intelligence emerges not only through meetings or teamwork, but through ongoing, distributed interactions among people with different roles, responsibilities, and forms of expertise. For collective intelligence to be effective, it requires constructive, healthy conflict where ideas and assumptions are actively challenged and built-upon, to generate better decisions and actions (Hamilton 2024).
In today’s digital environments, collective intelligence increasingly includes more-than-human actors – such as data systems, algorithms, and machines – which shape how information is produced and decisions are made (Leimester 2010). However, to truly benefit from this intelligence, organizations must intentionally design the conditions that support it: diversity of input, trust, mutual accountability, and structured collaboration.
By intentionally foregrounding embodied and collective intelligences, we aimed to challenge this imbalance. Rather than assuming which insights matter, we began with the premise that all organizational members possess forms of intelligence critical to the success of a digital system – and that these insights can only surface through inclusive, situated research practices.
Toward a Proactive, Data-Driven Factory Floor
The forest industry company’s objective was to implement a new software solution that would give production line operators a broader, real-time view of factory operations – an effort aligned with the larger Industry 4.0 movement toward smart, interconnected manufacturing (Lasi et al. 2014). Historically, operators had access only to their immediate workstations through localized control interfaces. Their awareness of upstream or downstream processes depended on indirect cues, such as fluctuations in material flow, or sporadic communication via phone calls or physical visits. However, the fast pace and fragmentation of factory work often discouraged such communication, reinforcing a narrow, isolated view of the production environment.
The planned software aimed to overcome this by providing a unified digital layer across production – an “informating” system in Zuboff’s (1988) terms. Rather than simply automating tasks, the new platform would render work processes visible, interpretable, and shareable. By turning actions and events into structured information, the system promised to support more anticipatory, data-informed decision-making on the shop floor. For the company, this promised not only gains in efficiency but a fundamental shift in how production knowledge was created and distributed amongst the employees.
Yet, while the technological capability to build such a platform was well established within our team, the more challenging – and ultimately more crucial – question concerned content and design: what exactly should operators see? What kind of data would genuinely support their decision-making without overwhelming them or inducing decision fatigue? These concerns resonate with Zuboff’s (1989) insights on the shift from sensory-based to abstract, screen-mediated work: workers accustomed to using sound, touch, and movement to interpret their environment may find it difficult to translate numeric readouts into meaningful action. Yet unlike Zuboff’s example of traditional vs. high-tech pulp mills, we expected the adoption of digital tools to not just bring disorientation and technical learning curves (ibid), but also deep cognitive and cultural shifts in how work is known and done.
Moreover, as Lee and Edmondson (2017) argue, digital transformation efforts require not just new tools, but more flexible, less hierarchical organizational structures capable of responding to change collaboratively. The implementation of new technology in such environments is not just a software deployment – it is a transformation in how authority, expertise, and communication flow. The introduction of digitally mediated work changes not only how tasks are performed, but also how workers relate to one another, to the system, and to their sense of agency within it.
Recognizing these dynamics early on, we understood that technical implementation alone would be insufficient. As ethnographers and designers, we realized that designing a system that would be adopted, understood, and trusted required tapping into the full spectrum of the organization’s intelligence – particularly the embodied and collective intelligence of the operators themselves.
Proposing Ethnographic Methods
To fulfill the project goals, tackle the issues of adaptation difficulties, and gather embodied and collective intelligence, we proposed employing ethnographic methods (see Roberts and Hoy 2015). From the outset, we anticipated tension between two knowledge systems: the formal, top-down expertise of managers and the embodied, hands-on intelligence of operators. Despite our conviction that the perspectives of factory operators were critical to the project’s success, the client organization initially did not share this view. From their perspective, the development of a data platform was primarily a technological undertaking – one that could be fulfilled by technical experts aligning their work with the strategic goals defined by upper management.
This viewpoint was rooted in the company’s long-standing culture where leadership practices had historically followed conventional industrial logic, where expertise was seen to reside with formally trained professionals. In Finland, the company operates across a wide urban–rural divide: while office staff are based in major cities and typically have strong educational backgrounds, factories are located in smaller towns where they may be the primary source of employment. These differences in geography, education, and daily realities contribute to silos between factory and office personnel, often leading to tension and limited mutual understanding.
However, the company was also beginning to acknowledge a recurring challenge: new technologies and systems were often met with resistance or failed to integrate effectively into day-to-day operations. This recognition created a window of opportunity – a willingness to explore new approaches that might bridge the gap between strategic intent and operational reality.
Through ongoing dialogue with the project team, we were able to shift their perspective. Drawing on our experience in both factory settings and digital development, we asked practical, grounded questions – such as whether operators wore gloves – that highlighted critical knowledge gaps. These questions revealed that technical success alone would not ensure adoption; the system also needed to align with the real, embodied practices of the factory floor.
From ‘Top-Down’ Insights to ‘Collective’ Co-Design
In the end, our project involved three main phases: ‘top-down’ workshops with central office personnel and some factory management, ‘ground-up’ ethnographic research with workers on the factory floor, and finally a series of ‘collective’ co-design sessions with factory workers’ perspectives at center stage. The organizational transformation gradually took place as we took steps forward to involve the users in the design and implementation of the software.
The project began with a series of ‘top-down’ workshops involving primarily central office personnel, along with two participants from factory shift management. These sessions aimed to build a shared understanding of the project’s goals, and to generate hypothetical solutions for improving the work environment and creating new operational systems and ways of working that we could build from our ethnographic fieldwork. However, once we transitioned to ethnographic fieldwork at the factory sites, it became apparent that many of the proposed solutions were not viable on the factory floor, because they failed to include the contextual, embodied intelligence of factory-floor employees or the material realities of the production lines. While we expected to iterate on hypothesized solutions from the workshops, we did not expect them to be quite as unfeasible as they turned out to be!
For instance, the hypotheses did not take into consideration the layout of the production lines, heat and noise levels at the factories, nor the material constraints that wood presents. For the operators, some ideas seemed too “far-fetched” because wood as a material requires close attention: it can easily dry out, break, clog production lines, or even cause fire hazards. This also affected what type of information is relevant for the operators, how that information should be made available, and how it should be communicated onward.
Our fieldwork lasted four weeks and included in-depth engagements with 36 production line operators across various roles, as well as sessions with managerial staff. Each engagement consisted of two parts: a private, quiet interview followed by on-site observation in real work situations. Due to high noise levels on the factory floor, the full interviews could not be conducted at the operators’ workstations. However, this constraint turned out to be a strength. The initial discussions provided a space for trust-building, allowing skeptical participants to become familiar with us and the project. By the time we transitioned to their work environments – their everyday life – many operators were noticeably more engaged and forthcoming, offering deep insight into both routine practices and systemic challenges.
The Learning Curve: Toward Collective Intelligence
Following the ethnographic fieldwork, the project team became convinced that the best results would come from actively involving factory workers in the design process. Their participation would not only ensure that the software reflected their actual needs, but also increase their commitment to its use – by encouraging reflection on their own practices and highlighting that good design requires effort and ownership.
Finding the right co-design workshop format required multiple iterations. Many of the participants had never taken part in a workshop before and were initially skeptical of its purpose or potential. Some operators were new to the project, while others were among those we had previously interviewed.
Eventually we arrived at a model where each session was structured around two parts: an open discussion phase (documented by facilitators) and a task-based phase, where participants worked in small groups to complete design worksheets. As the teams worked through their tasks, facilitators could return to earlier points raised in discussion to clarify or connect ideas – ensuring continuity and coherence between experience and design. Each workshop included two facilitators – one external and one internal factory representative – as well as members of the software development team, to ensure that needs were heard directly, without distortion. After the first design concepts were developed based on these workshops, we reconvened the same groups for follow-up sessions to collect additional feedback.
Key Findings and Lessons Learned
Lesson 1: Constructive Conflict as a Catalyst for Collective Intelligence
A central insight that emerged through this project is that organizational transformation rarely begins with consensus – it often starts with tension. In environments shaped by hierarchy and tradition, participation does not simply happen; it must be enabled, often by creating space for constructive conflict. As Hamilton (2024) argues, constructive conflict is not oppositional but generative – it creates the friction needed to challenge assumptions, surface contradictions, and push groups toward better decisions. In our case, the ability to navigate and facilitate this tension – between formal expertise and lived experience, between strategic vision and operational reality – was key to unlocking the collective intelligence of the organization. The process unfolded in phases, each building trust incrementally while allowing new ways of knowing to enter the conversation.
Creating Trust – and Entry to the Project – through Constructive Tension
By facilitating constructive conflict from the outset, we opened the space for alternative perspectives to enter the project. At the time of engagement, the client understood the challenge as a technical one, centered on building a real-time data platform. Ethnographic and participatory methods were unfamiliar – and initially seen as irrelevant. Instead of presenting these approaches as theoretical imperatives, we began by asking simple, grounded questions that the project team could not fully answer. For example: How would operators use the system if they are wearing gloves? How does heat or noise affect information flow?
These questions did not challenge authority directly, but they created productive discomfort by revealing key knowledge gaps. They introduced friction (Marini et al. 2023; Tsing 2005)– not as resistance, but as a necessary tension that highlighted the limitations of a purely top-down approach. This was our first act of constructive conflict: not confrontation, but a carefully designed inquiry that exposed the need for perspectives beyond the central team. Trust began to form as the team saw that our methods addressed not abstract risks, but very practical ones. This established a foundation of credibility and positioned ethnographic research not as a “soft” extra, but as a necessary component of a robust, user-driven solution.
Surfacing Contradictions by Listening to the Unheard
During interviews, constructive conflict emerged as operators voiced truths the organization had not previously heard – or had chosen not to confront. Factory operators entered the research process cautiously. Participating in interviews felt unfamiliar, and many expected their insights to be ignored. But as they were given space to speak freely, deeper issues surfaced – ranging from unspoken workarounds and process inefficiencies to emotional frustrations and exclusion from decision-making. Through deep listening, we did not just gather data – we allowed tensions to surface safely, revealing friction points between how the system was imagined and how work was actually done. What emerged was not just feedback, but a more complete picture of the organization’s internal contradictions. These moments of honest, grounded critique were not disruptive – they were generative openings that pointed toward better design and deeper change. Similarly, while facilitating the co-design workshops, we learned that the key to a successful session was allowing the group to first express their frustrations freely – something they were rarely invited to do at work. Only after this open dialogue were they ready to shift into more structured design tasks. By allowing the operators to share and identify production-related problems through open discussion, this increased their motivation and focus to create detailed solutions reflecting the needs of the factory floor.
Presenting the Findings: Transforming Conflict into Shared Understanding
When we brought the ethnographic findings back to the project team, the dissonance between original assumptions and field reality created a necessary moment of rupture. The solution hypotheses developed during the early workshop phase – mostly shaped by managerial perspectives – did not hold up under scrutiny from the field. Rather than treat this misalignment as a failure, we framed it as a critical learning moment. Using both the language of the factory and the frameworks familiar to leadership, we translated field insights in a way that was credible and accessible.
Importantly, we did not present the findings as complaints or opinions – but as systemic, material insights grounded in practice. This helped shift the team’s mindset from defensiveness to curiosity. The constructive friction between levels of the organization became a space for alignment, not avoidance. It was from this point that the organization began to see participation – especially from operators – not as a threat to control, but as a strategic necessity for designing systems that would actually work.
Lesson 2: Collective Intelligence Requires Creating Space for Inclusion
Creating Conditions for Collective Problem-Solving
Through our ethnographic and participatory approach, we were able to reveal systemic issues that sparked the need to actively include operators in the development work. This shift toward participatory development required more than a change in methods – it demanded a transformation in organizational culture. The right to speak, question, or suggest improvements was typically reserved for managers. This is why open-ended ideation felt unfamiliar, even threatening to many operators during their first co-design workshops. A critical success factor was the presence of both an external facilitator and an internal facilitator in the co-design workshops. The internal facilitator – someone deeply embedded in the everyday social fabric of the factory – was essential for creating trust, translating context-specific nuances, and bridging the gap between facilitation and shop-floor realities. The factory insider helped legitimize the process and bridge social dynamics, while the external facilitator was able to ask “stupid questions” that surfaced simple but essential insights. Over time, this structure created a safe environment where even the most skeptical participants began to contribute actively.
During our fieldwork, we also realized that physical space itself played a role in reinforcing the organization’s silos. Factory floors were designed for efficiency and noise control – not for collaboration. Loud machinery, isolated workstations, and lack of dedicated meeting spaces discouraged both formal dialogue and informal exchange. These environmental constraints mirrored the top-down nature of communication in the organization.
To fully support the co-design process, it became clear that the physical environment also needed to change. Collaborative work required dedicated spaces where operators, developers, and managers could meet on equal footing – spaces that signaled a break from routine and invited co-creation. Without such spaces, participatory practices risked remaining temporary interventions rather than becoming embedded habits. Since the quiet rooms within the factory were either too small or reserved for management use, we identified and set up a new, dedicated space for the co-design workshops. The room offered ample wall space for collaborative work and, importantly, was new to everyone – free from hierarchical associations or prior labels, such as being a “management space.”
New Roles and Shifting Hierarchies
As co-design practices matured, their value became increasingly evident. Factory workers – many of whom had never participated in development work – began to see the tangible impact of their contributions. Iterative feedback loops helped reinforce the idea that participation was not symbolic; it shaped the platform itself. This recognition fostered a sense of agency and ownership, transforming passive users into active co-creators (see also Cuciurean-Zapan and Hammel 2019). Operators moved from discouraged participants to active contributors, some even taking on new roles within the development process.
Indeed, this shift even opened new career pathways. Some operators moved into facilitator roles, helping lead future workshops by bridging the gap between shop floor experience and design dialogue. As the emergent platform allowed for low-code or no-code configuration, others took on responsibilities in the software development process, bringing their hands-on knowledge directly into the technical solution by helping build dashboards and adjusting logic.
This reconfiguration of roles signaled a larger cultural change. Intelligence was no longer assumed to sit at the top of the hierarchy. Instead, it was treated as something distributed, embodied, and emergent – drawn from many types of experience and expertise.
Organizational Transformation as a Shift in Ways of Knowing
One of the most profound realizations during the project was the extent to which factory operations relied on tacit, embodied knowledge and informal communication, which often concentrated among a few experienced operators, creating both fragility and bottlenecks. Existing digital systems had not displaced this human-driven logic. Moving from a “felt” way of working to a data-supported model was therefore not just a technical upgrade but a cultural shift – from intuitive to codified.
Recognizing that a sudden transition would be disruptive, we adopted an iterative approach that gave operators time and space to experiment, build trust, and influence the tools. This allowed collective intelligence to surface, clarifying which data supported decision-making and which created friction. This iterative approach laid the foundation for a model of continuous development. Workers began to see that they could influence the system – and that change was not only possible, but welcomed. Through this process, the co-design sessions became more than a means to an end; they initiated a shift in how change itself was understood within the factory.
Factory workers – once passive recipients of top-down decisions – became recognized as knowledge holders and co-creators. Their embodied, situated expertise informed design decisions. Their feedback looped directly into development cycles. And most significantly, their participation was formalized into an ongoing model of co-design, with regular workshops, open feedback channels, and a visible seat at the table.
This transformation also changed the mindset of leadership. Strategic planning began to integrate not only business logic and technical feasibility, but also frontline insight and long-term usability. The factory began to resemble not just a site of production, but a living knowledge ecosystem – where innovation emerged through dialogue, collaboration, and trust.
Conclusion
This case contributes to ongoing conversations about how knowledge is constructed, valued, and operationalized in organizational settings. Through a participatory process situated within a deeply hierarchical industrial environment, we demonstrate how ethnographic methods can make visible the embodied and collective intelligences that are often excluded from digital development work. Our approach revealed the complex and often contradictory dynamics between formal expertise and lived knowledge, between strategic planning and hands-on practice.
Importantly, this work underscores that participatory ethnography is not only a method of inquiry, but also a mode of intervention. As external researchers situated between management and workers, we were able to serve as mediators – not just amplifying unheard voices, but reconfiguring relationships of authority and agency in the design process. This positionality allowed us to challenge dominant epistemologies embedded in top-down technology development and to co-create new spaces of legitimacy for factory floor knowledge.
The project also invites reflection on the structural conditions that enable or inhibit collective intelligence. While ethnographers often highlight the richness of local knowledge, such insight remains inert without organizational mechanisms that can absorb and act upon it. Our case illustrates that designing for collective intelligence requires not only methodological openness but also infrastructural support – social, spatial, and cultural – for dialogue, dissent, and iteration to take root.
Finally, the emotional texture of our encounters – workers crying, venting, and ultimately engaging – reminded us that ethnographic work is always deeply affective. These moments of vulnerability and friction are not complimentary to the research process; they are central to the creation of trust, mutual recognition, and collective agency. This raises critical questions about how we as researchers show up, what we choose to represent, and how we negotiate our own roles as both analysts and change agents.
About the Authors
Iina Juurinen works at the intersection of data, AI, business, and end-user experience, combining a research background in science with over ten years of experience in business and technical roles across international organizations. Her work focuses on designing AI and data-driven solutions that are grounded in real-world contexts through participatory methods and ethnographic research. At Solita, Iina helps organizations shape AI strategies that reflect both technological possibilities and the lived realities of end users, emphasizing solutions that are socially informed, purposeful, and contextually relevant. Contact at iina.juurinen@solita.fi.
Julia Granroth is a Design Anthropologist dedicated to creating inclusive, human-centered solutions – both digital and non-digital. She has worked in design at Solita for three years, following earlier experience in academic and consumer research. Julia specializes in human insight, service design, user experience, and generative AI, with a particular interest in human–technology interaction. Using a qualitative and ethnographic approach, she supports clients in uncovering the root causes of challenges and co-creating meaningful, value-generating solutions. Driven by curiosity and a commitment to positive change, she helps organizations navigate evolving sociotechnical landscapes to build a better future. Contact at julia.granroth@solita.fi
Research Ethics
Throughout the research process, we complied with EU GDPR regulations. Participants gave informed consent for interviews and observations, and research goals were clearly communicated with them. We anonymized interview data to protect privacy and prevent possible harm. We also adhere to the AAA Statement on Ethics.