1. Introduction
The water environment in England and Wales is facing increasing pressures from climate change, pollution, and population growth (House of Commons, 2022). Currently, a significant proportion of English and Welsh rivers are failing to achieve good ecological status, which threatens biodiversity, human well-being, and water security (Environment Agency 2024; The Rivers Trust 2024; Surfers Against Sewage 2024). This is one of multiple complex and converging crises currently facing our planet, with several breached planetary boundaries linked to the health of water ecosystems (Richardson et al. 2023). However, this crisis goes beyond environmental degradation; it impacts economies, biodiversity, and human livelihoods (Anderson et al. 2019).
In our work, we have applied the concept of river and coastal health to describe the conditions of various waterbodies, covering lakes, rivers, estuaries, and the sea.
Our rivers and coasts, often referred to as blue commons, provide shared resources to all living beings and sustain life on a planetary scale (Macfarlane 2025; Rockström et al. 2009; Anderson et al. 2019). The current situation of poor river and coastal health reflects what can be described as a “tragedy of the commons”, characterized by over-use, under-maintenance, and degradation of common resources (Ostrom 2007, 2008).
Understanding river and coastal health is complex, with silos and varying definitions among stakeholders and disciplines. To water companies, farmers, and other industries, waterbodies are healthy when they hold plenty of water ready for abstraction (Karr 1999). To freshwater ecologists, a waterbody is healthy when it’s capable of sustaining aquatic biota and natural flow (Norris and Thoms 1999). To local communities, a waterbody is healthy when it’s aesthetically pleasing and doesn’t make you sick when you drink or swim in it (Pinto et al. 2012). To indigenous communities, waterbody health is viewed through a lens focused on spirituality and connectedness (Drew 2013; Kawe 2021). How do we navigate this complexity, making sure we don’t address the health of waterbodies from a narrow perspective, but do something that can improve it for all?
We believe that without integrating diverse perspectives on river and coastal health, any efforts to improve water ecosystems and overcome their challenges will continue to reinforce silos and a lack of shared ownership rather than supporting systemic change.
Addressing waterbody health requires a shift in how we approach environmental challenges, where we move away from silos and instead focus on collaboration and shared ownership to overcome the degradation of common resources (Ostrom 2007). We need to change how we perceive our relationship with each other and the rest of life on earth to enable all to flourish (Wahl 2019).
Through this case study we want to showcase how ethnographic and participatory practices, combined with the notion of collective intelligence, has been leveraged to navigate complexity and design novel artificial intelligence (AI), aimed at addressing river and coastal health in England and Wales.
2. River Deep Mountain AI
River Deep Mountain AI (RDMAI) is an innovation project funded by the Ofwat Innovation Fund aimed at enhancing the health of rivers and coastal waters by leveraging collaboration and artificial intelligence (AI). RDMAI is made up of a broad consortium of partners from inside and outside the water sector, including water companies, NGOs, charities, technologists, and ecologists.
RDMAI was awarded funding in 2024, with research beginning in February 2024. This case study details the research and collaborative work that took place during the discovery phase from February to October 2024. We explore how our approach has helped facilitate the emergence of collective intelligence, played a key role in designing solutions, and laid the foundation for the rest of the project, which runs until March 2026. Following the discovery phase, the project will go through two additional phases, focused on the development and validation of technical solutions.
From the beginning, RDMAI has been focused on the lack of shared ownership and fragmented solutions in river and coastal management. By bringing together a diverse set of stakeholders, internally and externally, RDMAI is collaboratively developing solutions aimed at addressing the multiple and complex challenges facing river and coastal health. By publishing all insights and AI models open source, RDMAI intends to democratize the use of AI across the water sector, encouraging collaboration and accelerating impact.
3. Approach
Over the course of 9 months, we carried out a series of ethnographic and participatory engagements with over 100 participants across multiple organizations and disciplines. Our approach combines in-depth interviews, observational fieldwork, and multiple co-creation sessions.
In this section, we go through the three pillars of our approach: 1) our systemic view, 2) ethnographic research, and 3) participatory design methodologies.
3.1. Systemic View: Collective Intelligence
River and coastal health is inherently complex, with conflicting perspectives and diverse stakeholders. Systemic design provides a framework for breaking down this complexity and integrating multiple different perspectives to design solutions that work for all (Jones and Van Ael 2022). Rather than zooming in on an individual stakeholder or a part of a larger system, systemic design encourages a holistic view, where multiple stakeholders, perspectives, and their interconnectedness are embedded into design solutions.
Drawing on systemic design and the concept of collective intelligence, described by Wahl (2019, [2016] 2022) and Atlee (2004, 2008), we believe that integrating multiple perspectives is essential to drive meaningful action and ensure interventions have an impact on complex systems.
“(…) multi-disciplinary collaboration in information sharing and sense-making can help us to co-create a world that works for all.” (Wahl [2016] 2022).
Collective intelligence emerges when stakeholders cooperate, have a shared vision, understand their collective purpose, and integrate multiple perspectives (Wahl [2016] 2022; Atlee 2004, 2008).
This systemic lens and the notion of collective intelligence have laid the foundation for how we have approached river and coastal health and guided how we have applied ethnographic and participatory practices to understand challenges and design solutions.
3.2. Ethnographic Research: Interviews and Fieldwork
It has been central to River Deep Mountain AI to get a deep understanding of different perceptions of river and coastal health, as well as explore the shared views, conflicts, and tensions that exist between stakeholders.
To achieve this, we conducted semi-structured interviews and observational fieldwork with a broad range of stakeholders and organizations involved in river and coastal health. This included local community groups, business owners, water companies, environmental NGOs, farmer associations, and regulators from England and Wales. Over the course of our research, we carried out more than 39 hours of interviews and 10 hours of observations, involving 67 participants from more than 30 organizations.
The interviews were conducted in two phases to progressively deepen our understanding of river and coastal health: The first phase explored the meaning of river and coastal health in detail, while the second phase focused on understanding ongoing initiatives and collaboration between stakeholders. The interviews were designed with semi-structured questions, keeping the tone conversational but allowing flexibility to explore emerging themes and insights. Most interviews were held online to accommodate geographical and time-zone differences between participants and interviewers. These sessions provided a rich understanding of different perspectives, reasonings, and challenges faced by individual organizations and stakeholders (Kvale and Brinkmann 2009).
In addition to our interviews, we carried out observational fieldwork to get first-hand and embodied knowledge of how river and coastal health is managed and experienced by different stakeholders (Hammersley and Atkinson 2019). This included visiting wastewater treatment plants, attending a local meet-up between farmers and water companies, meeting a local community group working to improve the health of their local bathing water, and spending the day with a catchment manager from an environmental NGO.
This has enabled us to understand the challenges of river and coastal health through complementary lenses of individual perceptions and embodied knowledge.
3.3. Participatory Design: Co-Creation And Collaborative Sense-making
Beyond understanding diverse perspectives and worldviews, an important part of designing solutions that work for all is to create a space for different perspectives to meet, make sense of the world, and co-create solutions (Wahl [2016] 2022).
Inspired by participatory design methodologies, we have hosted several co-creation sessions, inviting diverse stakeholders to collaboratively inform and shape the direction of RDMAI (Robertson and Simonsen 2012). Co-creation sessions have been held regularly during the discovery phase and have since become an integral part of how the project continues to deal with diverse perspectives in collaborative and creative ways. Sessions were often semi-structured, included open-ended and creative elements, and involved internal and external stakeholders, including water companies, regulators, ecologists, NGOs, and technologists. Creative elements, such as empathy maps, storyboarding, prototyping, and systems mapping were employed strategically during the sessions to encourage participants to collaboratively immerse themselves into the systemic challenges of river and coastal health. Sessions have been focused on two main topics: collaborative sense-making and exploring technological opportunity spaces. The latter has helped give all stakeholders a shared understanding of limitations and possibilities with novel technologies, such as artificial intelligence. Throughout the project, the size, frequency, and duration of these sessions have varied, ranging from a few participants for 2 hours to more than 70 participants for a four-day session.
Our combined approach has allowed us to get a deep understanding of the complexities at play in river and coastal health while at the same time supporting collaborative sense-making and co-creation between diverse stakeholders.
Together, this has enabled us to formulate and integrate diverse perspectives, and as we will argue later, facilitated the emergence of collective intelligence, which has been integral in shaping solutions that address the complex challenges of river and coastal health.
4. Reflections: Living the question together
In this section, we explore how our approach has helped uncover diverse perspectives, shared goals, systemic complexities, and establish a shared vision and trust between stakeholders. We reflect on how the ethnographic and participatory practices applied in this project have invited people to “live the questions together” (Wahl [2016] 2022) and, through that, facilitated collective sense-making and multi-disciplinary collaboration.
4.2. Uncovering diverse perspectives and complexities
Through our research, we have explored the broader stakeholder ecosystem involved in river and coastal health to understand the diverse perspectives and complexities at play. In the following sub-sections, we will expand on the insights gained through this exploration. Firstly, we outline four key stakeholder types. Secondly, we introduce three different perspectives on river and coastal health that emerged during our research.
4.2.1. Stakeholder types
Rivers and coastal areas are shared by all, which inherently means that many different stakeholders and users engage with them, are impacted by them, and play a central role in their wellbeing, which can lead to tensions. In the following, we outline four overarching stakeholder types that came out of our research: Industries, regulators, environmental NGOs, and local communities.
Industries cover stakeholders that primarily engage with waterbodies as part of their operations, relying on their resources to run their business. In our research, this type was primarily represented by water companies and farmers, but it also covers mining, pharmaceuticals, and other industries relying on water in their operations. For industries, a primary challenge lies in balancing operational needs with environmental impact, often constrained by regulatory frameworks. Farmers take water for irrigation of their fields, and at the same time, they are responsible for the pesticides, nutrients, or sediments being carried from their land back into the river during rainfall. Similarly, water companies abstract drinking water and distribute it to customers, while also bearing the responsibility of treating sewage before sending it back into the river or sea.
For this reason, industries are often under pressure from other stakeholders to improve or reduce their impact on river and coastal health. However, in many cases, their potential avenues of doing so are restricted by industry-specific regulations. This sometimes leads to industries feeling forced to prioritize siloed solutions that comply with regulations and reduce the risk of financial penalties, rather than pursuing broader positive impact.
Regulators are governmental and public bodies responsible for protecting and improving the water environment. Regulators conduct inspections, water quality monitoring, and provide financial incentives to regulate stakeholders and organizations. One of the biggest challenges faced by regulators is funding and resources. With focus and finances being split between national and local activities, resources to do local inspections, follow-ups, or thorough monitoring are often limited. Regulators often face criticism from industries claiming that their decisions to penalize are based on insufficient or incorrect data, or that funding schemes and other incentives are too difficult to manage. Simultaneously, NGOs argue that penalties are not harsh enough, that regulations favor businesses over the environment, and that regulations in general aren’t enforced.
Environmental NGOs are an essential and diverse part of the stakeholder ecosystem. They work with different environmental purposes in mind, such as improving the conditions for wildlife or waterbodies to thrive. Improving the ecological conditions is one of their main objectives, and they are often the voice of non-human actors such as birds, fish, and rivers. A key challenge they face is generating the necessary funding to do so. They rely on donations, public awareness, or partnerships to generate funding to do their work, and they often play a role in generating political and public pressure on industries and regulators.
Local communities are the residents and small businesses, living and operating in the area around rivers and coastal areas, directly impacted by the health of these waterbodies. Local rivers and coastal areas provide many benefits to local communities in the form of recreational, financial, spiritual, and cultural connections. In our research, we engaged with local community groups trying to improve the health of local waterbodies, but we found that they often struggle to understand their own impact. Regulatory monitoring is often technical and hard to interpret, making it difficult for local communities to understand if the health of their local rivers or coasts is improving and whether their work has had any impact. At the same time, the labeling of a local waterbody as unhealthy, either by regulators, NGOs, or media, has a direct impact on communities’ self-perception and their local businesses.
Beyond these four key stakeholder groups, there are also the waterbodies themselves, as non-human stakeholders providing habitats for aquatic biota. In recent years, to elevate their role in the system, there has been pressure to give rivers the same legal rights as people and businesses (Barkham 2021) and designers are arguing that water should be treated as an active stakeholder in every step of the process when designing solutions focused on managing it (Henning Larsen and Ramboll Water 2025). By integrating the many perspectives and voices that exist around the river and coasts, we aim to work with water and create solutions that address challenges holistically.
4.2.2. Perspectives on river and coastal health
In addition to mapping key stakeholders involved in river and coastal health, we also explored the different perspectives and meanings of health that are held by stakeholders. River and coastal health mean different things for different people (Karr 1999; Norris and Thoms 1999; Pinto et al. 2012). A point also raised during our interviews:
“River health (…) is a dynamic measure. It’s a bit elusive and it means different things to different people.” (Excerpt from interviews).
During our research, three overarching perspectives on river and coastal health emerged: Regulatory, Ecological, and Community (Table 1). Together, these perspectives provide a richer framework for understanding the worldviews of different stakeholders, supporting an increasingly shared understanding of the problems and how to shape solutions that work for all.
The regulatory perspective focuses on legal and regulatory compliance, often looking towards physical, chemical, and quantifiable thresholds as the main guidelines for determining river and coastal health. From a regulatory perspective, health is defined by policies and legislation. These instruments provide several ways for determining the health of a waterbody, but it often comes down to water quality and quantity, using readings of phosphate, oxygen, temperature, water levels, and other elements to determine the state of waterbodies. If these parameters exceed or fall below a certain threshold, a river or coastal waterbody is considered to have poor health.
Ecological perspectives on river and coastal health highlight the natural state of a waterbody as a key focus point, where things like a thriving biodiversity, natural flow regimes, and minimal human alterations are preferred. This perspective is focused on a waterbody’s ability to sustain life, rather than regulatory standards or swim-ability.
From an ecological perspective, focusing on chemicals or bacteria alone won’t provide the full picture. A river or coast might be considered healthy based on regulatory standards, but still have little to no life in them, nor provide the habitat needed to sustain life. As one participant argued:
“(…) if you’ve got fish in the river, that’s a good indication of river health. […] you can have no chemicals, and no nutrients, and you still may have low populations of fisheries and invertebrates.” (Excerpt from interviews).
Similarly, from an ecological perspective, human alterations such as weirs and dams are seen as disrupting the natural habitat needed to support fish and other aquatic life – thereby reducing the health of rivers. A perspective also represented by Michael Malay in the anthology “By the River”:
“The rivers are different now. There are more sluice gates, more dams, more tidal flaps, more concrete and more steel, and what chance do the elvers (ed. young eels) have against concrete?” (Malay, 2022/2024).
Lastly, community perspectives focus on the relationship between humans and waterbodies, often centered around perceived cleanness, aesthetics, and safety. Trash, odor, fear of getting sick, and sewage pollution play a central role in the perception of river and coastal health. In their study of community perspectives on river health, Pinto et al. (2012) similarly found that the visual appeal of rivers is perceived as a key indicator to determine river health, more important than specific water quality parameters.
Another focus in the community perspective is human accessibility to waterbodies. A healthy waterbody is accessible by humans to enjoy and connect with, recreationally, culturally, or spiritually (Deakin [1999] 2024; Pinto et al. 2012). As a participant put it, a healthy river is:
“a river that (…) can be enjoyed recreationally by people because it’s really important for a healthy river, is the connection with the people in the local area” (Excerpt from interviews).
Through our research, it became clear that the system around waterbody health is very complex, with individual hurdles, diverse perspectives, and tensions that hinder collaborative efforts, ownership, and solutions.
4.3. Establishing a Shared Vision and Understanding
Our use of interviews and observations provided a rich understanding and insights into the diverse perspectives and complexity surrounding river and coastal health. However, multiple perspectives on their own are not enough; they need to be synthesized into a complete and more coherent big-picture reality (Wahl [2016] 2022; Gharajedaghi 2011). Utilizing methods adopted from participatory design, we have created spaces of co-creation, where these insights inform the collaboration between stakeholders.
Both Wahl (2019) and Ostrom (2007) argue that overcoming complex problems with multiple diverse perspectives is possible through collaboration, trust, and by creating a shared understanding of the problems at hand and how to solve them.
Through recurring co-creation sessions, we have invited partners in the project to iteratively shape, make sense of, and re-shape the vision of River Deep Mountain AI. This includes using the expertise and diverse ways of knowing in the project to collaboratively explore technological opportunities, freshwater ecology, regulations, and dive deeper into the learnings from our ethnographic engagements. As the project includes a broad range of organizations representing the water sector, regulators, NGOs, freshwater ecologists, and technologists, these sessions have enabled us to “live the questions together” (Wahl 2019) and supported collective sense-making. This openness and collaborative approach have also helped foster trust between partners in the project, where everyone is involved, heard, and plays a role in shaping the project.
This was particularly evident during our largest co-creation session: A four-day innovation sprint that brought together more than 70 participants from 47 different organizations, inside and outside of the project. Rather than a stand-alone event, the innovation sprint became a natural extension of our collaborative process, providing a space to deepen the shared understanding of waterbody health and co-develop the vision for RDMAI. Through presentations from practitioners and technologists, stakeholder mapping and creative methods, we explored different aspects, perspectives and directions for the project together.
The broad range of organizations and multi-disciplinarity of the project has allowed for in-depth discussion and sharing of different approaches to addressing the shared problems at hand. These sessions played a crucial role in generating a shared understanding of the opportunity space represented by novel technologies, such as artificial intelligence (AI), allowing organizations not normally focused on AI and data science to think beyond their traditional ways of working. Similarly, these sessions enabled technologists to understand the world they are working within through a different lens, building a richer understanding of river and coastal health. This has facilitated a dance between disciplines, where boundaries between ways of working have been open for negotiation, exploration, and re-shaping together. This allowed for creative collaboration between disciplines, encouraging freshwater ecologists, catchment managers, and regulators to see the world through the eyes of data scientists – and vice versa.
These sessions have been instrumental in shaping the direction of the project, and as mentioned earlier, they have since become a recurring element in the project beyond the discovery phase, continuously creating a space for aligning on the vision of the project and solutions.
5. Impact: Navigating Complexity
Our approach has helped facilitate and operationalize collective intelligence, which has been critical in shaping the direction of RDMAI in a way that goes beyond traditional human-centered practices.
5.1. Operationalizing Collective Intelligence
Combined, our use of ethnographic and participatory practices has functioned as catalysts for integrating diverse perspectives and facilitating collective sense-making. This has facilitated the emergence and operationalization of collective intelligence in River Deep Mountain AI.
Wahl ([2016] 2022) argues that to facilitate the successful emergence of collective intelligence, stakeholders need to, amongst other things, deal with diverse perspectives and integrate multiple stakeholder perspectives into a “more complex big-picture reality”. Through our ethnographic engagements, we have generated a rich understanding of diverse perspectives and stakeholders in river and coastal health, enabling us to get a broader and more systemic view of the challenges and interconnectedness of the problems. This has created a more holistic and comprehensive view of the world, which we actively draw on in the project to get a wide array of perspectives integrated into solutions.
Beyond diverse perspectives, facilitating collective intelligence requires a shared vision, understanding of collective purpose, and an ability to encounter more of the whole (Wahl 2019). Using participatory methodologies, we have invited stakeholders to think and make sense of the world together through dialogue, deliberation, and co-creation (Atlee 2008). The diverse perspectives and complex big-picture realities developed through our interviews and observations fed into these sessions. This created a shared space where stakeholders involved in the project come together to integrate multiple perspectives, create a shared vision of the project, and develop solutions and mental models together (Atlee 2008). In these shared sessions, multi-disciplinarity and diversity are used creatively to overcome silos, see a bigger and more complete picture, and develop better solutions than any stakeholder could alone (Atlee 2008).
By operationalizing and leveraging collective intelligence, we have been able to move beyond traditional human-centered practices.
Relying on our understanding of individual perspectives, we can collaboratively zoom in viewing challenges through a human-centered lens, understanding the drivers, needs and perspectives of local communities, farmers, water companies or other stakeholders individually. At the same time, we can use our broader understanding of dynamics in the system and interconnectedness between stakeholders, to zoom out and view challenges, opportunities, and interventions through a systemic lens.
This provides a novel approach to navigating complexity, diagnosing real problems, and addressing complex challenges, enabling us to design effective interventions that drive positive change in the system, rather than continuing to do “the wrong thing righter” (Wahl 2019; IDEO 2025).
5.2. Moving beyond Discovery
The approach and methodology described in this case study have had a fundamental impact on the direction of River Deep Mountain AI and are playing a key role in the project beyond the initial discovery phase outlined above. Our ethnographic and participatory foundation continues to play an important role in shaping the AI models we are developing, sustaining collaborative ways of working, and helping stakeholders transcend silos and tensions.
All AI models developed as part of RDMAI have been designed with versatility in mind. Instead of narrow, tailor-made solutions, we are developing AI building blocks that can be leveraged by multiple stakeholders to address multiple challenges. Through collaborative synthesis, we have specified key elements of river and coastal health, where AI has the potential to play a central role in unlocking new pathways for stakeholders to understand and solve challenges. This has led to the development of AI models focused on predicting specific water quality parameters, understanding flow patterns, mapping pollution hotspots, and more. All of which offer new ways for stakeholders to understand river and coastal health from multiple perspectives. All models have adaptability as a core element in their design. Local communities might adapt a set of our AI models to get a better understanding of the impact they are having in their local river, while regulators might use the same models to reduce their reliance on time-consuming sampling routines, enabling them to cover larger areas faster. Similarly, NGOs might use these models to reduce the time spent doing river walkovers and instead focus their efforts where they can have the biggest impact. In addition, the models are built with adoption in mind, releasing all models open source and predominantly building models using openly available and existing datasets. This helps democratize the use of our models, ensuring that communities, NGOs, and industries alike have the same access to the models. This also enables the AI models, once developed and released, to integrate as part of the collective intelligence that has emerged during the project and play a role in river and coastal health beyond RDMAI.
Additionally, the openness and transparency that comes from actively engaging and integrating multiple perspectives has helped build an unprecedented trust and willingness to share between diverse stakeholders. Creating a continuously expanding network, where new stakeholders reach out, volunteering to collaborate, share data, and take part in shaping solutions. Since our discovery phase, we have been in contact with regulatory teams willing to share their experiences working with predictive modelling, preservation initiatives, and NGOs offering to share the data they are collecting, and academics offering insights from their research. This expansion helps to support, elevate, and improve the outcomes of RDMAI. Accelerating collective intelligence beyond the research phase, and beyond the project itself.
5.3. From Collecting Insights to Weaving
Our approach has also led to a reflection about our own role as researchers, especially the role we play after a discovery phase. The original intention was to reduce research allocations after the initial phase of the project and instead increase the focus on development. But it was quickly noticed that the diversity of perspectives, different ways of knowing, and mutual understandings began to fade. It became clear that collective intelligence requires continuous nurturing throughout the project – not just in the beginning. This presented a real risk of the project falling back into siloed solutioning. To address this, it was decided to pivot and keep researchers as a core part of the project, continuing to bring multiple perspectives together and facilitate the connection between different stakeholders. A role described by Wahl (2018) as weavers, where people, who through the initial research process have become capable of speaking the language of different disciplines, effectively broker conversations between diverse stakeholders. During the discovery phase, our use of co-creation sessions successfully connected previously unconnected stakeholders, but moving forward, it has become the role of weavers to facilitate the continuation of collaboration through these sessions.
Building on insights gathered in the discovery phase, we began to shift our focus towards more hands-on collaboration where development work and contextual knowledge would meet directly. Our role evolved from collecting insights to facilitating and nurturing the collaboration between practitioners and technologists.
Early sessions brought together large groups of domain experts to reflect on preliminary findings and explore ways forward. While these sessions gave valuable insights, we soon found that the group size unintentionally limited participation. Some voices dominated, while others withdrew, and important nuances were easily lost.
In response, we moved to smaller, more focused groups of four to five people. These sessions were higher in frequency (fortnightly) and brought together a mix of domain experts and data scientists to reflect on raw findings, review early model output, and challenge assumptions around methods and limitations. This format created a more open and constructive arena for knowledge exchange – continuing the dance between disciplines, where knowledge moved back and forth, and ideas and strategies were shaped collaboratively.
Importantly, these sessions gave practitioners, domain experts, and stakeholders across organizations a seat at the table where development decisions were made, not just reviewed. This helped keep development grounded in real-world needs and built trust and shared ownership along the way.
This ongoing weaving has become an integral part of the way of working in RDMAI, offering a recurring space for bridging diverse perspectives and ways of knowing – to build alignment and nurture collective intelligence.
5.4. More-than-Human
Upon finishing our discovery phase, we have had time to reflect and dive deeper into the complexities of designing solutions for living systems such as rivers and coastal areas, where non-human actors and humans are entangled together. Through our approach, we have dealt with multiple and diverse perspectives in creative ways to understand the complexities of river and coastal health and design solutions that work for all. We have arguably also introduced additional non-human actors into the system in the form of artificial intelligence.
AI is increasingly becoming an integrated part of how we make sense of the world together. By connecting and synthesizing vast volumes of data that would otherwise exceed human capacity, AI systems can reveal patterns, dynamics, and interdependencies that are difficult to grasp through human observations alone. It can become another voice of the river, enabling us to understand waterbodies through data. When developed in close dialogue with domain expertise, AI does not replace human judgement, it extends it, supporting more nuanced, contextual, and holistic ways of understanding complex systems. In this sense, AI becomes not just a tool, but a non-human collaborator in how we engage with complex waterbody data, expanding our capacity for understanding and elevating our collective intelligence.
If we were given the opportunity to start over, we would embrace the relationality between humans and non-human actors even more. Novel approaches that go a step further than us are emerging in academia and industry. Henning Larsen and Ramboll Water (2025) present what they describe as a new water paradigm, offering a set of tools to include water as an active stakeholder in every step of the design process. Nielsen and Neuhoff (2025) introduce non-human personas as a tool for designing more inclusive and ecologically responsible solutions, and Harre et al. (2025) argue that artistic practices can be applied as a tool to shift perspectives, and through that, extend the involvement of non-human actors.
Tools and approaches that offer new ways of immersing ourselves into the needs and challenges of rivers and coasts, and other non-human actors, treating them as active stakeholders in the design process.
Through our case study, we have illustrated how ethnographic and participatory practices can, and perhaps should, be embedded in projects and organizations to play a much larger role, beyond the discovery phase. By weaving together diverse perspectives, these approaches offer a framework for navigating complexity, enabling practitioners to leverage collective intelligence, reduce risk, and drive more effective actions in increasingly complex systems.
Moving forward, we encourage researchers and designers to take this even further when working with living systems that are inherently more-than-human and use emerging practices to explore the relationality between human and non-human actors in new ways, enabling them to design solutions that work for all.
About the Authors
Nicolai Traasdahl Tarp is a qualitative researcher at Cognizant Ocean, working in the intersection between systemic design and technology to address challenges in freshwater ecosystems. He holds an MSc from the IT University of Copenhagen, specializing in Digital Democratic Citizenship.
Elisabeth Haaland Sund is a designer and strategist at Cognizant Ocean, working on leveraging AI technology to restore the health of our aquatic systems. With a background in systemic design, she takes a holistic approach to addressing complex water management challenges. She holds a Master’s in Interaction Design from the Norwegian University of Science and Technology (NTNU).
Research Ethics
Our research adheres to established ethical standards recognized across academic and industry domains, ensuring integrity, transparency, and the protection of participant rights throughout the study. By adhering to these principles, our research not only addresses the multifaceted challenges of the water sector but also exemplifies a commitment to ethical research practices that prioritize inclusivity and balanced representation.
We outline our ethical considerations in the following paragraphs.
Equitable Selection of Participants
To achieve a diverse and representative engagement, we proactively included a broad spectrum of stakeholders from multiple organizations and backgrounds. To get access to the relevant subject matter experts (SMEs), we employed snowball sampling, initiating contact with key stakeholders who then referred additional participants. This method was particularly effective for accessing the sectors´ interconnected community, enabling us to examine the sector challenges from a wide range of perspectives. Understanding the diverse needs has been crucial for defining how AI can benefit the entire sector.
Potential Harm
Our research brings together perspectives from the water sector, regulatory agencies, local communities, industry, NGOs, and more. Consequently, the insights collected during our research contain opposing perspectives, conflicting values, and contrasting agendas. The risk of potential harm and reputational damage to participants cannot be neglected. To take the necessary steps to protect participants, we have made sure that all participation has been voluntary and informed, and we have ensured anonymity and confidentiality of all collected data.
Voluntary Participation and Informed Consent
All participants in our research have engaged with us on a voluntary basis. When participants were referred to us through their workplace, we made sure to make it clear before and during sessions that participation was voluntary and that they could withdraw if desired. Furthermore, we got verbal consent from participants at the beginning of each session to ensure they agreed to be recorded and accepted our intended use of the session. In relation to this, we made sure to inform participants about our research, objectives, and affiliations before starting the sessions.
Anonymity and Confidentiality
To ensure anonymity, all notes, quotes, and observations have been anonymized prior to synthesis of the data. We have removed any relations to specific company names, agencies, and geographies. Instead, we have chosen to add industry labels to participant data (academia, water company, NGO, regulator, etc.) to give context to insights. Furthermore, we have maintained strict confidentiality of all data collected. Access to sensitive information was limited to authorized research team members. Data and recordings have been stored securely to avoid unauthorized access.
Notes
This project would not have been possible without the full consortium of partners involved in River Deep Mountain AI and their willingness to collaborate closely and share openly. Furthermore, we are deeply grateful for the participants in our interviews and observations, offering their valuable time to support our project. River Deep Mountain AI, and our related research, was funded by the Ofwat Innovation Fund, managed by the Water Services Regulation Authority for England and Wales (Ofwat). We sincerely thank Therese Kjellerup Thorstholm for her invaluable support and editorial feedback.