Sign Up
ENعربي
All articles

Project Management

Human & AI

  • Introduction

The burgeoning field of Human-Artificial Intelligence (AI) collaboration has garnered considerable attention as organizations across various sectors increasingly recognize the imperative to leverage the transformative potential of AI while effectively mitigating associated risks and negative consequences. This paradigm shift acknowledges that augmenting human capabilities with AI, rather than pursuing full automation, can lead to more robust, reliable, and trustworthy outcomes. The specific relevance of this collaborative approach is particularly pronounced within the domain of case management, especially in public sector organizations that operate under constant pressure to deliver high-quality, trustworthy services amidst constrained resources. Effective case management, which involves handling processes and documentation for individual matters according to established regulations and procedures, is critical for public bodies to ensure fairness, consistency, and timely processing in areas such as citizen services, permits, and applications.

This literature review aims to summarize, synthesize, and critically evaluate the insights presented in the research paper "Human-AI Collaboration for Case Management: Exemplars, Required Capabilities and Transition Management" by Tveita et al. (2025). Drawing on an empirical study involving public sector organizations and case management enterprise systems providers in Norway, the research identifies specific tasks well-suited for AI augmentation, delineates the capabilities required for involved organizations (including regulators), and suggests tactics for successful transition management. By examining how public sector organizations can strategically leverage AI capabilities in conjunction with human judgment, this study addresses a gap in prior literature and provides valuable insights for capitalizing on AI's benefits for increased efficiency while maintaining essential human oversight and expertise to ensure high-quality service delivery. This review will delve into the main themes identified by the authors, explore points of agreement and debate, highlight research gaps, analyze the methodology employed, discuss the development of the field, and offer critical insights on the study's contribution to the understanding of Human-AI collaboration in case management.

  • Main Themes in Human-AI Collaboration for Case Management

The study by Tveita et al. (2025) identifies several key themes crucial for understanding and effectively deploying Human-AI collaboration in case management. These themes encompass specific areas where AI can augment human capabilities, the competencies required for successful integration, and strategies for managing the transition process.

  • Streamlining Case Management through AI Augmentation: The research highlights five key exemplars where AI can significantly contribute to streamlining case management processes.
  • Rapidly Compiling Information: Participants emphasized AI's ability to swiftly gather and synthesize large volumes of case-related data from diverse sources. This capability can substantially reduce the manual effort required by case workers to build a comprehensive understanding of a case. As Informant 1 explained, "today, a case manager needs many years of experience to find what is needed. AI will be a good support for retrieving and finding information". Similarly, Informant 4 noted that AI could quickly provide an overview of numerous case papers through compilation. The suggestion to use AI for developing a timeline of case progress further underscores this benefit. Informant 11 highlighted the time-consuming nature of understanding a case with unsorted documents, suggesting that AI could summarize information, create timelines, or consolidate essential documents. This type of AI-driven augmentation acts as a "smart assistant," enhancing both efficiency and work quality.
  • Resolving Routine Matters: For routine cases with clearly defined criteria, AI can undertake tasks entirely, automating their resolution. Informant 6 illustrated this with the example of simple inquiries about time or opening hours that could be automated. While Informant 9 suggested that rule-based systems might be simpler for repetitive operations, Informant 10 provided examples of regulated areas like construction and liquor licensing where AI, trained on Norwegian regulations, could apply given criteria. This delegation of repetitive tasks opens the door to automated case management, which was viewed positively by some informants but also met with caution, as highlighted by Informant 9, who emphasized AI's role in assisting rather than making decisions. Informant 2 suggested automating the more procedural aspects of case management, with human oversight at the end. However, Informant 11 pointed out that cases requiring practical experience and common sense might necessitate human-…
  • Assuring Quality and Preventing Errors: AI systems can analyze compiled case information and offer recommendations to case workers, thereby reducing the risk of human errors. Informant 4 suggested that AI could provide useful tips where human error is a concern. Several informants desired AI integration to provide real-time tips and adjustments, acting as a quality assurance mechanism to ensure case managers do not overlook important details, write correct legal citations, and meet deadlines. Informant 6 envisioned AI as a "support tool" for quality control before inquiries are sent. Informant 10 believed AI could reassure case managers that they are on the right track and within the bounds of legislation. Informant 11 highlighted AI's potential to identify crucial legal sources, a significant challenge given the volume of legal information. Informant 8 noted the long-term potential of AI to enable preventive case management at an organizational level by identifying patterns in data t…
  • Redacting Sensitive Data: AI can be employed to identify and redact personally sensitive information from case documents before publication or external sharing, ensuring privacy and compliance with data protection regulations. Informant 6 described this as a time-consuming manual process prone to oversight. Informant 7 corroborated the significant workload associated with public access requests and the need to compile documents. Informant 5 emphasized the potential of AI to facilitate citizens' right to access public information by efficiently checking for and redacting sensitive data. Informant 4 further suggested that AI could proactively alert case workers within processes about the inclusion of potentially sensitive personal information.
  • Strengthening the Evidence Basis for Decisions: With AI assistance, case workers can make more well-informed, evidence-based, and consistent decisions. Informant 8 questioned whether a greater knowledge base provided by AI would lead to different decisions. Informant 10 believed AI could contribute to fairer proceedings through machine learning. While Informant 8 acknowledged the potential for increased efficiency, they also raised the question of cost-effectiveness, emphasizing that enhanced legal certainty should be the primary driver for AI integration.
  • Building Essential Competencies for Human-AI Collaboration: The successful deployment of Human-AI collaboration necessitates the development of relevant competencies across solution providers, public organizations, and regulatory bodies.
  • Competence of Solution Providers: Given the relatively recent proliferation of AI technologies, it is crucial for suppliers to cultivate the necessary expertise and gain confidence in their AI solutions before offering them to customers. Informant 3 stressed the importance of internal experimentation to determine AI's maturity for solving customer problems. Informant 1 echoed this, highlighting the need for product owner competence before client rollout. Informant 2 described internal AB testing with a few customers to gather test data and build expertise.
  • Competence of Public Organizations: Client organizations adopting AI-enabled case management must develop both technical skills to utilize AI features and, more importantly, AI literacy to understand how to effectively integrate AI into their workflows. Informant 3 emphasized that customers need to change their processes to leverage the software effectively, rather than trying to fit new technology into old ways of working. Informant 11 noted that automating simple tasks might lead to more complex remaining tasks, requiring greater professional demands on case managers. Informant 4 highlighted the importance of intuitive solutions that minimize the need for extensive training. Informant 8 stressed the need to involve and respect experienced professionals in the change process, while Informant 5 suggested that experienced employees might appreciate the novelty of AI.
  • Competence of Regulators (Legal Maturity): Ensuring AI systems comply with all relevant laws, policies, and regulations concerning data privacy, security, and citizen rights is paramount for public services. Several informants expressed concerns about whether current legislation adequately addresses advanced AI applications, with regulatory uncertainty potentially hindering adoption. Informant 1 suggested the need for a more digitalization-friendly regulatory framework. Informant 5 believed that laws are not keeping pace with technological progress. Informant 10 emphasized the necessity of clear rules to safeguard privacy and enable safe experimentation. Informant 6 noted that fear of legal conflicts regarding privacy and a lack of clarity on obtaining good training data are causing hesitancy in AI adoption.
  • Effective Transition Management for Human-AI Collaboration: A thoughtful and well-managed transition is essential for the successful integration of Human-AI collaboration in case management.
  • Balancing Trust and Discretion: Public sector case workers need to develop an appropriate level of trust in AI systems' recommendations while also understanding their limitations and exercising human discretion when necessary. Informant 9 emphasized the need to provide good information and training, acknowledging that AI can make mistakes. Informant 7 pointed to the difficulty of determining an acceptable margin of error for AI models before deployment, comparing it to human error rates. Informant 2 highlighted that Norwegian legislation often involves discretion, suggesting AI should propose rather than decide in such cases. Informant 11 underscored the ultimate responsibility of humans in the loop and the state's non-outsourcable accountability.
  • Limited Scope and Pilot Projects: Adopting a phased approach, beginning with limited scope and pilot projects, was recommended before wider-scale deployment. This allows for the evaluation of AI solution performance, identification of unforeseen issues, refinement of processes, and building confidence through successful use cases. Informant 8 suggested starting with a specialist area. Informant 7 recommended starting with quality control before expanding to document production. Informant 4 emphasized the frequent use of piloting to test reception and make necessary changes. Informant 5 advised starting with one well-functioning aspect before moving to others to avoid overwhelming the process.
  • Engagement between Solution Providers and Customer Organizations: Continuous dialogue and close collaboration between AI/IT teams and client organizations are crucial for redesigning processes to effectively integrate AI capabilities. Managing client expectations regarding the current and future potential of AI is also important to avoid rejection due to over-expectations or limited adoption due to underestimation. Informant 5 emphasized customer meetings to define needs and challenges. Informant 4 highlighted the importance of incorporating customer feedback during the development process. Informant 5 further stressed starting with the specific challenges and focusing on solutions that provide the most value to the customer.

III. Points of Agreement, Debate, and Gaps in the Research

The study reveals several points of agreement, debate, and gaps in the current understanding of Human-AI collaboration in case management. There is a general agreement on the potential of AI to enhance efficiency and streamline operations within public sector case management. Participants and the researchers concur on the critical and continued importance of human involvement to ensure outcomes are trustworthy, ethical, and accountable, reinforcing the notion of collaboration rather than complete automation. Furthermore, there is a shared acknowledgment of the challenges inherent in integrating AI, including the necessity for developing new competencies, effectively managing trust in AI systems, and navigating the complexities of the regulatory landscape.

Points of debate emerge regarding the optimal extent and nature of AI's role. Some informants suggested that simpler, rule-based systems might suffice for certain repetitive tasks, questioning the necessity of more complex AI solutions in all instances. An implicit tension exists between the desire for increased automation and the fundamental need to retain human discretion in complex cases, particularly within the context of Norwegian legal frameworks that often require discretionary assessments. Concerns were also raised by informants regarding the pace at which legal and regulatory frameworks are evolving to keep pace with the rapid advancements in AI technology.

The study identifies several significant gaps in the existing research. Prior literature, while acknowledging AI's potential in the public sector and case management, has lacked specific and concrete insights into how to effectively enact Human-AI collaboration models in this particular domain. This study directly addresses this gap by providing empirically grounded exemplars and practical recommendations for implementation. Another identified gap is the need for further and more in-depth exploration into the development of AI literacy within public organizations, extending beyond mere technical proficiency to encompass a deeper understanding of how to judiciously integrate AI into human-driven processes. Furthermore, there is a recognized lack of research on how to effectively balance trust and discretion in Human-AI decision-making processes within the specific context of case management. Finally, the study's context-specific focus on the Norwegian public sector suggests a gap in research across different national, cultural, and organizational contexts, limiting the generalizability of the findings.

  • Relevant Methodologies

The research conducted by Tveita et al. (2025) adopted an interpretive qualitative approach to investigate how AI-human collaboration can be enacted for case management. This methodological choice aligns with the exploratory nature of the research and allows for the development of rich insights from the participants' experiences and perspectives. The primary method of empirical data collection was through semi-structured interviews conducted with 11 employees holding diverse roles within a leading Norwegian technology organization providing case management enterprise systems and services, as well as their public sector clients. These interviews, with durations ranging from 43 to 72 minutes, were recorded and transcribed verbatim. An iterative process of thematic analysis was employed to identify key recurring themes across the collected data. To ensure the quality of the interpretive research, the researchers applied principles proposed by Klein and Myers, including exploring multiple viewpoints, remaining open to reframing interpretations, and establishing trustworthiness through chains of evidence. Furthermore, the researchers adhered to processes for responsible research, ensuring informed consent, privacy protection, and responsible data management.

  • Development in the Field Over Time

The introduction of the study highlights that Human-AI collaboration has witnessed a significant surge in attention in recent years. This growing interest reflects a broader trend in the field of AI research and practice, moving from a focus on autonomous AI systems towards understanding and designing systems that can effectively work alongside humans. The study explicitly positions itself as addressing a crucial gap in the existing body of knowledge, which, while acknowledging the potential of AI within the public sector and specifically in case management, has largely failed to provide detailed and context-specific insights into effective models of Human-AI collaboration in this particular domain. This suggests a developmental trajectory in the field, where initial enthusiasm for AI's capabilities is now being tempered by a recognition of the necessity for human oversight, control, and ethical considerations, leading to a greater emphasis on collaborative approaches. The increasing societal concerns regarding the potential adverse impacts of AI technologies, such as accountability diffusion and job displacement, have further fueled the focus on collaborative AI models that aim to harness AI's benefits while mitigating these risks. Moreover, the informants' expressed concerns about regulatory frameworks lagging behind the rapid technological advancements in AI indicate a temporal misalignment between the development of AI capabilities and the legal and policy infrastructure intended to govern their use.

  • Critical Insights and Contextual Alignment/Divergence

This study by Tveita et al. (2025) offers a valuable contextualized understanding of Human-AI collaboration specifically within the public sector case management context in Norway. This context is uniquely shaped by a strong emphasis on established legal frameworks, the protection of citizen rights, and stringent public accountability measures. The research findings align with broader literature that acknowledges the significant potential of AI for driving automation, enhancing efficiency, and improving decision-making processes across various organizational settings. However, it diverges by its explicit focus on the collaborative paradigm of Human-AI interaction and by providing empirical evidence and specific examples derived from the distinct domain of public sector case management. This contrasts with more generalized discussions of AI adoption and its implications in organizations, which may not adequately address the unique challenges and requirements of public service delivery.

The study makes a significant contribution by providing a granular perspective on the specific tasks within case management that are particularly well-suited for AI augmentation. By offering concrete exemplars such as rapidly compiling information, resolving routine matters, assuring quality, redacting sensitive data, and strengthening the evidence basis for decisions, the research moves beyond abstract discussions of AI capabilities and provides practical insights for practitioners and policymakers. The strong emphasis on the development of AI literacy for public sector employees highlights a crucial enabler for successful AI adoption that may receive comparatively less attention in AI literature primarily focused on the private sector. The considerable attention dedicated to the competence of regulators and the pressing need for greater legal maturity in the face of rapidly evolving AI technologies underscores a critical consideration that is particularly pertinent within the inherently regulated environment of the public sector.

VII. Conclusion

This study by Tveita et al. (2025) provides significant empirical insights into the underexplored domain of Human-AI collaboration within the context of public sector case management. The research effectively identifies key exemplars of how AI can augment human capabilities to streamline processes while maintaining quality and trustworthiness. It also highlights the essential competencies required for solution providers, public organizations, and regulators to facilitate successful integration. Furthermore, the study offers valuable guidance on transition management strategies, emphasizing the importance of balancing trust, adopting a phased approach, and fostering continuous engagement between stakeholders.

The findings have important implications for both research and practice. For researchers, the study points to the need for further investigation into areas such as AI literacy, the dynamics of trust in Human-AI decision-making, and the evolution of regulatory frameworks in response to advancing AI capabilities. For public sector practitioners and policymakers, this research serves as a valuable reference for strategically pursuing AI integration in case management, highlighting priority areas for deployment and underscoring the critical need for workforce planning and training programs to build necessary competencies. The identified regulatory uncertainties also emphasize the importance of reassessing current laws and policies to provide greater clarity for the ethical and effective utilization of AI in public service delivery. Ultimately, this work reinforces the fundamental importance of harmonizing the powerful technological capabilities of AI with human-centric governance and accountability to ensure that AI adoption in public sector case management serves the best interests of citizens and upholds core public sector values.

Frequently Asked Questions: Human-AI Collaboration in Case Management

  • What is Human-AI Collaboration in Case Management?

Human-AI collaboration in case management refers to a partnership where Artificial Intelligence (AI) tools and systems are used to augment human capabilities, rather than replace them entirely. It's about leveraging AI's strengths (like processing large datasets, identifying patterns, and automating routine tasks) to support human case workers, who retain final decision-making authority and handle complex, nuanced situations. The goal is to improve efficiency, accuracy, and overall quality of service delivery.

  • Why is this approach important, especially in the public sector?

Public sector organizations often face high caseloads, limited resources, and strict regulatory requirements. Human-AI collaboration offers a way to:

  • Increase Efficiency: Automate routine tasks, freeing up case workers to focus on more complex cases requiring human judgment.
  • Improve Accuracy: Reduce human error and ensure consistency in applying rules and regulations.
  • Enhance Decision-Making: Provide case workers with better information and insights, leading to more informed and evidence-based decisions.
  • Maintain Accountability: Ensure human oversight remains central, upholding public trust and ethical standards.
  • Improve service delivery: By supporting case workers and providing a greater knowledge base.
  • What are some specific examples of how AI can be used in case management?

The article highlights several key applications:

  • Rapid Information Compilation: AI can quickly gather and summarize relevant information from various sources, saving case workers significant time.
  • Routine Matter Resolution: AI can handle simple, rule-based inquiries or tasks, automating their resolution.
  • Quality Assurance & Error Prevention: AI can act as a "second set of eyes," identifying potential errors, missing information, or inconsistencies.
  • Sensitive Data Redaction: AI can automatically identify and remove personally identifiable information (PII) from documents, ensuring privacy compliance.
  • Strengthening the Evidence Base: AI can analyze data to provide insights and support more evidence-based decision-making.
  • What are the challenges of implementing Human-AI collaboration?

Several challenges need to be addressed:

  • Developing AI Literacy: Case workers and organizations need to understand how to effectively use and interpret AI-powered tools.
  • Building Trust: Case workers need to develop an appropriate level of trust in AI recommendations, while still exercising their own judgment.
  • Regulatory Uncertainty: Laws and regulations may need to be updated to address the unique challenges of AI in public service.
  • Data Privacy and Security: Protecting sensitive citizen data is paramount.
  • Transition Management: Organizations need to carefully plan and manage the transition to new ways of working.
  • Competency of solution provider, organisation and regulators.
  • What skills are needed for successful Human-AI collaboration?
  • For Case Workers: AI literacy, critical thinking, data interpretation skills, and the ability to integrate AI insights into their decision-making process.
  • For Organizations: Technical expertise to implement and maintain AI systems, change management skills, and a commitment to ongoing training and development.
  • For Regulators: A deep understanding of AI technology and its implications, and the ability to develop clear and effective regulations.
  • For Solution providers Developing AI solutions, experiment with internal systems.
  • How should organizations approach the implementation of Human-AI collaboration?

The article recommends a phased approach:

  • Start Small: Begin with pilot projects in specific areas to test and refine the approach.
  • Focus on Collaboration: Emphasize that AI is a tool to support, not replace, human workers.
  • Provide Training: Ensure case workers have the necessary skills and knowledge to use AI effectively.
  • Engage Stakeholders: Involve case workers, IT teams, and regulators in the planning and implementation process.
  • Monitor and Evaluate: Continuously assess the performance of AI systems and make adjustments as needed.
  • What is the role of human discretion in this collaborative model?

Human discretion remains essential. AI is designed to provide recommendations and support, but the final decision-making authority rests with the human case worker. This is particularly important in cases involving complex ethical considerations, nuanced situations, or where legal frameworks require human judgment. The human "in the loop" ensures accountability and prevents AI from making inappropriate or biased decisions.

  • What is the long-term vision for Human-AI collaboration in case management?

The long-term vision is a more efficient, effective, and citizen-centric public sector. AI will handle routine tasks and provide valuable insights, while human case workers focus on complex cases, relationship building, and ensuring fairness and equity. This collaborative approach aims to improve the quality of public services while upholding core values of accountability, transparency, and citizen rights.

  • Is this approach applicable only to the Norwegian public sector?

While the study focused on Norway, the principles and challenges of Human-AI collaboration are relevant to public sector case management worldwide. However, the specific implementation may need to be adapted to different legal frameworks, cultural contexts, and organizational structures.

  • Where can I find more information about this topic? The study provides insights drawing on an empirical study involving public sector organizations and case management enterprise systems providers in Norway, the research "Human-AI Collaboration for Case Management: Exemplars, Required Capabilities and Transition Management" by Tveita et al. (2025).

This FAQ provides a good overview of the key issues and considerations related to Human-AI collaboration in case management. It's designed to be informative and accessible to a broad audience.

(Tveita et al., 2025)

References:

Tveita, L. J., Kotsialos, A., & Vassilakopoulou, P. (2025). Human-AI Collaboration for Case Management: Exemplars, Required Capabilities and Transition Management. Procedia Computer Science, 256, 166–173. https://doi.org/10.1016/j.procs.2025.02.109