Rasmus Ulfsnes

Tita Alissa Bach, Aleksandar Babic, Narae Park et al.

The maritime industry requires effective communication among diverse stakeholders to address complex, safety-critical challenges. Industrial AI, including Large Language Models (LLMs), has the potential to augment human experts' workflows in this specialized domain. Our case study investigated the utility of LLMs in drafting replies to stakeholder inquiries and supporting case handlers. We conducted a preliminary study (observations and interviews), a survey, and a text similarity analysis (LLM-as-a-judge and Semantic Embedding Similarity). We discover that while LLM drafts can streamline workflows, they often require significant modifications to meet the specific demands of maritime communications. Though LLMs are not yet mature enough for safety-critical applications without human oversight, they can serve as valuable augmentative tools. Final decision-making thus must remain with human experts. However, by leveraging the strengths of both humans and LLMs, fostering human-AI collaboration, industries can increase efficiency while maintaining high standards of quality and precision tailored to each case.

Tor Sporsem, Rasmus Ulfsnes

This short paper explores how a maritime company develops and integrates large-language models (LLM). Specifically by looking at the requirements engineering for Retrieval Augmented Generation (RAG) systems in expert settings. Through a case study at a maritime service provider, we demonstrate how data scientists face a fundamental tension between user expectations of AI perfection and the correctness of the generated outputs. Our findings reveal that data scientists must identify context-specific "retrieval requirements" through iterative experimentation together with users because they are the ones who can determine correctness. We present an empirical process model describing how data scientists practically elicited these "retrieval requirements" and managed system limitations. This work advances software engineering knowledge by providing insights into the specialized requirements engineering processes for implementing RAG systems in complex domain-specific applications.

Anastasiia Tkalich, Nils Brede Moe, Rasmus Ulfsnes

With the increasing availability of software usage and the influence of the Lean Startup mindset, more and more companies choose to innovate through internal software startups. Such startups aim at developing new business models while at the same time relying on the resources from the companies where they emerged. The evidence from both researchers and practitioners indicates that driving internal software startups is challenging. This paper seeks to address this problem by asking the research question: how to make internal software startups work? We examined a unique case of a venture builder, a company primarily focusing on building internal software startups and launching them as independent companies. Applying a Grounded Theory approach, we analyzed data on four internal software startups at the case company. The results suggest that four strategies drive the examined startups, cultural, financial, personnel, and venture arrangement. We interpret our results by drawing on earlier literature on intrapreneurship and internal ventures and suggest four recommendations to succeed with internal software startups 1 establish shared arenas for the employees, 2 provide necessary resources for experimentation in the initial phase and increase them incrementally, 3 build up in-house product management competence through coaching, and 4 harness employees own motivation to develop their own ideas.

Nils Brede Moe, Rasmus Ulfsnes, Viktoria Stray et al.

Software development companies organize hackathons to encourage innovation. Despite many benefits of hackathons, in large-scale agile organizations where many teams work together, stopping the ongoing work results in a significant decrease in the immediate output. Motivated by the need to understand whether and how to run hackathons, we investigated how the practice affects productivity on the individual and organizational levels. By mapping the benefits and challenges to an established productivity framework, we found that hackathons improve developers' satisfaction and well-being, strengthen the company culture, improve performance (as many ideas are tested), increase activity (as the ideas are developed quickly), and improve communication and collaboration (because the social network is strengthened). Addressing managerial concerns, we found that hackathons also increase efficiency and flow because people learn to complete work and make progress quickly, and they build new competence. Finally, with respect to virtual hackathons we found that developers work more in isolation because tasks are split between team members resulting in less collaboration. This means that some important, expected hackathon values in virtual contexts require extra effort and cannot be taken for granted.

Rasmus Ulfsnes, Viktoria Stray, Nils Brede Moe et al.

Hackathons are events in which diverse teams work together to explore, and develop solutions, software or even ideas. Hackathons have been recognized not only as public events for hacking, but also as a corporate mechanism for innovation. Hackathons are a way for established companies to achieve increased employee wellbeing as well as being a curator for innovation and developing new products. Sudden transition to the work-from-home mode caused by the COVID-19 pandemic first put many corporate events requiring collocation, such as hackathons, temporarily on hold and then motivated companies to find ways to hold these events virtually. In this paper, we report our findings from investigating hackathons in the context of a large agile company by first exploring the general benefits and challenges of hackathons and then trying to understand how they were affected by the virtual setup. We conducted nine interviews, surveyed 23 employees and analyzed a hackathon demo. We found that hackathons provide both individual and organizational benefits of innovation, personal interests, and acquiring new skills and competences. Several challenges such as added stress due to stopping the regular work, employees fearing not having enough contribution to deliver and potential mismatch between individual and organizational goals were also found. With respect to the virtual setup, we found that virtual hackathons are not diminishing the innovation benefits, however, some negative effect surfaced on the social and networking side.