Adoption of Large Language Models in Scrum Management: Insights from Brazilian Practitioners

Scrum is widely adopted in software project management due to its adaptability and collaborative nature. The recent emergence of Large Language Models (LLMs) has created new opportunities to support knowledge-intensive Scrum practices. However, existing research has largely focused on technical activities such as coding and testing, with limited evidence on the use of LLMs in management-related Scrum activities. In this study, we investigate the use of LLMs in Scrum management activities through a survey of 70 Brazilian professionals. Among them, 49 actively use Scrum, and 33 reported using LLM-based assistants in their Scrum practices. The results indicate a high level of proficiency and frequent use of LLMs, with 85% of respondents reporting intermediate or advanced proficiency and 52% using them daily. LLM use concentrates on exploring Scrum practices, with artifacts and events receiving targeted yet uneven support, whereas broader management tasks appear to be adopted more cautiously. The main benefits include increased productivity (78%) and reduced manual effort (75%). However, several critical risks remain, as respondents report ‘almost correct’ outputs (81%), confidentiality concerns (63%), and hallucinations during use (59%). This work provides one of the first empirical characterizations of LLM use in Scrum management, identifying current practices, quantifying benefits and risks, and outlining directions for responsible adoption and integration in Agile environments.

💡 Research Summary

This paper investigates how large language models (LLMs) are being incorporated into Scrum management activities, focusing on a sample of Brazilian software professionals. The authors conducted a structured online survey with 70 respondents, of whom 49 currently practice Scrum and 33 report using LLM‑based assistants in their Scrum work. The study is organized around four research questions: (RQ1) the level of knowledge and usage of LLMs among Scrum practitioners; (RQ2) which Scrum artifacts, events, and roles are supported by LLMs and how helpful they are perceived to be; (RQ3) the benefits practitioners experience; and (RQ4) the risks and challenges associated with LLM use in Scrum contexts.

Survey design followed a Goal‑Question‑Metric (GQM) approach, and the questionnaire was validated through expert review and a pilot with five Scrum practitioners. It comprised nine sections covering demographics, LLM familiarity, frequency of use, specific Scrum‑related applications, perceived benefits, and perceived risks. Recruitment employed convenience and snowball sampling via LinkedIn Agile groups, professional mailing lists, and direct invitations, ensuring a diverse but non‑probabilistic sample.

Key findings show a high level of proficiency: 85 % of respondents rate their LLM skills as intermediate or advanced, and 52 % use LLMs daily. LLM adoption concentrates on “exploring Scrum practices” – searching for guidelines, summarizing standards, and clarifying terminology. In concrete management tasks, LLMs are used for backlog refinement, sprint planning, meeting agenda creation, and retrospective summarization, but support is uneven across artifacts, events, and roles. Product Owners and Scrum Masters report the least perceived replacement potential, indicating that LLMs are currently viewed as augmentative rather than substitutive.

Perceived benefits are strong: 78 % cite increased productivity, 75 % note reduced manual effort, and additional advantages include improved documentation quality, enhanced communication, and decision‑making support. However, significant risks are also reported. “Almost correct” outputs are a concern for 81 % of users, confidentiality and privacy worries affect 63 %, and hallucinations (fabricated or misleading information) are experienced by 59 %. Other challenges include over‑reliance, bias, organizational resistance, and a lack of systematic prompt‑engineering practices.

The authors discuss that while LLMs hold promise for mitigating the knowledge‑intensive nature of Scrum, their current deployment is still limited compared with technical coding‑oriented uses. Trust, verification, and governance mechanisms are essential before broader adoption. The paper acknowledges limitations such as the non‑representative sample, self‑report bias, and a focus on mainstream LLM platforms (e.g., ChatGPT). Future work is suggested to include longitudinal studies measuring objective productivity gains, deeper analysis of cultural and organizational factors influencing adoption, and the development of best‑practice guidelines for prompt design and risk mitigation.

In conclusion, this study provides one of the first empirical characterizations of LLM use in Scrum management, mapping current practices, quantifying perceived benefits, and highlighting critical risks. It offers a foundation for both practitioners and researchers aiming to integrate LLMs responsibly into Agile environments.