Contemporary advanced business services and products increasingly incorporate artificial intelligence components, such as chatbots and generative AI, across various domains. These are delivered through dedicated, complex, innovative software programs and projects initiated by virtually all industry sectors. Complex project management include challenges such as: the complexity of product backlogs with a number of requirements, highly dynamic changes in customer expectations impacting product backlog quality and requirements engineering, labour shortages, advanced tool adoption and automation, predictability of deliveries, insufficient transparency of processes applied to product backlog management, communication barriers between business and project teams. The primary objective of this paper is to address a key research gap related to the insufficient quality of product backlog management in complex agile software project environments. The paper addresses the research question regarding the potential application of chatbots and generative AI as a methodology to conduct reliable agile product backlog evaluation and subsequently enhance refinement processes through dedicated Prompt Engineering Language (PEL). This paper emphasizes the importance of a structured description of product backlog items and its impact on the overall quality of agile product backlog and delivered software products. Following the literature review, the author's empirical research presents a detailed analysis of the research gap and focuses on applying chatbot and generative AI solutions to evaluate agile product backlog items and to improve related agile refinement processes. Research results demonstrate that agile product backlog refinement processes can be supported by chatbots and generative AI utilizing dedicated Prompt Engineering Language (PEL). These tools are not designed to create business value directly but rather enhance the efficiency and automation of product backlog management processes to respond rapidly to stakeholder expectations within agile environments, ultimately achieving superior project outcomes. Nevertheless, numerous challenges must be addressed, particularly related to AI governance and compliance with numerous policies including data privacy, intellectual property, legal, security and internal ones.

The paper presents practical applications of AI assistants in everyday software engineering: requirement formalization, code scaffolding, SQL/migrations support, log‑based fault localization, test generation, documentation upkeep, and CI/CD optimization. We summarize observable benefits (time savings, defect reduction), typical risks (hallucinations, data leakage, licensing), and a lightweight evaluation method based on before/after measurements on real backlog items. Implementation guidelines for phased adoption are provided.

As agile development has become mainstream, requirements are increasingly managed via issue tracking systems (ITSs). These systems provide a single point of access to the product and sprint backlogs, bugs, ideas, and tasks for the development team. ITSs do not clearly separate requirements from work items. We first tackle a knowledge problem concerning how requirements are formulated in ITSs, including their categoriziation and granularity, the presence of multiple requirements, and the existence of a motivation. Second, to assist practitioners in finding requirements in poorly organized ITSs without changing their way of working we investigate the potential of automated techniques for identifying and classifying requirements in backlog items. Through quantitative content analysis, we analyze 1636 product backlog items sampled from fourteen projects. To explore automated techniques for identifying requirements, we experiment with large language models (LLMs) due to their recent significance in NLP. The labeling of backlog items is largely inconsistent, and user-oriented functional requirements are the prevalent category. A backlog item often contains multiple requirements with different levels of granularity. The experiments with LLMs reveal that encoder-only models (BERT and RoBERTa) are most suitable for extracting and classifying requirements in backlog items compared to decoder-only models (Llama 3, Mistral 7B and ChatGPT with GPT 4). We reveal knowledge and patterns about requirements documentation in ITSs, leading to a better empirical understanding of Agile RE. The experimental results with LLMs provide the foundation for developing automated , unobtrusive tools that identify and classify requirements in ITSs. • We perform content analysis of 1636 product backlog items sampled from 14 projects. • Backlog items are labeled inconsistently, making locating requirements difficult. • We apply five LLM architectures for identifying and classifying requirements in backlog items. • We compare the performance of encoder-only and decoder-only neural architectures. • Encoder-only models are most suitable for locating requirements in backlog items.

This paper examines the automatic extraction of customer pain points from open reviews using the “Review to Pain Matrix” pipeline. The objective of this study is to develop a systematic approach for extracting customer pains from unstructured reviews, ensuring the reproducibility, transparency, and scalability of the analysis, while preserving the contextual usage and emotional valence of the statements. The relevance of this work is driven by the growing volume of user reviews and the need for product teams to rapidly focus on real customer issues without manual processing, which is constrained by subjectivity and the labor-intensive handling of large datasets. The novelty of the proposed “Review to Pain Matrix” workflow lies in the integration of product discovery and delivery stages within a single automated pipeline: from text cleaning and normalization, deduplication, thematization and aspect-level analysis, through to the generation of a canonical table of pain points with measurable attributes of frequency, intensity, impact on key metrics and effort required for resolution. Of particular importance is the ability to cyclically reprocess successive batches of reviews for prompt evaluation of product release effects and for updating marketing and service scripts. The outcome is a Pain Matrix: a machine- and human-readable matrix of verifiable pain-point formulations linked to audience segments and customer-journey stages, with priorities computed as a weighted combination of frequency, intensity, impact, and effort. This format enables the transformation of fragmented user feedback into a governed decision-making system, accelerates the conversion of insights into backlog items and experiments, and optimizes communication among product managers, designers, engineers, and front-office teams. This paper will be of interest to researchers and practitioners in product management, UX research, and data analytics.

The integration of Artificial Intelligence (AI) and Machine Learning (ML) into agile project management has ushered in transformative capabilities for sprint planning optimization. This review explores the role of AI-powered solutions in enhancing sprint planning through dynamic backlog prioritization and proactive risk mitigation. By leveraging supervised and unsupervised machine learning models, project teams can analyze historical sprint data, stakeholder feedback, and realtime project dynamics to predict task complexity, optimize resource allocation, and adapt backlog items based on shifting priorities. The paper systematically examines various ML algorithms, such as random forests, support vector machines, and neural networks, that support decision-making in backlog grooming and risk forecasting. Additionally, it evaluates AI-driven tools and platforms capable of automating sprint estimations, velocity tracking, and identifying potential blockers before they impact delivery timelines. Key challenges in data quality, model explainability, and integration with existing agile tools are also addressed. Through comparative analysis and case study synthesis, this review underscores the value of embedding AI intelligence into agile frameworks to drive efficiency, enhance team responsiveness, and reduce delivery risks. Ultimately, the paper provides a roadmap for researchers and practitioners aiming to implement intelligent sprint planning systems within modern software development lifecycles.

Lelang Ikan mobile application is an online auction in the marketplace platform of Pasar Iwak based on Android platform. Scrum framework is applied and consists of determining the product backlog, creating sprint planning and sprint backlogs, and conducting sprint reviews and sprint retrospectives. The product backlog resulted 14 backlog items based on the results of system and user requirements for user auctioneers. Sprint planning and sprint backlog are divided into four sprints, namely front-end and back-end development, system integration process and system implementation. Sprint reviews are carried out by implementing two types of testing, namely blackbox testing and user acceptance testing (UAT). Blackbox testing emphasizes testing application functions or features, while UAT is applied to measure the level of user acceptance. The results of blackbox testing showed that the features provided by the application are in accordance with the predetermined requirements. Whereas UAT showed the result of 66.8%, which means that the application is in the appropriate category and can be accepted by users. The application development process ends at the sprint retrospective stage which is a suggestion or feedback after the application testing. The suggestions obtained are in the form of adding tracking features, payment features with payment gateways, and application development with the iOS platform.

Agile Software Development Effort Estimation based on Product Backlog Items

ABSTRACTAgile Methodology requires specialized roles like Product Owners to act as the bridge between the business and the product delivery teams. This calls for specific skills such as a customer‐centric, design‐thinking mindset and a good vision of the overall system, its capabilities, and the ability to provide the right information at the right time.The primary duty of a product owner is to maintain a well‐refined, prioritized backlog of work items. Systems Engineers are enabled with the right methods and tools to perform solution architecture, design synthesis, system verification, and validation. Systems Engineers have visibility into the overall system, interfaces between sub‐systems, and external interface requirements. Enabled with the knowledge of Technical Processes, Systems Engineers are able to describe System Elements, their behavior, and interactions in the best possible detail. This allows them to ensure that functional and non‐functional backlog items are defined in unambiguous adequate detail.By applying their knowledge and expertise as mentioned above, Systems Engineers can effectively perform Product Owner duties and enables Agile teams to work efficiently to deliver the correct system increment at the end of each pre‐defined time box.

The need for information technology at this time has become an obligation for every company or organization. One that applies the development of information technology is ABC University. ABC University is a university that was founded in 2012 and along with the development of educational infrastructure every year, in 2020 ABC University has just completed the construction of an integrated laboratory building that functions as a place for research, learning, and academic and non-academic activities. This Integrated Laboratory will be used by 17 study programs at ABC University, more than 3000 students as well as lecturers and the academic community. With the limited number of rooms and laboratory assets for a large number of laboratory users, a system that functions to facilitate the use of laboratories is needed, starting from scheduling laboratory use, borrowing laboratory space, and borrowing laboratory equipment, Therefore, an information system needs to be developed in order to meet the needs of the ABC University Integrated Laboratory, namely the ABC University Integrated Laboratory Management Information System (MIS). The method used in developing this Integrated Laboratory SIM is the Scrum method which includes sprint planning, daily scrum, sprint review, and sprint retrospective. Based on 5 iterations during the development process of the Integrated Laboratory MIS, the total backlog items obtained were 65 backlog items. The ABC University Integrated Laboratory MIS has also been hosted on ABC University's servers and can be accessed via the https://labterpadu.itk.ac.id link.

This paper presents a Process Automation Framework (PAF) that systematically improves procurement efficiency and transparency from requisition to payment. Integrating workflow orchestration, rules-driven approvals, robotic process automation, process mining, and API-based ERP integration, the framework reduces cycle time, minimizes manual error, and creates end-to-end audit trails. PAF operationalizes standardized data models, supplier master governance, dynamic risk scoring, and automated three-way matching to curb maverick spend and strengthen policy compliance. It embeds privacy-by-design controls, role-based access, and cryptographically verifiable event logs to deter fraud and enable defensible audits. Analytics services expose real-time KPIs for on-time approvals, first-pass match rates, and cost per transaction, while digital supplier portals increase visibility into orders, deliveries, and payments. Interoperability is achieved through reusable integration patterns, canonical payloads, and low-code connectors, enabling incremental adoption without disruptive rip-and-replace. Methodologically, the framework follows a four-stage pathway: discovery, design, deploy, and de-risk. Discovery maps current processes using event logs, conformance checking, and stakeholder interviews to quantify waste and control gaps. Design translates policy into BPMN workflows and micro-service contracts with embedded controls and exception paths. Deploy provisions bot workers, secure connectors, and telemetry to capture ground-truth events. De-risk applies continuous monitoring, anomaly detection, and post-implementation reviews to refine controls and prioritize backlog items. Expected outcomes include twenty to forty percent cycle-time reduction, improved first-pass match rates, higher on-contract spend, and materially lower processing cost per purchase order and invoice. Transparency increases via immutable logs, accessible audit trails, and supplier-facing status dashboards for cost, compliance, and service-level performance. The framework advances sustainability and ethics by integrating sanctions screening and ESG attestation into automated onboarding and renewal. To scale, PAF defines capability maturity metrics, a change-management playbook, citizen-developer guardrails, and value dashboards that link automation benefits to financial and control outcomes. It offers a practical blueprint for procurement leaders seeking measurable efficiency, resilient compliance, and trustworthy transparency, bridging policy intent and operational execution through disciplined automation and evidence-driven oversight.

The article focuses on the application of decision support systems for prioritization of product backlog items in IT projects implemented using the Scrum methodology. The study identified the features of prioritization of different types of the product backlog items — user stories, epics and themes. It is justified that high-level product backlog items (epics and themes) require comprehensive prioritization, due to the following reasons. First, high-level product backlog items are particularly important because they determine the planning and implementation of detailed user stories within individual sprints. Second, any high-level item can be considered in terms of different criteria. Third, the implementation of epics and themes takes longer time compared to the implementation of user stories, so it is necessary to take into account possible future states of the project's environment. Fourth, prioritizing epics and themes requires increased objectivity and validity, so group decision making with participation of several experts seems reasonable. Taking into consideration the aforementioned features the conclusion regarding limitations of existing methods of prioritization is made. It is argued that prioritization of high-level product backlog items (epics and themes) may be performed using multi-criteria decision making methods with availability of several problem situations (possible future states of the environment), as well as involvement of several experts. The idea of applying decision support methods and systems is illustrated on the appropriate example. It is also argued that increased consumption of time and resources related with setting and solving decision support tasks may be considered as acceptable for high-level product backlog items.

The paper is basically dedicated to the problem of effort estimation for the Product Backlog items of IT projects led accordingly to the Scrum framework. The effort estimation issue is important, because low quality estimation decreases the efficiency of project implementation. The paper proposes an estimation method for the Product Backlog items of Scrum-based IT projects (which can be adapted also to other projects), which has two original elements with respect to the state of art in Scrum estimation: the usage of fuzzy numbers and strict rules for consensus forming, combined with a space for human interaction. The assumptions of the method should be complied with and were formulated on the basis of literature and authors experience. Two case studies were used for an initial method validation. The case studies confirmed a high potential of the method to increase estimation quality in Scrum-based projects, as well as in other project types. The case studies were conducted using research methods fulfilling the symmetry principle. The paper is thus an example of symmetry in management research.

Software development has been inevitable in the modern era. In the olden days, organisations followed intense traditional software process; but currently, the focus has turned greatly towards agile methodologies. In agile methodology, scrum is the mostly followed process. But it comes with a bunch of technical and generic issues. For instance, assigning, prioritising and integrating product backlog items prove to be difficult to deal with in agile methodology. On the other hand, it poses several other generic issues ranging from environment problems due to idle team participants and developer-tester issues. In this paper mainly concentrates on overcoming the technical issues mentioned above with the assistance of a framework which is perfectly refined in addition to the introduction of a new term called RScrum which is the extension of scrum which will greatly help to overcome the glitches.

Software development has been inevitable in the modern era. In the olden days, organisations followed intense traditional software process; but currently, the focus has turned greatly towards agile methodologies. In agile methodology, scrum is the mostly followed process. But it comes with a bunch of technical and generic issues. For instance, assigning, prioritising and integrating product backlog items prove to be difficult to deal with in agile methodology. On the other hand, it poses several other generic issues ranging from environment problems due to idle team participants and developer-tester issues. In this paper mainly concentrates on overcoming the technical issues mentioned above with the assistance of a framework which is perfectly refined in addition to the introduction of a new term called RScrum which is the extension of scrum which will greatly help to overcome the glitches.

Abstract Scrum teams aim to deliver products productively with highest possible value and quality, so they try to deliver high priority and high value product backlog items in earlier sprints. Making size estimation of product backlog items correctly is one of the most prominent factors for effective sprint planning. Retrospective meetings are an opportunity for teams to improve product quality, their productivity, and estimation capability. Enhancing in those areas requires empiricism as agility requires; hence, measureable indicators should be inspected and adapted at regular intervals. In this study, we assessed how and what kind of historical data is required to be collected for monitoring, and how statistical analysis can be investigated for inspection and adaptation in retrospective meetings. We experimented that statistics of “Correlation between Story Point and Actual Effort” and “Consistency of Relative Estimation” were convenient for inspection and adaptation of estimation capability of teams. Past retrospective meetings also showed that statistics of “Team's Actual Effort on Product,” “Team Velocity,” “Actual Effort for One Story Point,” “Innovation Rate,” and “Velocity vs Unplanned Effort Rate” were helpful to control and increase the productivity of teams. “Actual Effort Rate of Quality Activities” and “Subcomponent Defect Density” statistical results helped to enhance product quality.

The order in which you send your backlog items into construction determines when stakeholders can reap benefits from each piece of functionality. This can substantially impact market timing, enterprise earnings, and the project manager survival rate. There are several ways to order a backlog, and sophisticated methods and tools exist to do so-for example, during release planning. But no matter what backlog-ordering scheme you use, you ought to be explicit on the order in which you realize potential business value. To that end, researchers have developed methods to express business value relative to cost in your backlog. They also have devised methods to monitor how much potential business value you're realizing along the way-in addition to the cost expended.

Improving task breakdown comprehensiveness in agile projects with an Interaction Room