Software Engineering

From the recent literature review, it is evident that existing agile methodology lacks the method to evaluate the requirement understanding of agile team members for a given set of requirement chosen for agile software development. Hence, there is a need to introduce a requirement understanding check to ensure every agile team member follows the given requirement clearly without any ambiguity. To fill this existing gap, this research paper proposes to extend the usage of story cards to evaluate the understanding of the given requirement and to highlight any challenges and risks in the early stage of requirement understanding under agile software development methodology, if any. This paper primarily focuses to introduce a robust requirement understanding evaluation process in agile methodology. The research results were found to be motivating and were analyzed by comparing the data-points using time-series for performing agile query analysis, agile team velocity analysis and agile team involvement analysis for two agile teams where one team delivered the sprint output using agile traditional method while another team opted for proposed approach. A considerable decrease of 33.07% was observed in the number of queried raised and a significant increase of 26.36% in agile velocity was observed for agile sprint under proposed approach when compared to agile traditional approach. Also, a significant shift from 40%-80% team involvement under traditional agile method was uplifted to 80%-90% team involvement under proposed approach.

In this paper, we have extended our literature survey with experimental implementation. Analyzing numerous Artificial Intelligence (AI) techniques in software engineering (SE) can help understand the field better; the outcomes will be more effective when used with it. Our manuscript shows various AI-based algorithms that include Machine learning techniques (ML), Artificial Neural Networks (ANN), Deep Neural Networks (DNN) and Convolutional Neural Networks (CNN), Natural Language Processing (NLP), Genetic Algorithms (GA) applications. Software testing using Ant Colony Optimization (ACO) approach, predicting software maintainability with Group Method of Data Handling (GMDH), Probabilistic Neural Network (PNN), and Software production with time series analysis technique. Furthermore, data is the fuel for AI-based model testing and validation techniques. We have also used NASA dataset promise repository in our script. There are various applications of AI in SE, and we have experimentally demonstrated one among them, i.e., software defect prediction using AI-based techniques. Moreover, the expected future trends have also been mentioned; these are some significant contributions to the research

Background : Artificial Intelligence (AI) has recently emerged as a transformative innovation within the software industry, disrupting conventional approaches to application development by automating tasks, refining code, and enhancing resource efficiency. Prior research indicates the effectiveness of AI-powered tools across various domains. However, contemporary studies lack a detailed analysis of the diverse sectors utilizing AI tools for software development. Objective : This article aims to identify the potential benefits and impacts of AI in software development, specifically regarding time-to-market, productivity, code quality, bug-fixing rates, resource flexibility, and developer satisfaction. The goal is to present fact-based information about AI’s impact on multiple industries and scopes of work. Methods : A mixed-methods research design was employed to analyze quantitative data from 40 projects across healthcare, financial services, retail, technology, and e-commerce industries. Data were collected using various project management tools, automated testing environments, and online questionnaires addressed to developers. The study incorporated a comparative evaluation of AI-based projects and traditional projects, with statistical analysis. Results : AI-driven software development projects demonstrated a mean reduction in time-to-market by 34.6%, an improvement in code quality by 70%, and a mean reduction in bug-fixing time by 57.7%. Productivity per sprint increased by over 70%, resource flexibility was higher (90.2% in AI projects vs. 67.8% in traditional projects), and developers reported higher satisfaction levels. These findings reinforce the concept that AI significantly enhances workflow and the achievement of optimal results. Conclusion : AI substantially improves both the speed and quality of software development. Further research should expand to explore the experiences of different sectors, the application of AI-driven tools, their differentiation, and usage, as well as the ethical considerations to promote sustainable and innovative software engineering solutions.