Artificial Intelligence Techniques to Enhance Cost, Effort, and Schedule Estimates in Software Development Projects: A Systematic Literature Review
TL;DR
The findings indicate that techniques such as artificial neural networks, optimization algorithms, machine learning models, and hybrid approaches consistently yield improvements in estimation accuracy, with average error reductions reported in the literature ranging approximately from 15% to 30% when compared with traditional methods.
Abstract
The inaccuracy of cost, effort, and schedule estimates remains one of the primary factors associated with failures in software development projects, particularly in contexts characterized by high complexity and frequent changes in requirements. Given the well‐documented limitations of traditional estimation methods, this study aims to systematically analyze how artificial intelligence (AI) techniques have been applied to improve the accuracy of such estimates in software projects. To this end, a rigorous systematic literature review (SLR) was conducted, structured according to the PICOC protocol and established guidelines for systematic reviews, encompassing searches in the IEEE Digital Library, ACM Digital Library, SpringerLink, and ScienceDirect. In total, 108 primary studies published between 2015 and 2025 were analyzed, selected based on predefined inclusion and exclusion criteria as well as methodological quality assessment. The findings indicate that techniques such as artificial neural networks, optimization algorithms, machine learning models, and hybrid approaches consistently yield improvements in estimation accuracy, with average error reductions reported in the literature ranging approximately from 15% to 30% when compared with traditional methods. The reviewed studies also highlight challenges related to data quality and availability, model reproducibility, and the feasibility of deploying these approaches in real‐world environments. As a contribution, this SLR provides a structured synthesis of the state of the art, identifies research gaps, and offers valuable insights for both the academic community and industry practitioners in the development of more accurate and reliable estimation models and tools.