Prompts Blend Requirements and Solutions: From Intent to Implementation

TL;DR

This paper redefines prompts in AI coding assistants as evolving requirement artifacts, proposing a model that decomposes prompts into components related to requirements and solutions, aiming to improve prompt engineering and software development practices.

Contribution

It introduces a conceptual model for prompts as requirement artifacts, assesses this model with existing prompts, and formulates hypotheses for empirical evaluation to enhance prompt engineering.

Findings

Prompts evolve toward increased specificity over time

User characteristics influence prompt evolution

Engagement in prompt refinement improves code quality

Abstract

AI coding assistants are reshaping software development by shifting focus from writing code to formulating prompts. In chat-focused approaches such as vibe coding, prompts become the primary arbiter between human intent and executable software. While Requirements Engineering (RE) emphasizes capturing, validating, and evolving requirements, current prompting practices remain informal and adhoc. We argue that prompts should be understood as lightweight, evolving requirement artifacts that blend requirements with solution guidance. We propose a conceptual model decomposing prompts into three interrelated components: Functionality and Quality (the requirement), General Solutions (architectural strategy and technology choices) and Specific Solutions (implementation-level constraints). We assess this model using existing prompts, examining how these components manifest in practice. Based on this model and the initial assessment, we formulate four hypotheses: prompts evolve toward specificity, evolution varies by user characteristics, engineers using prompting engage in increased requirement validation and verification, and progressive prompt refinement yields higher code quality. Our vision is to empirically evaluate these hypotheses through analysis of real-world AI-assisted development, with datasets, corpus analysis, and controlled experiments, ultimately deriving best practices for requirements-aware prompt engineering. By rethinking prompts through the lens of RE, we position prompting not merely as a technical skill, but as a central concern for software engineering's future.