The Development of a RAG-Based Artificial Intelligence Research Assistant (AIRA)

Abstract

ABSTRACT The increasing volume of accounting research presents challenges in efficiently navigating and synthesizing information. We introduce AIRA—an artificial intelligence (AI)-based application developed to assist scholars in consolidating accounting research papers using generative AI models. Traditional literature review methods are time-consuming. AIRA streamlines this process with natural language prompting. To ensure the application’s reliability and usefulness, we use the design science methodology to validate it meets the designed objectives. Overall, faculty found AIRA achieves its objectives of being useful and easy to use, and they plan to continue using it in the future. A large sample of employees from the Securities and Exchange Commission (SEC) also believed the tool would be useful for their work. In examining the first 500 prompts that were entered by anonymous users, we find that it is used for literature search and retrieval, summarization of research findings, and literature review creation, among other tasks.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.