Incorporating prompt engineering into the accounting curriculum

The rapid emergence of generative artificial intelligence (AI) tools, including large language models (LLMs) such as those used by ChatGPT, has begun to transform work across industries, including the accounting profession. As these tools become increasingly embedded in practice, a new digital competency has become vital to learn: prompt engineering.

Prompt engineering is the process of crafting effective instructions, or “prompts,” you can input into LLMs to produce relevant and reliable outputs.

Because LLMs are highly sensitive to the phrasing, context, and clarity of user inputs, the quality of a prompt directly affects the structure, relevance, and accuracy of the output (see “GenAI for Accountants: 10 Prompt-writing Tips,” JofA, April 1, 2024). As a result, prompt engineering is an increasingly important digital communication skill in professional practice.

The Big Four accounting firms are investing heavily in AI capabilities and prompt literacy. PwC, for example, has committed more than $1 billion to AI and cloud infrastructure, including workforce training programs that emphasize prompt engineering as a core skill. Deloitte’s AI Academy features modules dedicated to effective prompting, while EY has rolled out its internal generative AI initiative — powered by the EY.ai platform and its proprietary LLM, EYQ — to 400,000 employees across 150 countries.

As accounting educators, we need to prepare students for this evolving environment. Prompt engineering not only prepares students to use LLM-based generative AI tools effectively but also reinforces essential competencies such as professional skepticism, ethical reasoning, and critical thinking. By integrating prompt engineering into accounting curriculum, we prepare our students to become not just users of AI tools but also informed collaborators who can engage with these systems thoughtfully, assess their outputs critically, and understand their limitations.

Popular generative AI tools (for education) based on LLMs:

• ChatGPT, developed by OpenAI, is one of the most popular tools. Its educational version, ChatGPT Edu, is a secure platform specifically designed for universities.
• Microsoft Copilot, integrated into Microsoft 365 applications such as Word, Excel, and PowerPoint, brings AI-powered support to everyday academic and business tasks. It is available to academic institutions through Microsoft’s education programs.
• Claude, developed by Anthropic, is valued especially for its ability to summarize lengthy documents and provide thorough responses to open-ended questions. Anthropic now offers Claude for Education, a free version for universities.
• Google Gemini (formerly Bard) is closely integrated with Google Workspace tools such as Docs, Sheets, and Search. U.S. college students currently have free access to Gemini Advanced through Google’s educational offerings.

As generative AI tools continue to evolve, exposing students to multiple platforms helps them recognize the importance of prompt design across different interfaces and equips them to engage critically and effectively with AI in varied professional settings.

Resources for teaching prompt engineering

Technical guides:

• OpenAI’s Prompt Engineering guide is a practical resource that teaches users how to write effective prompts. Though this is designed for ChatGPT, the ideas work for all LLM-powered generative AI applications. Instructors can easily substitute the examples provided with accounting context to guide students.
• Learn to Use Microsoft Copilot is a tutorial by Microsoft that guides users through Copilot.
• Anthropic’s Claude prompting engineering overview is best for Claude-specific prompt structures and safety-aware instructions.
• Google’s Gemini Prompting Guide is best for Gemini users and those integrating with Google apps.

Also available are other high-quality resources, such as LearnPrompting.org and YouTube videos about how LLMs work.

Resources from Big 4 accounting firms

• KPMG’s University Connection offers all educators access to course materials in various accounting topics, including AI. Under “AI” in the site’s “Topics” field, there are several case studies including: Accounting Research Using Generative Artificial Intelligence, GenAI in the Audit, Generative AI Use Cases, and Forensic Case Study. Those cases introduce basic AI terminology, prompt engineering, and how generative AI can be used in audit.
• EY’s Academic Resource Center provides educational materials, including support for AI curricula. Under AI curriculum, you can find a generative AI-enhanced professionalism framework, a generative AI competency matrix, and a generative AI competency curricula map. The AI curricula also include generative AI case studies, generative AI ethics case studies, and presentation and video studies.
• PWC’s faculty resources site lists recordings and other AI resources. For example, in April 2025, PWC hosted “Understanding AI with PwC,” which showcased materials being used to upskill PwC employees.
• The Deloitte Foundation offers resources and curricula for the classroom, including AI ethical case studies that simulate real-world scenarios.

Challenges and considerations

Despite its benefits, incorporating prompt engineering into the accounting curriculum presents several challenges.

First, the technology is evolving quickly, and educators may feel uncertain about their own familiarity with AI tools. Faculty development and institutional support will be critical. Second, LLM-powered apps like ChatGPT can produce inaccurate or incomplete outputs, which requires instructors to guide students in critically evaluating AI-generated responses.

Students need to understand that even effective prompting will not substitute for core accounting and auditing knowledge. In using LLMs, AI skepticism and fact verification are important.

Academic integrity is also a concern, as students may use AI to complete assignments. Clear guidelines about ethical use can mitigate this risk.

— Jian (Jane) Zhang, Ph.D., is a professor of accounting at San José State University. To comment on this article or to suggest an idea for another article, contact Jeff Drew at Jeff.Drew@aicpa-cima.com.

MEMBER RESOURCES

Articles

“5 Advanced Ways to Use ChatGPT,” JofA, June 1, 2025

JofA Artificial Intelligence (AI) Coverage

CPE self-studyEthics in the World of AI: An Accountant’s Guide to Managing the Risks