Students love using GenAI to see its capabilities and answer accounting and business questions!  

I taught GenAI this past semester using these five steps:

  1. I had students do 66 accounting and business analytics labs in Excel, Tableau and Power BI (from my analytics texts) early on in the semester.  They quickly embraced the AMPS model and the process of going from “Asking the Question” to “Mastering the Data” to “Performing the Analysis to finally “Sharing the Story” for many different accounting and business questions.

  2. Fall semester 2025, I asked the students to learn to use GenAI to perform five of the same labs, choosing from the 66 labs they had just worked on.  To do so, they used my GenAI labs (with prompts and videos) at each book’s instructor resources and my website at RichlyAnalytic.com (and they will all be assignable within Connect by April 2026).
    I specially asked them how I could make the GenAI labs better, and if they were able to get the same answer as they got using Excel, Tableau or Power BI they had done earlier in the semester.

  3. I then asked the students to create their own GenAI lab with prompts and a video using a question of their own making.  

  4. After writing up their own lab, we had other students evaluate their GenAI labs and videos and make suggestions. 

  5. Finally, after receiving feedback from other students, the students then refined their own lab prompts and videos and turned it in for grading.  It was subsequently placed on the AccountinGenAI.comwebsite to share with others.  Take a look and let us know what you think.

What did the students learn?  Here’s five comments:

Student Comment #1: One of the challenges I faced was figuring out how to write prompts that ChatGPT could understand. If I wasn’t specific enough, it sometimes gave answers that weren’t exactly right. This taught me the importance of being clear and detailed when asking questions or giving instructions, just like in school or any project.

Student Comment #2: Unlike Excel, where errors are often a direct result of user formulas, missteps in ChatGPT were likely from broad prompts. This experience emphasized the critical role of human input in guiding AI tools effectively.

Student Comment #3: The other conclusion that I came to after this project has been that AI, like a calculator, is only as good as the information that you provide it with.

Student Comment #4: ChatGPT always responds confidently, whether the answer is correct or not. This makes it crucial to double-check the results, especially for technical or high-stakes tasks.

Student Comment #5: This project has been very eye opening for me, because I think that previously I had been under the impression that AI, especially the paid version, would automatically get it right when it comes to doing the kind of work needed. However, I saw first-hand how even a premium generative language model could misinterpret the information provided.

Well worth the extra effort!  The students absolutely loved learning how to use GenAI.