Case 1: Nisha's AI Team

How a chartered accountant used Claude and ChatGPT to plan a book, and where human judgment mattered

Panaversity case study · November 2026

Where this case fits

This is the fifth section of Claude and ChatGPT 2.0, and the first case study in the Claude and ChatGPT 2.0 series. The first four sections explained the ideas. This case shows them at work in one real evening.

• The brief. The New Way of Working showed that every job starts with a brief. Watch how Nisha's first prompt works as a brief, and what she had to add later.
• The AI Worker. Tour of the Apps and Behind the Screen showed what Claude and ChatGPT can do by themselves. Here, one AI Worker writes and grades its own work, and a second one reviews it.
• The second signature. Staying safe showed that a person makes the final decision. In this case, Nisha is that person at every step.
• Saving work somewhere lasting. The borrowed desk showed that the desk can be cleared. Nisha saved each version as a file and kept a diary of the whole evening.
• The architecture. The Architecture explained how the apps run. You do not need it to follow this case.

This case stays at the first step, a helper used by a person who is present. The lab that follows takes the same work up the ladder.

About this case

This is a fictional teaching case, prepared for class discussion. It is based on a real practical exercise carried out on November 8, 2026, with Claude (claude.ai) and ChatGPT (chatgpt.com).

Made up for teaching: Nisha, her background, her money (including the subscription costs), her conversations and her thoughts. These come from the class scenario.

From the practical: the prompts, AI replies, grades and drafts. They are recorded in the practical log, the six draft files and ChatGPT's review file. Word counts and save times come from those files. The final book file (version 6), with its full list of 15 sources, is in the resources of this case.¹

The case follows the style of a Harvard Business School case but is not connected to Harvard. It does not say whether Nisha's choices were right or wrong.

How this case is organized

The story comes first, in four parts. Read them in order.

1. Nisha's problem: who Nisha is, why she wants to write a book, and how she sets up her AI team.
2. Steps 1 to 3: Claude writes and grades its own work
3. Steps 4 to 6: a second opinion and the human check
4. A company of one: what Nisha's diary showed, and what she began to imagine.

Then come the questions for discussion and the exhibits, which give the details: the method, the prompts, every grade, every version, who caught what, the fact check, the final table of contents, the time and cost, and the words to know.

After the discussion, the lab lets you run Nisha's method again for one chapter, with a project, an overnight draft and a second signature.

For instructors, the teaching note gives a discussion plan. The practical log records every prompt and reply, so a class can run the exercise again.