Over a year later and I still do not understand what the use case for this is.

A lot of the examples/documentation that was made by Microsoft for this seems to focus on data analysis and data visualization. Anyone in those fields would probably prefer to get the data out of Excel and into their tool/pipeline of choice instead of running their Python code in Excel. That also makes the big assumption that the data being used is fully contained within the Excel file and that the libraries used within the code are avalaible in Excel (including the library version).

For anyone looking to learn/use Excel better, I doubt the best use of their time is learning a new programming language and how Excel implements that programming language. They would likely be better off learning Excel's formulas, pivot tables, charts, etc. They could even learn Power Query to take things to another level.

For anyone looking to learn Python, this is absolutely a terrible way to do so. For example, it abstracts away library maintenance, could provide modified error messages, and makes the developer feedback loop more complicated.

If you want to automate Excel then this realistically allows for very little new functionality that did not exist prior to this feature. Using other Python libraries like OpenPyxl and xlWings will still be required to automate Excel.

I am sure there are edge cases where this iteration of Python in Excel is perfect. However, this feels like a checkbox filler ("yeah, Excel supports Python now") than an implementation of an actual useful feature. A fully featured and supported Python library that manipulates Excel/Excel files would have been a much more exciting and useful feature - even if it had to be executed outside of Excel, like OpenPyxl.