I am going to assume PK means "Primary Key".
Data segregation is done at the Company level as you have noted (112: Row Level Access), where each Company has an associated entry in the "Group" form.
The "Group ID" is what is set in field 112 to control the data access, not the actual Company name.
Although you may be able to change the Company name directly in the Company form, if this does not update any related data you may end up with orphaned or "corrupt' data as you put it.
I would always recommend using the Data Wizard as it updates all related data across the application and related data.