It reminds me of a previous life, at a previous time, when everybody reinvented the content management system based on slightly different approaches, because they thought existing CMSs were too big. It starts with serving pages, then you want multiple templates, then you want accessibility, then you want to authenticate with LDAP, then you want to use the LDAP groups as your groups, and then you need to tie groups to roles, because groups mean one thing to the org, and roles mean another just to the CMS, then you realize a relational database isn't that good a match, then you realize it would be convenient to attach some logic, and a document database is no longer a great idea, then you need multiple datastores for content, then you need new workflows on top, then...
The devil is in the corner cases. They have been solved in different ways, by different parties, for decades, not always documented ("because it was just a simple thing, that once, and we fixed it"). Unless you've been there, you are doomed to make the same mistakes, and nobody was everywhere to avoid all mistakes.
Why not dedicate an effort to migrate all the COBOL code currently running almost every banking transaction? That should be easy ;-)