One of the strange emerging factors of the current world is that higher stakes, larger dollar, fully custom projects tend to be messier and more disorganized than regular common transactions.
It makes sense, when you think about it -- a credit card company with millions of users can streamline, test, and design an interface to make their online payment and review near perfect. There's many competitors, it's easy to benchmark against what works and doesn't, and many people will use it.
As project scopes and dollar amounts get larger, the whole thing gets murkier. There's more ambiguity -- large custom projects, by definition, have unique elements to them -- but also less benchmarking (because people tend to be more confidential and not share their documentation, technology, and workflows) and finally, the scale just isn't there to invest heavily in to automate away small amounts of time and friction.
Thus, it becomes really easy to apply for and receive a basic credit card online, buy some ecommerce product online, and then link up your online banking to review to pay your credit card online smoothly in a nice interface.
But transactions in the hundreds of thousands of dollars or more are handled with a mix of physical paperwork which is custom and must be reviewed manually by multiple parties, and then a much less descriptive and less user friendly wire transfer.
The same thing applies to workflows, project support materials, and so on. Some of the world's most important documents and processes are somewhat crude and dated, because due to confidentiality, they're only seen by a few people and not given broad exposure for feedback, review, benchmarking, and iteration.
Organizations that need very rapid response times to complex problems will often build their own custom tools, and some of those might be quite good. But if a delay is acceptable, then it's nearly guaranteed that the lowest quality automation, technology, and workflows belong to some of the most important projects.
An odd state of affairs, to say the least.