I can't help wondering what any company that suddenly finds itself in need of digital transformation has otherwise been doing over the last 20+ years?
Managing change incrementally with (surprise, surprise) digital bolt ons. Nobody is ever in a hurry to really shake up a company's core processes and systems, so when customers want a way of seeing their bill on line, that's not achieved by a new system, but by a quick and dirty kludge to interface between the customer and the CRM, linked via the company's website (originally designed largely as a marketing toy). Then, when they want to pay bills on line, that becomes a different kludge, that has to tie between the online bill, the CRM, and the customer accounting element of the back office systems. Then it turns out customers want to do this from phones, so these two kludges get replicated in Android and Apple versions, which then makes the company a software publisher and supporter. By now, even in this very high level approximation we've got six different software packages all having to interface with business critical systems, all with different security models, and possibly PCI DSS implications. If the customer pays by direct debit, and you want them to be able to change that, there's another three packages. Changing their personal details, another three, and so it goes on. Each element needs the full set of sign off, business case, business analysis, design, code test, release, bugfix updates. Now consider the movement in third part systems eg your web development platform such as Adobe, MS or whoever doesn't stay still, so you need to update your web apps for whatever the vendor supports. Then you've got the PITA lifecycling of the main ERP, and so it goes on.
All these companies hoping to be transformed are starting from systems architecture of immense complexity built over time, almost none of which was actually designed. So in terms of "what have they been doing for twenty years", the answer is "spinning plates".