Banking at its core is the process of matching illiquid loans against liquid deposits. Blockchain is very focused on tracking individual items, perhaps loans for example; probably less on tracking individual deposits.
Yet this process of matching illiquid loans against liquid deposits does not start by matching loans and deposits individually; one has to assess what the problem looks like in total.
Blockchain adds nearly nothing to solving this problem, and in some measure because of the complexity of the data structures involved, it might in fact detract significantly from solutions that are in place.
The systems that solve this problem are the finance, risk, and other systems at the core of the bank. The objectives of blockchain are very applicable there–shared data, reduced reconciliation–but something quite different from today’s blockchain is needed.