Banks are generally very careful about which new feature they would deploy into production. In contrast to the crypto project, which iterates by deploying changes to see whether it works or not, banks cannot allow themselves any failures — applications need to work from the first time. There are many examples where crypto platforms were hacked and needed to roll back the chain. Banks cannot afford something like that; their reputation would be ruined. 

Banks thinking about applying blockchain technology should find a very simple use case and start from there. Once they gain more experience, they can add new features in a step-by-step approach, building confidence and complexity in parallel. For example, a bank can create a digital version of an investment fund, then allow clients to withdraw to their own whitelisted wallets, afterwards automate distributions, then connect legacy cash accounts for subscription and redemption process, etc.

Another piece of advice would be to avoid vendor lock-ins and stay flexible with the technology. Quickly changing and new innovative projects are coming every year. There is still a debate about public vs. private blockchains, proof-of-work vs. proof-of-stake, etc. It will take some time until the technology fully matures. However, there are already tools, which are abstracting away complexity of so many different chains, allowing companies to build multi-chain applications and remain flexible.