Blockchains aren't just for currency, but their best use case is surely currency, for only the currency project can recuperate the massive externalities (costs) of running the system, for the profit margin is almost 100 percent. What that means is, for a blockchain to make sense for you, you really, really need to need the anti attacker parts of it for it to make sense over any other system.
Current real world uses are timestamping, (recently Peter Todd timestamped all of archive.org, so people can't go back and make changes to fool people.) Bitcoin (currency). And as a tier 1 trust layer to support tier 2 protocols that have faster, cheaper, more stateful transactions (counterparty, mastercoin, rsk, sidechains, lightning, etc. a few of those need opcodes added to the btc network.)
Trust is expensive, bankers hours are comedic, much of the world loses large percentage of currency value every year, bitcoin can solve much of that. Databases and merkletrees already do a great job if you don't have advanced attackers. Blockchains are much less exciting than bitcoin.