A lot of large companies, especially those that operate semi-autonomous business units use internal accounting to handle transactions between these units. So a project manager might have a budget of 1M and he purchases marketing services from the mkting dept, engineering from the dev group, etc. But obviously money is not actually moved around...
You can see how a private intra-company blockchain could be very useful here, for auditing and real-time tracking purposes. For external services, you would even "pay" CoCoins to the financial dept. and they would cut and mail a check externally.
This BNY effort is probably the first step towards a consulting business setting this up for companies.
EDIT: yes you have a great point about gold. And hopefully the same will happen with diamonds (grown in a lab). Sometimes I think that the best way to handle some of these areas with horrible human rights records is to leave them alone. But, as we learned with conflict diamonds, realistically and unfortunately that only works if they don't have anything the first world wants.
the real question is why that internal accounting wouldn't continue to be more efficient and even more secure using a traditional SQL database? are they really having any internal accounting fraud to incentive them to set up an entirely foreign means of accounting that could be subject to unknown types of hacks from the outside or even inside? what drives security of this model? can't be POW. w/o POW why couldn't it be gamed?
This is a good question. I will ask around. We need to first divide the problem into 2 systems:
1. Internal accounting for an external client. Lawyers and many government sponsored projects need to carefully account for their time, for example. The ability to track all the coins backwards lets you do that in a way that can't be retroactively "massaged"
2. Straight internal accounting.
I can make some guesses right now though... the point of keeping the actual $ away from the internal txns is to eliminate internal fraud. However, there can be lots of other non-fraud issues -- like why is it always such a shocker when projects go over budget? Examination of spending on an internal blockchain can show these trends before the annual monster audit.
In terms of security, there are lots of choices. The most likely candidate would be where everyone runs SPV wallets, and the finance department runs a few trusted miner nodes. We don't need to solve trustless consensus for this application -- the finance dept is not going to be 51%ing their own network because no mining reward, and blacklisting addresses is essentially a feature (if someone gets fired their CoCoin is automatically reaped for example). What they really need is a database with immutable per-record signatures tied to company employees.
Difficulty is "nice to have" because it makes it very hard in practice to rewrite the blockchain with a fake one starting from six months ago (say). But even if the finance dept bought 100x its baseline mining capacity to do some massive 6 month ago fork of the blockchain to a completely fraudulent record, it would take massive company-wide participation in the fraud since everyone would have to surrender their private keys.