Databases don't work with names well because it's really freaking hard even for a human sometimes. People have funky spellings, middle initials, sometimes full middle names, phone operators mis-type or mis-hear names, or assume a different spelling. People have accents which messes up people hearing names and bad handwriting which messes up people reading names. People have names that look like words that should be cleaned from names like "Mr. John Trust". While "Trust" should usually be cleaned from a list of bank trusts, it can't be cleaned here. Some people do last name first. You want to add another language and address standard in there, good freaking luck.
As far as I know, doing data processing as a job, if you want good results, you MUST look at a very small percentage of an average database manually (maybe 0.01%-1% depending on how dirty your DB is). But for any global company, that could still be tens of thousands of checks you have to make and it will be very time consuming.