Software that recognises a name even if it varies from different information sources is becoming a vital tool, used by law enforcement, border control agencies and even financial institutions.
For example, one database can have a surname spelled Rodgers, while another has it as Rogers.
The problem gets much worse when dealing with foreign names, as language becomes a factor.
In an interview at IBM's Information On Demand (IOD) conference in Las Vegas, the company's manager of linguistics Frankie Patman spoke about the importance of name recognition software.
She referred back to 1993, when an investigation to find Mir Aimal Kansi for shooting three CIA agents was hampered because his name was spelled differently on his passport and visa.
"In some records it was Kansi. In others it was Kazi. Our technology played a part in finally figuring out that those two names were the same," she said.
Financial potential
So far, the biggest clients for such global name recognition technology were in intelligence, law enforcement and border control, but it was increasingly being taken up by financial institutions.
This was because this can reduce the chances of fraud and money laundering. She also highlighted healthcare as an area where it could be useful.
The software is also of benefit when two companies merge and need to combine separate sets of information together.
"To find the value in those merged sets of information, you need to make sure that when you refer to one person you find everything that's in there about them," Patman said.
"Our technology will allow you to find more of what you're looking for, and fewer distracting non-relevant records. So it's a huge asset for merging data," she added.
