Some vendors use multiple levels of automatic analysis, while others narrow down the number of documents likely to need redaction and then rely on human intervention to help improve the software's automatic redaction capabilities.

"It's a new technology, but a proven technology," said Paul Miller, president of Aptitude Solutions Inc. in Casselberry, Fla. Miller said Aptitude's aiRedact software looks for specific numbers, words or combinations of related words, such as "account number" or "Social Security number."

On big jobs involving millions of document images, several thousand pages are culled and manually analyzed by a worker who can verify that data should be redacted, Miller said. The software then automatically adjusts to redact the remaining records based on the manual choices. It typically costs between US$200,000 and $300,000, he said.

ImageTech Systems Inc. in Camp Hill, Pa., has built a plug-in redaction module for Kofax Ascent Capture, a tool from Kofax Image Products Inc. in Irvine, Calif., that finds data in documents and forms.

R.J. Oommen, ImageTech's principal, said the plug-in module uses several methods, including on-the-fly input from users, automatic processing of data in standard forms and an intelligent algorithm.The module starts at $5,000, but the total cost can exceed $100,000, Oommen said.