Full text retrieval and semantic analysis software are useful software applications in the legal market. Software developers are making available increasingly sophisticated techniques to work with in very large document collections to identify potentially relevant ones and classify or otherwise automatically analyze them. These tools are very useful in knowledge management and litigation support applications, but law firms using them or considering doing so must keep in mind the need for human intervention and review.

Yesterday, the Wall Street Journal carried an article on page B1 titled “Online Job Hunting Is Tough. Just Ask Vinnie.” The article is about how large companies use software to screen resumes for job applications submitted on their Web sites (and that for some companies, the majority of their hiring is via the Web). A consulting firm tested the Web-based job application process by inventing a fictional job applicant named Vinnie Boombotz and sent the resume to 400 big companies. The resume was “patently ridiculous,” containing many misnomers but with some “plausible” work experience as well. To make a long – and very amusing – story short, the made-up resume made it through quite a few automated screens, whereupon a human rejected it.

I claim no expertise on how automated resume screening software works, but the challenge with screening thousands of resumes is similar to the challenges of identifying useful work product or discovery documents among thousands or more. The article ribs the companies where the resumes passed the screen. I agree, however, with a quoted spokesman of one company. He argues the technology is doing its job and that the humans did theirs, saying that’s the way the system is supposed to work. I think the same is true for the application of technology in KM or litigation support. It is helpful, arguably essential, to screen documents and automatically do a first-pass analysis, but that does not eliminate the need for review by people who know what’s going on.