"It's as easy to get online as to dial a call"
By Pankaj Maru
Mumbai, Mar 18, 2008

Matthew Siegelman, CEO of Burning Glass, in an e-interview with Pankaj Maru, talks about AI, his Indian experience and the new job portal, whereismyboss.com (WIMB)

HR is not somewhere you would normally expect to use Artificial Intelligence (AI). Probably, this is the first instance where AI has been applied in real life situation involving people. Your views on this.

Burning Glass was founded in order to bring to HR, the kind of analytical rigor and the insight regularly applied in other disciplines. Many of us come from the financial services industry, which has long been at the forefront of decision automation. By contrast, HR has always depended on qualitative decisions and intuitive judgment. Time and again we were told that it's simply not possible to apply artificial intelligence to HR because these are human decisions requiring intuition.

But, what is intuition and brain's awesome ability to discern patterns? We built Lens in order to replicate the same kind of intuition that people use in assessing choices. A real recruiter has seen hundreds or thousands of placements happen, so he/she knows what kind of person an employer is likely to favor and what kind of person just won't cut it, even when their skills look very similar on paper. Lens can tell the difference because it is applying the same kind of pattern-based judgment.

What has been your experience working with a job portal?

The Indian recruitment landscape is ripe for change. With so many job seekers and job opportunities, there are tremendous inefficiencies in the way the two sides find one another. An employer may put an ad on a job board, receive thousands of responses, even then only a handful may be suitable for the post. Even if you disregard the cost of reviewing all of those irrelevant applications, how do you go about finding the needle in the haystack? At the same time, it's a bad process for job seekers as well. With WIMB we want to make sure that recruiters reach the most suitable candidates.

With Lens, job seekers can upload their resumes automatically without the need to pre-fill a registration form, and will be directed to and notified of jobs that fit them well. Similarly, when an employer posts a vacancy, Lens will immediately identify the most relevant talent from the WIMB database, the people who look similar to those who have gotten the same kind of job in the past.

WIMB offers live-web interviews, but most part of India does not have high speed Internet. What's the future of such innovations in India?

There's a corollary of Moore's Law that seems to apply to connectivity. In 1998, I lived in Mumbai, and it was nearly impossible to connect to the Internet. I used to fax letters back and forth with my colleagues in the US. Today, it's as easy to get online as it is to dial a call. Now in India, when a plane lands, half the passengers whip out their Blackberries, and start checking email; even six months ago, that wasn't the case. It is true that many people don't have the level of connectivity at home currently, which will allow them to participate in online interviews, but that will catch up soon.

Frankly, leading edge services are important not only because of the value they create for users, but also because they stimulate consumer demand. If there are all of these great facilities available to me if I have broadband Internet at home, then I will demand it and then it's only a matter of months before the telecom providers deliver it affordably. Until then, a trip to the local browsing center is still a lot more convenient than having to take a day off from work or even travel to another city just to interview.

What is LensMatch and how does it work to select a suitable resume for recruiters?

LensMatch is patented AI for matching people and jobs. Unlike conventional tools that rely on keywords matching or that look for pre-defined terms or criteria, LensMatch learns from past patterns of placement to determine what kinds of jobs and candidates make intuitive sense.

Going beyond whether they mentioned the right words, it finds out, does the applicant look like the kind of guy who will get a job like this or is this just someone who listed a certain keywords a dozen times without actually having relevant experience.

The Predictive Matching technology works on statistical probability. Then, how can it actually select a suitable candidate for any recruiter on the basis of his past work experiences and descriptions? Also, how much will the recruiters rely on the predicted candidate and his skills?

At any moment, any individual person might choose to do something entirely unpredictable. I might decide on a lark to go to a store I have never visited before. Someone who has been working as a financial accountant might choose to quit it all to try to land a part in a Bollywood film. On the mien, however, human behaviour adheres to knowable patterns. Someone who has been working as a Junior Auditor is likely to move next to be a Senior Auditor. This is the broad concept behind Burning Glass's Predictive Matching platform.