Javed Muhammedali: What has happened in the industry, in general, is that a lot of those insights are available, but it’s locked away behind a huge reporting application. The end-user has to leave what they’re doing, move to a different interface, and then run this report that you have to wait for.
What we’re trying to do is figure out what elements of those insights can we deliver in bite-sized pieces directly to the recruiter at the time they need it. We’re hinting and nudging them along.
Sramana Mitra: Algorithmically, what is AI doing in here that is particularly interesting?
Javid Muhammedali: That is a great question. If you look behind the curtain, the three major areas of capabilities that we’re focused on are natural language processing, which is feature extraction, tokenizing requirements, identifying what the key requirements are, and natural language generation.
Then the middle layer is symbolic AI or AI, which is a decision tree basically. What are the decision points? How do you codify those decision points in a workflow that is both interesting and meaningful? We’re building out some configurability there.
The third pillar is typical machine learning. So going back into what inputs produce the outcomes we care about, which are submissions and placements. Even for the placements, the duration of placements and perhaps the pay rate of the candidate who got placed in the job. How do we optimize our matching algorithms?
Sramana Mitra: How do you quantify ROI?
Javed Muhammedali: Especially on the analytics side, it’s very hard to quantify the ROI. It’s really what you do with it. I see your point there. On the flip side, it’s very easy on the workflow and the sourcing to quantify ROI.
On the sourcing side, if we can bend the curve and shift a few points in expense, that’s savings, the less they have to spend on finding new candidates and sourcing them from other sources. That represents an immediate benefit to their bottom line because that’s a hard dollar cost saving.
In the middle bucket, you’re improving the productivity of your recruiters. There are soft dollars savings in ROI where you’re saving a recruiter some time. Now they don’t have to do 10 steps to get the phone number and contact them.
They can just text Cleo, our assistant, and say, “Hey Cleo, what is Javid’s phone number?” Those are productivity enhancements. We’ll be quantifying those in terms of time saved and the first one in terms of actual dollars saved.
Sramana Mitra: Are we at a point where an agency doesn’t need as many recruiters to do their job?
Javed Muhammedali: I doubt that the impact will be on desks in the current environment. For years and years, we had a skills gap. People who worked in healthcare know that that’s not a year-long problem. That’s a decade-long problem. We’ve had that in our economy for quite a while.
Now with unemployment at 3.5% for such a long time, we’re actually reaching at actual bodies gap. In as much as it’s true, there’s a gap there for our customers. It is also true for our actual customers.
It’s hard to find a programmer or maintenance worker. It’s equally hard to find a recruiter nowadays. The immediate savings will be in helping our customers satisfy the demand that’s right in front of them.
This segment is part 2 in the series : Thought Leaders in Artificial Intelligence: Javid Muhammedali, VP of Artificial Intelligence at Bullhorn
1 2 3