In this twenty-minute presentation, John Sumser covers AI and its role in the Recruiting Process at Elmhurst College.

We’d love to hear your thoughts about this video. Post your comments on or John’s YouTube Channel »





Important: Our transcripts at HRExaminer are AI-powered but there are still instances where the robots get confused and make errors. Please expect some inaccuracies as you read through the text of this conversation. Thank you for your understanding.



JS – John Sumser, Principal Analyst, HRExaminer

00:00:00:27 – 00:00:19:24
Good morning everybody. How are you everybody? That’s good. That’s loud enough. I’m John Sumser and I’m the principal analyst at HRExaminer. And I am going to try to cram a lot of stuff into a little time. I hope that you can follow along.

00:00:20:09 – 00:01:04:16
At the HRExaminer. I run a small industry analyst firm that’s right now almost exclusively focused on intelligent tools emerging in HR technology. Recruiting is by far the largest subset of that. You might take a look at the HRExaminer site sometime, we have a weekly newsletter, there’s a bunch of industry analysis so if you wanted to go deep on these topics we’ve got reports. There are a couple of podcasts a week and we’ve got 4 to 25 editorial advisors who produce cutting edge content about the state of H.R. and H.R. technology every weeek.

00:01:06:12 – 00:07:22:05
So I wanted to start by giving you a sense of the overall landscape. It’s been the case for either I’ve been lucky I’ve been looking at recruiting and recruiting technology for maybe 25 years now and that’s been the case almost all of that time that you could imagine recruiting as a stand alone function that’s not related to the rest of the H.R. department and what’s happening is the result of the price drops in processing and storage is at all of H.R. technology is sort of becoming one thing because for the company the data set is becoming the most important asset. So I just want to give you the big picture of each party as we get started talking about the little piece of talent acquisition and then we’ll get down to the implications for recruiting for college so the picture is that talent acquisition feeds the organization data that goes into talent management scheme which is how people work develop and all of the related hygiene factors in what’s happening in this entire arena is that employees and people who come in touch with the organization are being measured more and more and more and more. And one of the things that it’s really important to be aware of is the fact that measurement is the name of the game. The term so recruiting has exploded over the last decade from sadness and resonate well. When I started just before the job war really took off recruiting was the college recruiting department which was sort of in the loosest terms. It was it was a way to go out and networked in a very select small group of schools. The regular recruiting department which was sort of a reservoir a processing operation that led to interviews and hires and today there’s technology in all of these different facets of the recruiting process. You’ve probably seen some or maybe even most of these things on this spiral but it has become a complicated technical discipline that run by specialists whose expertise may not overlap. So one of the things that you’ll encounter is for every one of these boxes it’s just like a department in a university right. The one thing that’s right next to the department is the wall between nuts and everything else. And the the integration of these tools is coming along. It’ll be great in four or five years. But right now it’s sort of chunky and you get different experiences for everybody who comes along not because it’s personalized which is something that people can’t do but because the integrations are we’re learning how to do the integration devolved into pieces I guess is what I’d say. What’s notable here and I’ll talk about this more over the course of the conversation what’s notable here is that the requirements for a job are getting increasingly precise. Maybe the single biggest consequence of the introduction of I don’t like the term artificial intelligence but that’s what people are following. If I say intelligence tools because it’s not much more than math. The consequence of what’s going on is that requirements are getting more rigid but really great. Machine learning learns from lots of examples what the precise requirement. Machines are not flexible machines are really really good at providing precise rules and constraints. So so what you have to watch is that that the the lines that mark whether or not you’re in and out for consideration when you go through the chain are very stringent and increasingly so. That’s the downside of the new technology. The other thing that’s worth noticing about all of these boxes is that you can understand almost all of them as a messaging device right. So assessments meaning sourcing for sure but social recruiting and then social media becomes a measurement device it’s part of the data accurate Facebook account with job distribution tools to help the employer understand what works and what doesn’t work for or job ads employment branding. If you were to look at the at the outside control rooms it’s probably more descriptive than the reality. But but but where employment branding which is all that other jump about why you should come to work for the company in operations would be when you get there is measure and monitor and then look at to see if it results in people doing things. And so so I will go one step through all of the boxes on the chart. But every one of these things is measured and monitored and control trying to tweak the result out of the entire telegraphed system system.

00:07:22:06 – 00:07:31:00
This is my favorite slide. Beautiful.

00:07:31:06 – 00:08:20:21
I don’t know about you but I love it. So that’s part of why this is my favorite slide. A good bowl of fruit salad is an amazing thing my family my dad’s family was from Ohio and there was there was like a family competition to see with. So and So. So I have a fondness in my heart for fruit salads that thing that’s great about the fruit salad is isn’t part the way that begins to see as you get closer to the Bible the ball. This is what the computer sees and you should notice several things about this computer sees no Jews in the Bible. It just can’t stop. It doesn’t have pleasure pain in comparison. Anything like that. It’s just God.

00:08:23:14 – 00:08:31:07
Numbers see the ball do the numbers see the ball see the numbers. Now

00:08:31:07 – 00:16:11:07
Now what’s really interesting here is this is the kind of floor that you’ll find in almost every model and almost everything that you interact with these days is going to be a model you see that PowerPoint that’s at you end with some fruit. That’s the polka dots off the ball. And so you get this sort of it’s almost right. You can get a pretty good fruit salad by having a machine so you that all of these things by two thirds to get the warmer servings except the machine would tell you that you only need a third of the poker darts at the end of this evening. So there’s but the point of this image is that when you deal with artificial intelligence you’re dealing with mathematical models. And the mountain level mathematical models would say things are probably more precise that they need to be and they are probably less useful than they purport to be. So one of the things that people on the other end of this the recruiters at a certain point are having to learn about is how to argue with the machine because it’s the machine we’re evaluating this fruit salad and it had some fruit salad camera that’s coming along. And this was the set of requirements for the fruit so it might disqualify somebody who only had half the number of poker dots or who had a yellow ball or who had it for it or didn’t use whipped cream. The the the requirements become quite precise and the machines are scoring great big piles of recipes. And so when something is left out you fall down into the right things and it may be that what’s being asked for a measure is why took out of a bowl where there’s the problem of it’s already happened. Part of why I’m not at all clear and maybe we get a little bit of this at the person questions but I bet you don’t know that recruiting has already transformed that every aspect of recruiting is already different. And these 30 categories base of tool there are software vendors in the market many six hundred and fifty eight who are selling these sort of tools with intelligence and better intelligence terms to mean it’s a cheaper machine learning which is prediction based on the accumulation of cases or natural language processing saying which is churning Tech student bath and with those two tools. All of these things get done. There are some amazing things you might because it’s interesting take a look at text video text field that come it’s a tool that audits your job description for it’s implicit it’s not really bias but it’s implicit capacity to attract a general level girls or females. And so it’ll score the job description in a way that allows you to improve the mix. And so at that very fundamental level what is the job description we’d like perfect job. We’d like intelligent tools are making it so that the test itself has measured and monitored the production capabilities so so the game is changed then and here’s here’s what you need to know about the fact that the tables. If it isn’t in every employer that you interact with today it’ll be an every employer that you interact with tomorrow. That’s that’s bias mitigation as a technical challenge is front and center. And that means all of the ways that people have been evaluated in the past are now being shifted to account for desired levels of diversity. It isn’t as overt as calculations for quotas but but there is a distinct and strong technical emphasis to take the bias out of decision making and give you something that resembles an objective standard. And one of the things that gets talked about every single time is the fact that the name of the college is a biased marker. So so you should understand that that in the world of recruiting technology the college that you come from is some kind of bias its most acute in the Ivies. But this thinking is being applied to colleges in general and it’s part of a larger trend to look at skills rather than degrees we can debate really hard. Whether or not that’s a good idea. But then what’s happening is not what this means is that the rules for competition who’s going to be the person to get the job who’s going to make it so that the public while they’re changing and they’re changing rapidly what a resume a means is changing very very rapidly and for the time being maybe the next five 7 years exactly how your resumé will be treated with it fits the system is going to be in flux and they’re going to be grand experiments with automated recruiting that result in train wrecks. And so so we’re going to learn in a kind of a lumpy process as a culture how the social contract for work we’re doing to Burt’s part of what’s really important here is that data come from many many places. It’s not just your credentials as you submit them but social media and all sorts of other components does it. There’s a tool out there called the gates tower and that’s that’s amazing. And it will measure the likelihood that you are willing to take an offer and it makes the likelihood that you are willing to take an offer by taking hungry variables and assembling them into a prediction about your relationship to this particular offer.

00:16:11:26 – 00:16:30:18
And so when people look to make the decision not start with lists that are that are determined by who is most likely to accept the offer and I can’t tell you today how you would hold the machine’s perception of the likelihood that you will offer.

00:16:31:04 – 00:17:19:05
I don’t know if you’ve been teaching your students to use search engine optimization in their resume ways. But that’s a fail going forward that kind of thinking with results and being motivated being tough with keywords doesn’t make any sense any longer because that’s not how the machine processes things anymore that work in an era where search was operated with keywords but searches and operated with keywords in each new tool search it’s operated with machines assessment of those so that means that as you teach people how to assemble their resumé over time you’re going to have to change that it can be a fixed component of a curriculum.

00:17:19:05 – 00:17:57:22
It has to be market sensitive and about what’s in the market today. All of the places that are real you Christine are going to have rapidly evolving with moving technology. And I’m sure your you’re already doing this but the new kinds of work that are reporting which have to do with the management and utilization of the output of these Intel systems is going to it’s going to be it’s going to be an interesting time. We’re going to have the job for we know what they are.

00:17:57:22 – 00:19:51:01
I was just talking with somebody the other day who’s who’s hiring for an H.R. storyteller at the age of storyteller looks at all of the aggregated people where of Olympics data and uses it to tell stories to employees about how they’re doing and uses it to imagine new ways of employee experience and then leads a technical team to execute that experience so that so the job is imagining and experience from data talking to people about it and then putting that experience in place with the technical team. I don’t believe I’ve ever heard of a job like that before. But I think there are going to be a lot of so. So what you have to do if you’re going to have a career career is probably a bad word. But if you’re going to have a career meaning the aggregate of your life experiences that work it’s going to be pretty bouncy and pretty have to change that. So the most important thing that you know there is I would navigate sort of territory the cleanest thing that you can have when you’re navigating uncharted territory. There’s some idea of what you work in some method for getting to figure out what you work when you’ve forgotten what you when you move forward here. You’re not talking to recruiters anymore you’re talking to machines and the machines get your material in advance and you bring it into the machine. Those are likely to go through this means starting earlier in the development of rentals and manicure and social media. It probably means having separate professional and personal facilities. I’m not sure that’s that your clients.

00:19:51:08 – 00:20:25:21
It means that people who have ruled themselves by the rules you know the classic high achiever over 4.0 high school grade point average really do well in college with all the boxes. Some of those people are going to fail in ways that are surprising. And we need to be perfect for that because the anti bias mechanisms are going to start to watch for things like that and then matching is going to be harder.

00:20:25:23 – 00:20:32:28
So that’s the 20 minute version of A.I. in recruiting I can give you a five hour one some day if you like (audience laughing).


Watch more of John’s talks on his YouTube Channel »


Tagged with:  
Read previous post:
Video: New Management Approaches for HR in AI Enabled Enterprises

In this video presentation, John Sumser discusses new management approaches for HR in AI-enabled enterprises.