TAtech: AI Recruiting - Part 2


This is Part II of a two-part podcast LIVE from Tempe Arizona and the TAtech AI Summit. It’s a hype-free discussion around “AI and Automation” Aaron Matos – CEO of Paradox Olivia Yongue – Director of Client Strategy at KRT Marketing Sahil Sahni – Co-Founder of AllyO.

Enjoy, and support our sponsors: Sovren, JobAdX and Canvas.

PODCAST TRANSCRIPTION sponsored by:

Disability Solutions helps support and educate your workforce through disability awareness and inclusion training.

Announcer: Hide your kids. Lock the doors. You're listening to HR's most dangerous podcast. Chad Sowash and Joel Cheesman are here to punch the recruiting industry right where it hurts. Complete with breaking news, brash opinion and loads of snark. Buckle up boys and girls, it's time for the Chad and Cheese Podcast. (music)

Chad: Hey it's Chad and this is part two of a two-part podcast, we're just finishing up in Tempe, Arizona at the TA Tech AI Summit. It's a hype free discussion around AI and automation with Aaron Matos, CEO of Paradox, Olivia Youngue, Director of Client Strategy at KRT Marketing and Sahil Sahni, Co-Founder of AllyO. And, of course, some snark and opinion from Chad and Cheese. Enjoy.

Joel: It's commercial time.

Sovren: Sovren is known for providing the world's best and most accurate parsing products. And now, based on that technology comes Sovren's artificial intelligence matching and scoring software. In fractions of a second, receive match results that provide candidate score by fit to job. And, just as importantly, the job's fit to the candidate. Make faster and better placements. Find out more about our suite of products today, by visiting sovren.com, that's S-O-V-R-E-N.com.

Sovren: We provide technology that thinks, communicates and collaborates, like a human. Sovren, software so human you'll want to take it to dinner.

Chad: It's show time.

Joel: One of the things I loved early on, Aaron, when we spoke about Olivia and ... It's not necessarily a question for him, that's fine ... Is, you're trying to solve the mobile apply issue, which was the heart of the whole movement. For you, anyway. And as anyone in this audience knows, we've been trying to figure out the mobile apply thing for a long, long time. Just pushing apply through LinkedIn and then going to LinkedIn and then coming back, it just didn't work. And you created this mobile chat bot, or whatever you want to call it, to create an apply system through-

Chad: Assistant.

Joel: Yeah. Through a conversation-

Aaron: He's getting it.

Joel: ... and I thought that was fantastic, so a year or so since launching, talk about the mobile activity versus desk top. Maybe what's next for this, are we going to see voice assistants? Walk someone through an apply process.

Aaron: We got to the concept of this behind two things, one was, yes we were trying to solve mobile. We had made it not pinch and zoom ugly, but there was still crazy forms, and so this has been an obsession. How do you fix mobile apply, where candidates want to raise their hand in a mobile environment. The other piece that was driving us was we saw the problem that when I walked in to large TA teams, where they were all staring at their computers.

Aaron: And I just did not believe that the value of recruiting is in us playing with software. And so, it kind of got to be a pebble in my shoe of like, I think there's the 80/20 problem of, we're spending 80% of our time on admin stuff and 20% on people stuff and we should flip that. So, those two things juxtaposed to drive. Today, mobile on the candidate experience side is utterly amazing. Conversion rates in the hundreds of percents better because the process is easier.

Aaron: And that piece, I think, will just continue. The computer of the future is a mobile device. And that's still working it's way through the enterprise. I joke about this all the time, go take a Millennial and put him in front of [Toyo 00:03:46]. No offense to Toyo, but they look at it and they're like, "Where the hell do I Skype?" Like it's the most confusing stuff in the world and so, our enterprise software also has to take a transition from the last 20 years we built software purely for the enterprise.

Aaron: For the recruiters, for the HR departments, and we have to change that. We have to start building software first and foremost for the candidates, because if you build great candidate experiences, you'll actually have a great experience for the company.

Joel: Kyle, for the record, he drew first blood on Millennials, just for the record.

Aaron: I love Millennials. They're our favorite.

Joel: How are you working with ATS's, are integrations that paramount? Do you care?

Aaron: oh absolutely, I think it's ... We have a huge network of partners on the ATS and the CRM side for ... All the enterprises have these. We have to plug into this, there is a system of record and we have to help facilitate communications in a different way. We view that as our strength, our strength is to help facilitate communications, be able to talk to candidates in a new way, do that through different channels; through the mobile, mobile web, whether that's their Facebook, WhatsApp. Any kind of channel.

Aaron: But the ATS is still core, that's not changing. The CRM systems have a different layer, someone talked earlier about the CRM penetration, and the enterprise, I think, is pretty high. And that idea is, how do we get more candidates into the CRM to actually start doing work?

Chad: On the data side, some of the data that we have to work with is garbage. I mean, take a look at job descriptions. They haven't changed for five years, or even Plus, right? So, if you're trying to build algorithms off of data that's garbage ... We've always heard, garbage in, garbage out. How do you get ... And not only that, from a resume standpoint, when I have an asshole who puts frickin' 'ninja' on their resume, right? How does that even correlate to what I'm looking for?

Chad: Wizard. Something of that nature, right? So if you're working on ... Yeah ... If you're working with garbage data, I mean, how do you get passed that? How do you build ontologies that gets passed garbage?

Sahil: Well that is the challenge, and that's the IP. You cannot take the Google AI and assume that it applies to recruiting. We, Google as a customer and a partner for us, you just can't do that. You have to build it. And even recruiting as a generalization is very broad. Different kinds of skills, different kinds of challenges you're solving. So when you say "role" even the world "role" could mean very different in the usual Google sense, which is very broad, versus what you're looking in recruiting.

Sahil: To your point, we built the whole ontology to parse all the jobs recs. You guys know of how LinkedIn and Indeed scrape every job rec an ATS and just show the job recs. We took that a step forward and we said, "Hey, can you parse the information in there to extract the job title? The location? And the requirements?" And it's-

Chad: The requirements in many cases are still junk.

Sahil: Oh 100%.

Chad: You'll see, you'll see, you'll see "needs a bachelor's degree," and it's like, "Oh yeah we haven't needed that for two years, because it's a tight job market." So it's like, yeah we haven't needed that for a couple years now. What the hell do you do?

Sahil: Well, this is what you do. A, you use technology to parse that and surface what's on there, because the recruiters are not looking at it. Right? The hiring managers posted it, and no one's looking at it. That's number one.

Chad: Because they're too busy scheduling shit.

Sahil: Right, they're too busy doing other stuff. Who has the time to go change the years of experience from three to two? When you've got 80 candidates to schedule? Second, what we push for is rapid deployment. When you deploy and you're scheduling interviews with McDonald's in San Jose, and they're getting candidates who have one year of experience, within a week or two weeks, they will text or email AllyO and say, "Hey this needs to be two years."

Sahil: So you've taken what they have, you deploy, you let them review it. They need to sign off, we don't take liability for that, but you deploy because it's publicly out there. And then they, over time, optimize. That's the way you do it. You could spend all your time doing material, what's the perfect job description? But the reality is, a job description should vary by region because the label market changes.

Chad: So, Olivia, is this a major point that you talk to your clients about constantly? We were talking about this for decades, have we not? We've been talking for at least 20 years, as long as we've been doing goddamn job postings, right? Before Monster was Monster. Why hasn't it changed? And will it? And will these types of platforms start to surface the need to be able to do that?

Olivia: Absolutely. I mean, job descriptions are stale. All of my clients are like, when we talk about low-converting jobs and why they're not converting, we take a look at the job title, the description, maybe there's something in there that we could change and tweak to increase SEO, to increase traffic. And so, those are conversations we're having with them daily that, ultimately, impacts larger discussion of going back and updating all job descriptions.

Olivia: But just thinking about ATS integrations and, I want to talk about programmatic a little bit, we've been doing a lot of work of integrating with ATS's.

Chad: That's gotta be hard, though, from a programmatic standpoint, if you've got garbage, right? To be able to target.

Olivia: And the reason why it's important is right now in current state, we, essentially, get a live feed of completed applications feeding back to our system. So we can optimize programmatic campaigns, based off how many completed applications are coming through per job. However, we'd like to take this a step further and optimize per qualified applicant.

Olivia: And so, having that live feeding integration set up with ATS's is so important. I mean, not only driving the cost-per-application down for our customers, but just having that real-time data that we're not getting today. So, it's just something difficult that we're working towards to, essentially, offer in the near future.

Chad: Yeah, this is the bane of your existence right now. The job description.

Aaron: I mean, the job description's always going to be a challenge. I think that's part of the challenge in matching. We're trying to take incomplete job descriptions and match them with incomplete resumes. People and jobs are complicated. They're very, very complicated. I mean, there's companies they're doing incredibly interesting, innovative stuff. Companies like Intel that have kind have gone reckless, not reckless, but "reckless". And are managing talent pipelines and looking more at how do I ... look at people and instead of them applying for our jobs, we look at what they can do and match them to jobs in our company.

Aaron: So it's a lot of innovation that's happening around this. There's no perfect answer for that.

Joel: Olivia, quickly, what percentage of your clients are using an AI tool, currently?

Olivia: I'd say, probably, it's grown over the past year. There's definitely more awareness, some people just getting buy-in from leadership, but it's probably 30, 40%. So definitely room to increase that.

Joel: How big do you think it will get? Do you think it will be 100% at some point? And if so, how long is it going to take us to get there?

Olivia: I'd hope it'd be 100%. I mean, I want my clients to be using technology to make smarter decisions, to just work on areas of challenges; whether it's bandwidth or cost savings. A lot of the points I think Eric talked about today on one of the, I think four or five factors that you focus on, but I think it's going to take a little while. I think as awareness continues to grow and customers learning and understanding how it can help, probably i