TAtech: AI Recruiting - Part 1
This is part one 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
- Standard AI definition - AI Transparency vs. Explainability - Recruiter adoption for RPA or AI. WHY?!? - Is chatbot a dirty word?
PODCAST TRANSCRIPTION sponsored by:
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.
Chad: Hey it's Chad, this is part one of a two part podcast we just finished in Tempe, Arizona, at the TAtech AI summit. It's a hype free discussion around AI and automation. Again, this is part one of a two part podcast, with Aaron Matos, CEO of Paradox, Olivia Yongue, director of client strategy at KRT Marketing, and Sahil Sani, co founder of AllyO, and of course, some snark and opinion from yours truly Chad and Cheese.
Joel: It's commercial time.
JobAdX: This is the sound of job search (computer mouse clicking). This is the sound of job search defeat (computer mouse clicking and a groan).
JobAdX: Job search can be frustrating. Job seekers run into the same irrelevant ads, page after page before they find a match. When job seekers aren't engaged, conversions are low, budgets are wasted, jobs go unfilled. No one wins. But job search doesn't have to be defeating. JobAdx's smart search exchange references 400 data points to select the most targeted jobs, and delivers what job seekers really want, to premium add units across our network (computer mouse clicking, a man saying “score”). That's the sound of JobAdx's relevant results, attracting a qualified candidate and filling your job faster. Find out how to improve your job advertising campaigns, and increase candidate attraction and engagement by emailing us at; firstname.lastname@example.org. JobAdx. Together we can save job search
Chad: It's show time.
Joel: I don't know how I feel about this reclining set up here. Like two old men. Where's the cigars and the smoking jacket.
Chad: Just do your job.
Joel: Hello everyone, I am Joel Cheesman.
Chad: And I'm Chad Sowash.
Joel: And we are the Chad and Cheese podcast. I feel safe saying if this room doesn't listen to our show, they probably should, raise a hand so you know who we are, anyone. Outstanding. Well I guess my intro's over. You can find out more about us at chadcheese.com, if you have any questions or comments about our little panel here, you can hashtag chadcheese, and we'll get to questions or try to reply at some point, and you're doing the intro here right? Everyone's doing sort of the 140 characters or less who you are and why you're here? And then we'll get to the Q and A. Aaron, would you like to start us off?
Chad: Aaron starts.
Aaron: Hey I'm Aaron Matos, I'm the founder and CEO of Paradox.
Joel: Just a heads up, we're recording this Aaron, so choose your words wisely.
Chad: You said 140 characters.
Joel: I mean after the 140 characters.
Olivia: Hello, I'm Olivia Yongue, and I am director of client strategy at KRT Marketing.
Joel: Did you make those pants yourself or did you buy them like that?
Chad: He wants to know cause he wants a pair.
Joel: I do want a pair. Do they have jackets like that?
Chad: Oh Jesus.
Sahil: Are you guys done? Okay good. My name is Sahil, I'm the founder of AllyO, just wanted to take my 140 characters and thank both Aaron and Peter for having us here. I think we've been in this space for four years, completely illiterate, still about the talent acquisition space, it's great to come here and get to know people, so thank you so much.
Joel: By the way, we like our shows to be interactive, if you have questions, shout out something, Chad loves exercise, unlike me, so he'll run out and give you the mike to ask questions if necessary. Chad, you wanna start us off?
Chad: So the AI issue, that was shut down, the Amazon algorithm, that was shut down because of bias, right, has that helped? Has it brought awareness or is it a pain in the ass, because now everybody's asking about Amazon, it's like, no we don't wanna do that AI thing because it's biased, Aaron.
Aaron: I think Athena tackled this a bit, you know I think it was good that Amazon was transparent, and they admitted it, and they said, “hey this is what happened”, I think it is kind of a cautionary tale, to the question of, "are clients bringing this up?" I think they brought it up in the first. We have a really short attention span news cycle today, so, those ten days afterwards, people brought it up, today, I think if you're on the sourcing and matching side, you care about this stuff, for us, we're really trying to focus on BFOQ kinda very clear criteria as an assistant, so we don't plan that matching game as much, so for us it's not as big of an issue.
Chad: Okay Olivia, I mean, clients, are they asking about this whole AI slash Amazon thing, or did it just go away real quick.
Olivia: You know there is limited conversation about it, I mean, at KRT our job is to bring these types of articles and updates to our clients, so we really brought it to their attention, and I think there's no such thing as bad press, so as bringing it to light, it starts a conversation, it gets them talking about what happened, what can we do, what should we be thinking about to fix it if we are to go this route.
Sahil: I think in my opinion, what we saw, coincidentally just before the news came out I was at a conference where I'd taken a poll asking the audience how many of you had considering an AI solution in the next one year. And the response was six percent. When the interest is so low-
Chad: Six percent?
Sahil: Six percent was the response. This was a hospitality conference, so take that filter. When the interest is so low, and Amazon comes out and does whatever it does and openly says it, it actually causes everyone to consider that, “hey, if Amazon, which is really the largest staffing company in the world, has been working so much into it, and is open about it, there must be something in it.” I understand the negativity around it, the negativity helps in differentiating vendors who are actually solving it, if it's a problem. It does, it surfaces that. But at the same time I think it, in a weird way, amplifies the hype, and so now if you go and ask, you'll have somewhere like seventeen to eighteen percent of the market is interested in exploring an AI solution in the next twelve months.
Chad: What you're looking at, you're looking at major enterprise organizations, right? So, most of them are federal contractors, and they're scared shitless, in most cases, because there are hundreds of millions of dollars, if not billions of dollars, that they're getting contracted by the US government, are you building platforms, to be as Athena would say, explainable or transparent, knowing that at the end of the day, the US government could regulate this and say you have to be transparent.
Sahil: Yeah, I'll give you examples, amongst like major defense contractors, customers that we have include Walmart, or Allied Universal, G4S, and if you go and apply on their website, or did a year or two ago, you might know this when you apply for a job and it's a defense contractor, you're required to give ten years of employment and residential history and they would include that in the application. Awful.
Sahil: What we've seen at a high level is the adoption in AI has been for reasons to improve their candidate experience, more from a conversion standpoint because end of the day they need to hire. I mean, how good is a defense contractor who can't sell the defense. Right, and so, we've seen them adopt it for that reason, from a concerned standpoint, absolutely, when we play mostly in the larger enterprise space, and we're going through all the scrutiny, whether it's your IT security check ins, and whatever it might be sometimes it takes a couple of weeks, sometimes it takes months and with defense contractors, it's much more severe.
Sahil: But having said that, at the back, you've got a very core need, it's not like they're saying, "hey if this passes, I will use it." They want this, and they want it to pass. And so they are making sure, that whatever is happening is not only compliant, but it's also moving efficiently to get to that point where it can start creating value.
Chad: You guys don't want to weigh in on it? Not really, it's just I don't like that compliant shit. I don't like that compliant shit. Go ahead.
Joel: Eric Kostelnik earlier from TextRecruit said there were only about a handful of AI companies in the world. Google, Microsoft, et cetera. Is he right? And if so, are all these AI solutions just using all these other big companies AI to power their stuff, is there any homemade AI out there in our industry, because I think if everyone's just using Google or Amazon, or whatever AI solution that there should be some transparency. Discuss.
Chad: Is Eric still in here?
Aaron: No, he jetted, so he left.
Joel: Are we gonna refute whether he said it or not?
Aaron: No he said it, I mean I think that, what's your definition of AI?
Joel: This isn't about questioning me, Aaron. This is me asking you questions. And I love that you're trying to dodge the question, but I'm not gonna let you do that.
Aaron: I'm not, do you think this whole room would agree with the definition of AI?
Chad: Well okay so question, because you use AI right, in defining what you guys do.
Aaron: Sort of.
Chad: Okay, well I mean if you use it, then that means your company's defining it.
Aaron: We actually call it assistive intelligence. Because we're trying to make a very clear separation that the goal of what we do is help assist, and get work done, that's our goal. To the question of what AI is, people talked about this before, the original Dartmouth conference was in 1956, and John McCarthy, they talked about what AI was originally, which was “how do we do work that humans are doing, how do we understand language, how do we understand messaging, how do we do translation.”
Aaron: Today AI is a bigger category where that's full machine learning and deep learning, includes things like NLP, partly I always joked that a rose by any other name would smell just as sweet, these companies are doing amazing work, that's changing recruiting, and people honestly sometimes just get too hung up on, “that's not AI.” John Sumser, who we all love, he's been around and his first report on AI said, that he kind of defined AI as having sentience. We're not there. This is not Westworld. Google built Alpha Go, it can play Go really, really well, it can't play checkers. It depends on what you define as real AI.
Joel: So AI is not AI is not AI. So Googles doing AI, and you're doing AI, but it's just a little different AI.
Aaron: I think Eric's point was there's a few big companies, and look China's even leading the way on just really great ML, and great power that is used in machines to see insights that we have never seen before, and Google's open sourcing some of this stuff, and those big tech companies are clearly leading the way. But, same point, there's companies in our space that are using amazing ML to source to match, there's NLP that's mincing conversations from all of the players in the space, that are ten, fifteen minute conversations, that pass quote, the Turing test. So again, a rose by any other name, would smell as sweet.
Joel: Olivia, when a customer says, what's AI, what do you tell them.
Olivia: You know I think there's so much noise out in this space, of just partners throwing out the word artificial intelligence, and so I strongly believe that there is, I know there's a lot of my vendors out there, but a lot of the tools are really smart, intelligent, I don't think they are truly doing artificial intelligence if you break down what AI is, but I think they are doing excellent work, and they're providing a ton of value when it comes to machine learning and augmenting, and providing that value, I just think when you go back and really define what AI is, in having that artificial intelligence come into play, it's not truly doing what it should be doing.
Chad: So the AI End-to-End Recruiter doesn't that make HR numb when they hear AI, AI, AI, and then you come in-
Joel: It just means nothing right.
Chad: Yeah and then you come in and you're like “Hey, I've got this End-to-End Recruiter Recruiter, we're using AI.” And they're like, “Bullshit.” So, how do you differentiate yourself, how do you actually prove beyond that. Is it because, is it RPA, or is it beyond that.
Sahil: So firstly I feel like when you think about where AI is going and the buzz around it, who's to be blamed for it is the market. Because vendors are searching for AI, so when vendors are searching for AI, sorry, buyers are searching for AI, and when buyers are searching for AI, vendors will use that. That's keywords under caption or whatever. First and foremost.
Chad: So you're blaming the buyers? Because that's what they're looking for?
Sahil: No, no, I'm blaming the [crosstalk 00:13:18].
Joel: The markets.
Sahil: Wherever it might come, but end of the day, the goal of a for profit business is to sell and make a profit.
Chad: It's not to lie though.
Sahil: So that's point number one. That's point number one, I think point number two, I'll take my definition of the AI the way we think about it, it's not automation only, you need to have a objective which comes down to either the process and what you're trying to improve, and then you need to have the math behind, which automatically learns and maximizes that objective. That's what the human mind does, when you grow up at the age of two you touch the hot stove, you learn that it burns, and over time you learn not do that, that objective is solving for not getting burnt.
Sahil: That's AI, we as a company have our NLP framework, we have our own ontology, we have patterns around it, we've taken AI in a very different way and we've patented that specially for the recruiting world. Having said that, I will confess that eighty, eighty five percent of the value to Aaron's point comes from automation, it's not that the industry needs AI, an AI is just a tool there and sadly so it tends to create more value in the presale process than wholesale. It might change later, I'm not saying it won't change, Google's the first company to come and say, “we'll be AI first.” You type in sentence in Gmail there's a one in five chance that it'll complete it for you. Right, it's pretty awesome. That is cool stuff.
Sahil: But today, in the recruiting world, AI is being done, I think it's being done by a few companies, it's been advertised by all, but that's because most of the value exists in the automation, not in the AI.
Joel: It's commercial time.
Canvas: Canvas is the world's first intelligence text based interviewing platform, empowering recruiters to engage, screen, and coordinate logistics via text, and so much more. We keep the human, that's you, at the center, while Canvas Bot is at your side adding automation to your workflow. Canvas leverages the latest in machine learning technologies, and has powerful integrations that help you make the most of every minute of your day. Easily amplify your employment brand with your newest culture video, or add some personality to the mix by firing off a Bitmoji. We make compliance easy, and are laser focused on recruiter success. Request a demo at gocanvas.io, and in twenty minutes, we'll show you how to text at the speed of talent. That's gocanvas.io. Get ready to text at the speed of talent.
Chad: It's show time
Joel: Should there be standards around AI? Cause I'm sure you guys just grit your teeth when you see companies you know aren't doing AI, but their site is dot AI, they have AI on their website, they're promoting it, I mean should there be a standard with which consumers can tell whether it's real AI or not? So like the old job board association?
Sahil: Well I think it is. So the buyer in the market is super smart. They're super smart, I think they gravitate with starting with AI as, “I'm interested in AI, but the sale process, or the selection process is all around what value I'm gonna solve and how I'm gonna track that.” And soon AI just becomes a minor ingredient in that. It just becomes a minor ingredient. Whether you solve that problem with AI or automation or you have humans doing it, if you are solving it efficiently, you are solving it. So as a result I think that standardization could happen, it might get some marketing buzz, but it will not make no difference in practice, because no one is buying a solution saying, “hey, I just want AI,” without knowing what the hell it's gonna do. And even if they bought it, at some point six months down, they'll realize, “hey, it's not doing anything.” And so they'll stop.
Chad: Olivia, earlier we saw some McKinsey data that showed fifty percent of jobs are automatable, right?. Tasks versus the actual jobs themselves. What the hell are companies waiting for? I mean if there is tasks that you can actually see, and even Fred showed it very nicely with the O net data right, and he highlighted some of the tasks that can be automated. What the hell are companies waiting for, I mean you're talking to companies, multiple companies every day. What the hell are they waiting for?
Olivia: It's change. It's being uncomfortable with change, I mean I study my client's funnel, their [phonetic candidate funnel, so studying how many applications are getting to generate a hire, and a lot of them are Fortune 500, and they're converting literally one to three percent of applications into hires, which is so low. And knowing that information and there are tools out there to help them, to shrink, or grow that percentage or you're looking at that ninety percent of talent that isn't even getting touched, there's never engagement, I mean trying to get them to think about different ways to get ahold of them, I mean it's difficult bringing new technology, adapting to change, going up the ranks talking to leadership, I mean there's so many layers that a company has to think about before even implementing anything like this, so it is pretty difficult, I think it's gonna take awhile to get there.
Chad: So Aaron what I'm hearing is Change management is a bitch, how do you get by that, is that like one of the biggest obstacles that you have to actually deal with?
Aaron: Change management is a huge issue, especially on the global enterprise. I mean you're in enterprises, these are not one workflow, this is not one recruiting process. You have to first and foremost solve business problems and then you gotta figure out how you're gonna make change. And if you change technology, then you have to also change people and processes, and what the work that gets done, the technology helps enable that. But large organizations that have huge departments, huge recruiting departments, huge hiring neural nets, it's a task.
Chad: So from a recruiter’s standpoint, and a recruiter adoption standpoint, who's trying to get recruiter adoption. So the question is why, why isn't this a part of the process methodology already, and it's taken away, just like from a scheduling standpoint, sourcing can be done automatically, there can be QA/QC, scheduling can be done automatically, there can be QA/QC. Why in the hell are we trying to get recruiter adoption? I don't understand. Why don't we, as leaders, actually adopt.
Joel: Because you're turning around the Titanic in most cases. It's not easy.
Chad: Yeah but if you're taking it off the table for your recruiters and you know it'll make them more effective because they're doing all this admin shit.
Aaron: I mean I think that is happening in some companies, I think the large organizations are trying to figure out how do you eat a whale, I mean it's a bite at a time, this is hard work. But in smaller midsize companies I think you can kind of dictate, “hey, all scheduling's gonna now happen through this technology.” And it is happening. We're still, and I mentioned this last night, I feel like, and for those of you who have been around and watched the internet grow, this feels to me I think there's this very large very large shift on how software is going to work for people in the future. That we're going through this transition, that the software is going to be intelligent, it's gonna have automation built into its core. And it feels a lot to me like the web in 2000 or 1998, where it's still so early, in 2000, everyone thought the web was done, or in 2001 it was like, “hey, the webs done.” I think we've had a lot of innovation since then, to me we're in that early stages of a lot of transition.
Chad: So you have recruiters?
Chad: So talk to me about that.
Olivia: We don't like to have the hard conversation. You have to say yes for things
to go your way.
Chad: It's not a hard conversation though.
Olivia: Then it slides. But that slide is, people think jobs and all the hype is jobs, those are the headlines-
Chad: It's a task. You don't have to take the garbage out. Yay, I don't have to take the garbage out.
Olivia: If you don't articulate that, in your plan, “here are the tasks for automating, and here's what you're going to do instead.” If you don't have that second part, you can tell them till you're blue in the face their job's not impacted, they're still not gonna believe you. Cause layoffs happen all the time. Somebody had a layoff somewhere.
Chad: If you stop taking, if you started having somebody else take out the garbage for them, okay, and they didn't even know they're like, “holy shit, the garbage is taken out.” What's the-
Olivia: So I say, “you don't have to take out the garbage”, but if you're taking out the garbage ten times a day, they're going, “hmm, what else am I gonna do?” And if I don't come and say, “now you're gonna mow the lawn, and you're gonna wash the windows."
Chad: Now you're gonna do more of the brand ambassador things that we should be doing in the first place, to actually ensure we have a hell of a brand.
Olivia: Yeah so now you're gonna do something else, but companies don't articulate that, they don't say, "here's what you're gonna do-"
Chad: Is it that hard?
Olivia: It's not that hard but it's a hard conversation. Cause they're scared.
Joel: Fear, yeah. We're human beings, we're trained to stay away from change. Someone said recently that this whole thing feels a lot like search in the early days. You do searches and the results were sorta kooky, but you knew eventually it'd be worked out, and to me that's to your point of early days, that's kinda how it feels. Is chat bot a dirty word, and I ask that, because it seems like all the chat bot folks are running away from being labeled a chat bot. Am I wrong on that? Is it a dirty word? From the agency, it feels like companies are getting away from that, and I wanna know why.
Chad: Well we actually did an interview with Quincy from AMS, and she said when she throws out chat bot in a conversation-
Joel: Platform automation. Automation platform.
Chad: The company pukes a little bit on themselves, and they're like, "no, we're not a chat bot." So is it that labeling thing that's a problem or is it not a problem?
Aaron: I mean first of all, I think Quincy's right, we've always called the concept an assistant. No one ever called Siri a chat thing... We don't do that.
Joel: Voice assistant is not a dirty word.
Aaron: The whole idea for me of what we're trying to do is, and yeah, it started with chat bot, and to me it was like, “hey there's a website.” Well is SAP a website? I mean come on, it's a lot more than that. And I think the whole idea, if you Google bot, chat bot, and say define chat bot, it will say a computer program designed to have a game of basically talk reply. And I don't think that's what we're trying to do. I think the market has moved past that very quickly. For us, it's a dirty word, but it's also a frame of reference, because people do understand chat bots today, so they understand the concept. So we think of it as fine as a frame, I think we all have to go past that, and the people in this room, we are going past that.
Joel: Do you feel like part of it is the risk of commoditization of chat bots? And if you're labeled a chat bot then you become a commodity, so we need to be more than just a chat bot. Is that part of why it's a dirty word?
Aaron: To me I don't know that that's the reason, I think that that it focuses on the wrong thing. To the point of buyers aren't trying to buy AI, they're trying to solve recruiting problems, and business problems. No one's ever came and said, "hey I wanna buy a bot." They don't do that. They come in and say, "hey-"
Joel: I've heard agencies say people call them and say we need a chat bot. We don't know what a chat bot is, but we know that we need one. Am I wrong on that or? So it is happening.
Chad: Olivia, what are you doing to stop this chat bot madness, this chat bot fucking madness.
Olivia: Oh, it gets my clients talking which is important, it talks about when they bring it up, what kind of challenge are you trying to solve here. And how can a chat bot help you. So, having these conversations is really important, if they want to call it a chat bot, if they wanna call it a virtual assistant, that's fine with me personally. I think my job is to help and this is with Laura's point earlier being a translator. Every single vendor out there calls their product different names, it's very, very confusing, so if I can help translate what does that mean to them for their business and how can it help them, then I'm doing my job. So if they wanna call it a chat bot, and I can communicate what they're trying to solve, then it's a win.
Chad: Okay so, on the AllyO side, is that incredibly hard because what we're saying is that hybridization of platforms and processes, right? So it's like you have a platform that does all of this, and it's really a hybrid of different point solutions. They're all different to Olivia's point, right? They're all different in their message, different... How do you get by that? What's the major... If you're talking to Olivia who is incredibly powerful from the standpoint of being able to touch and educate clients, how do you get past that? Because again, there's a lot of noise, and it's really muddy.
Joel: By the way, Olivia is Paradox's [crosstalk 00:26:14], and it's also her name. So just don't get too confused. I was a little bit confused.
Sahil: I'll confess, I've only known two Olivia's in my life and I have both on the stage.
Joel: Olivia Newton John?
Sahil: Don't ask me to pick who I'd prefer. I think, I'll speak a little bit to the previous one and then try to answer to Joel, to your point. My intuition, it's the opposite. Will chat bots commoditize by 2020 or 2021, yes. If we had to put money I'll put it on the yes bucket. But they haven't commoditized yet. What that means is when you say chat bot, you could mean something very different than what you say chat bot. As a result, if I'm doing something very different than what you think, but maybe more aligned with that, then why would I call myself a chat bot. So I don't think it's gotten to the ATSCRM world. But it's really a difference of ten percent, you're talking like eighty, ninety percent difference between products. And so I feel that's the key point.
Sahil: To your point on how do you deal with the challenge. If I'm talking with Olivia or her clients, it really all comes down to showing the value that you've created for customers. It's just the case studies. You go into a customer, they had a certain problem that hopefully resonates with who you're talking about, and then you show what you did, and then you show how that improved, and hey, talk with them, or see what they said in Forbes, or see what they said wherever. It comes down to that, and who cares whether it was a chat bot, whether it was an ATS, whether it was Joel and Chad sitting at the back and making it happen. It happened. And if Joel and Chad did it they should be paid for it.
Chad: Could be a Chad bot.
Joel: You heard it here first.
Sahil: So I think the honest truth just lies in-
Joel: Snark on demand.
Sahil: The honest truth just lies not on the website, it lies not in these presentations here, it just with when you are working with a customer, or folks who are sending that message out, you're not trying to create a ill formed category out there called chat bots because it's not reached that stage, but you're trying to just say, “hey, if you have these categories or problems, this is something that might work.” I'm guessing that's how you guess.
Joel: It's commercial time.
Sovren: Sovren Parser is the most accurate resume and job order intake technology intake technology in the industry. The more accurate your data, the better decisions you can make. Find out more about our suite of products today by visiting Sovren.com. We provide technology that thinks, communicates, and collaborates, like a human. Sovren. Software so human, you'll wanna take it to dinner.
Chad: And that's the end of part one. Check out part two for the rest of the story.
Ema: Hi, I'm Ema, thanks for listening to my dad, the Chad, and his buddy Cheese. This had been the Chad and Cheese podcast. Be sure to subscribe on Itunes, Google Play, or wherever you get your podcasts. So you don't miss a single show. Be sure to check out our sponsors, because their money goes to my college fund, for more, visit chadcheese.com.