Firing Squad: PredictiveHire's Barb Hyman


A new chatbot... Errrrrr.... "conversational AI: contender is coming to America and the rest of the world soon. They're called Predictive Hire and they have an army of PhDs that says their tech is better than yours.


CEO Barb Hyman can explain and has no problem staring down the barrel of our Firing Squad rapid-fire questions. Does Predictive Hire come out unscathed? Are run into the guns? Ya gotta listen to find out :)


Firing Squad is powered by Pandologic's programmatic revolution!


PODCAST TRANSCRIPTION sponsored by:

Disability Solutions is your RPO partner for the disability community, from source to hire.


pandoLogic Promo (-1h -1m -0s):

Hey HR, have you joined the programmatic revolution? If not, you're about to get schooled on how to take your organization to the next level. Check this out. 95% of CMOs use programmatic advertising, but only 5% of CHROs are using it to advertise their jobs. Where's the disconnect? PandoLogic uses powerful automation to drive higher performing job ads without the bloated budget. It's time to transform the way you think about job advertising and join the revolution. Learn more at pandologic.com/revolution


Shark Tank Intro (36s):

Like Shark Tank? Then you'll love Firing Squad! CHAD SOWASH & JOEL CHEESEMAN are here to put the recruiting industry's bravest, ballsiest, and baddest startups through the gauntlet to see if they got what it takes to make it out alive? Dig a fox hole and duck for cover kids the Chad and Cheese Podcast is taking it to a whole other level.


Joel (58s):

Yeah, the only thing standing between me and happy hour is this firing squad. What's up everybody. This is firing squad and you're listening to the Chad and Cheese podcast. I am Joel Cheeseman is always joined by my partner in crime, Chad Sowash and today our sucker is Predictive Hire. We welcome Barb Hyman from down under. She is the CEO Barb. Good day, mate.


Barb (1m 26s):

Good day.


Joel (1m 26s):

How's it going?


Barb (1m 27s):

Hi guys.


Joel (1m 29s):

So before we get into the company, tell us a little bit about you.


Barb (1m 34s):

So I'm the CEO and founder of Predictive Hire. I'm probably the least likely person to be leading this company as an AI company because I'm female, for one. And I am neither a data scientist nor an engineer, but I care a lot about the experience that people have when they're applying for a job. So that's me. I live in Australia, in Melbourne. I'm currently in day one of 14 days, quarantine. Yeah, I know, not fun. And I'm a mom of three kids as well.


Joel (2m 6s):

The wrong person, because you're a woman. We may have to get into that Chad and the Q and A, but until then, tell her, tell her what she's won.


Chad (2m 14s):

Well, Barb, you will have two minutes to pitch Predictive Hire. At the end of two minutes, you will hear the bell then Joel and I will hit you up with rapid fire Q and A and if your answers are starting to ramble or do you just get plain boring, Joel is going to hit you with the crickets. That's the signal to tighten up your game at the end of Q and A, you will receive one of these three grades from both Joel and myself. Big applause. Pile that shrimp up on the bar-bee and open some champagne because you have a winner. Oh,


Joel (2m 49s):

Open the Fosters baby, open the Fosters.


Chad (2m 52s):

Oh, that's golf club golf clap. Lay off all that shrimp pop a Fosters and get your ass off to work. The firing squad. That's not a knife. Seriously. Get back to the whiteboard. This isn't working.


Joel (3m 7s):

Gonna need a bigger boat. Barb are you ready to pitch the Firing Squad?


Barb (3m 10s):

Yeah. I just want a bit of what you guys have to raw me out, share it later.


Joel (3m 16s):

Listeners should know it's about six in the morning for Barb. So this is a little bit much for her, but she's a trooper and she's going to handle it. You have two minutes to pitch Barb. Are you ready?


Barb (3m 26s):

I'm ready.


Joel (3m 27s):

In three, two.


Barb (3m 30s):

Imagine for a moment that you're a store manager of a large national retail chain. You have 500 open positions. You get about 8,000 applicants a day, and you'd love to interview every one of them because you know that the people you hire, are your brand. So you sit down and you interview all of them. You let them share their story of who they are in their voice. You give them about 20 to 25 minutes and it feels really friendly and familiar cause it's just a chat and you don't go to them. You know, unlike recruitment you give every one of them personalized learning, which makes him feel amazing. It makes him feel heard, dignified and smarter because you're going to let down about 7,500, right?


Barb (4m 11s):

If you do the math, because you can only hire a few and you care a lot about the experience that everyone has, because they're all customers in your store. That's what we do, we make that possible. Where conversational AI, that's built for humans and backed by science. We call it Phai, PHAI. Predictive high AI. It understands you, it doesn't rush you. You don't need to do your hair and makeup for PHAI and PHAI is also blind. Doesn't know your race, age, or gender agenda. So it means that when you bring in those 500 people, they really truly represent the diversity of the community that you serve. That's pretty powerful.


Joel (4m 49s):

And where can they find out more Barb?


Barb (4m 51s):

Go to predictivehire.com.


Joel (4m 54s):

We'll end that right there. Okay. I'm going to start the questioning Barb. And I usually start with sort of a branding question usually revolving around your name, but your name makes a lot of sense to me. What's hard for me to get is the, the robot's name PHAI. So, so forgive, forgive the pun, but why PHAI.


Barb (5m 14s):

So part of what really has informed the whole product is the idea that everyone's given a fair go. You know, that's that mission. It's a great Aussi slogan. And that means that we don't really care that you're male or white or whatever your age is and we want it to have a chat bot with a name that was really, you know, neutral. It really gender agnostic. And actually our head data scientists came up with that. So that's how we came up with the name. We didn't want it to be a female like Alexa or Siri. Cause that just seemed too obvious. We wanted it to be something that represented us, but that also talked to our mission.


Joel (5m 48s):

So it's not a cute acronym or something. It's P H I or P H A I correct. That's right.


Chad (5m 54s):

Predictive Hire AI.


Joel (5m 56s):

Oh, it is an acronym. Okay.


Barb (5m 58s):

Part of what we're trying to do is be really authentic and honest and we don't pretend that it's a human. So when you give it a human name, you're kind of, we feel representing that this is something that it's not. You know, it is a robot, so to speak it is a machine, it is AI on the other end. And so we want to be honest about it and not give it a name that suggests it's a person.


Joel (6m 19s):

No job seeker confusion about PHAI. What is that? They get it, that it's the name of the robot.


Barb (6m 25s):

Some companies put their own name on it. You know, there are quite a few businesses now that really have built in chat as a vehicle to engage with their staff and they might have their own name. So they're welcome to apply that name, you know, part of our product is that the whole experience, right?


Joel (6m 41s):

Gotcha. Say no more Predictive Hire AI, P.H.A.I. Got it. Chad boom.


Chad (6m 46s):

There you go. I'd say it's a little later in the day for Joel. So he hasn't obviously taken his nap. So is this a high volume play, mainly?


Barb (6m 56s):

So it definitely is. You know, I think a couple of things that make it high volume, which is, you know, when you're bringing AI into your organization, it's a lot of change for the recruitment team for the organization. And so we find the bigger, the problem, you know, the easier it is for organizations to embrace it. So we tend to work with businesses recruiting at scale because they get the most payback, but you can use it in other roles as well. Because part of what the solution delivers is not just efficiency because we think it's much more important to also be really human. Is, you know, it just helps bring data and intelligence to every recruitment decisions. So whether you use it for engineers or for frontline retail staff, you know, it works in the same way.


Chad (7m 40s):

So when it comes down to a company actually buying into Predictive Hire, generally, what's their number one priority that you are filling right out of the gate?


Barb (7m 50s):

Absolutely recruitment efficiency. How do I get to the right talent fast? You know, you get to time to offer in less than 24 hours. And if you've got 500 open roles or any open roles now, because it's such a candidate first market, you got to move fast. So speed.


Chad (8m 6s):

Okay. So what about setup for that company? Because not everybody has the same questions. They don't have obviously the same assessment processes. I mean, do you come with canned assessments and canned questions that you just roll out automatically? Or is there usually a lead time? Weeks, months to be able to get it set up so that it can learn and then deploy?


Barb (8m 28s):

Yeah, no, look, it's actually a less than 24 hour set up, in theory. What we find though is usually there's an integration which might take a few weeks and the customer always wants to customize it. You know, they always want to bring in their voice. We're able to reflect their values and their culture in the questions. So there's usually a period of us collaborating with them to really amplify those aspects of the experience. But, you know, in theory, you can go live tomorrow.


Joel (8m 54s):

Talk about the job seeker interaction with, with P.H.AI. Is it, I go to the website and it's a little bubble in the bottom. Is it a text messaging as well? Is it like sort of mobile? Platform based? Does a recruiter send a link to a candidate and then they click it and then start the process of chatting? Is it, is it available on other messaging platforms? Like WhatsApp or others? Talk about that.


Barb (9m 21s):

Yeah. So it's simply a link that can get included in inserted in any sourcing channels so Snapchat, WhatsApp, et cetera. But usually what happens is we're working with really big companies who have a legacy system, like a Workday or a Success Factors. And so you apply for a job in the normal way, but pretty quickly you get to the point in that experience where it says, Hey, now we want to give you a first interview. We don't call it an assessment because it's like an interview. So everyone gets a chance at the job. And so they click on that and they start the conversation. And it's five questions, you know, that's really the key IP or an invention is that it's super quick. And within about 10 to 20 minutes of finishing, they get this feedback profile, which is really phenomenal in recruitment because, you know, there's a big difference between a normal chat bot that an engineer can code in a week versus conversational AI that's based on science where you really understand the person's real language.


Barb (10m 16s):

And we give that feedback back to the candidate. So it's pretty quick and simple, and it means no CVs, no phone screens. So from a recruiter perspective, it gives back just a huge amount of time to them.


Joel (10m 29s):

So let's talk about sort of global domination. Are you guys are based in Australia. Are you available in other countries? Which ones, if not, what countries are you hoping to grow into or are you just Australian or APAC company? Talk about the globe.


Barb (10m 49s):

Yeah. Yeah. So, round about, every 42 seconds to be precise someone around the world in one of 47 countries is having a chat with P.H.A.I. And our clients are global. Obviously we're, most of our team is based here, but we've also got a sizable team in the UK. We've got a team in the Philippines and some people in China as well. And you know, the one we obviously want to pursue global domination, you know, going into the US, is incredibly exciting for us. We've just got one client there at the moment, and we're looking to expand that, but the only constraint to our tech is it's English only, you know, AI is pretty sophisticated.


Barb (11m 29s):

You can't slap on Google translate and then, you know, provide the service. So we'll be in Spanish and French and Italian and German late next year. But for now we're in any English countries, English speaking countries.


Joel (11m 42s):

Gotch ya. Gotch ya. So in America, there's a lot of a lot of chatbots. It's very competitive. What's your differentiator?


Barb (11m 48s):

So, you know, I think the market is really confused about chat bot versus conversational AI. And, you know, the difference simply I say is, it's the difference between 2d and 3d, you know, a chat bot. If you engage with most of the chat bots out there, in fact, all of them, they don't give you anything back. You know, they're really just automating a process and it's very transactional. They're pretty crude and simplistic. They use keyword matching. They're not based on any kind of science, like natural language processing, which is what we have. So to be truthful, the only other technology that's similar to ours, but it's not really applied in this space is IBM Watson. Which was built through understanding people through language and then predicting whether or not they had a certain set of traits to be successful in a role.


Barb (12m 36s):

And so right now we're the only one that has something similar. It was built in a really different way. And so that's fundamentally the difference. It's truly a science-based experience. And you know, it's also very much a machine learning product and maybe I'll stop there and see what questions you ask. But I think there's also a fair bit of confusion around AI and a lot of puffery around AI versus something that's truly a learning technology.


Chad (13m 1s):

You said something earlier that I want to dive into, you said, it's really an interview it's not an assessment. Well, I mean, the assessments overall, personality assessments or what have you, they are different than an interview because of obviously the algorithm that tests, it goes against that. So is it an interview or is it an assessment? like a personality assessment?


Barb (13m 24s):

Yeah, look, it's designed as an interview. It simulates the best in class of what a good interview looks like, which is it's a structured set of questions where people are measured against a performance rubric. What we're discovering through the language is soft skills and communication skills and features like your propensity to be a job hopper, which is a kind of a formal term in IO world. And, you know, my view is it's exactly the same as an interview when you're interviewing me, you're assessing and judging me, you're listening to my responses to see whether I'm really driven, whether I'm really resilient, whatever it is that you're looking for. But you know, the problem is that you bring all of your biases, right?


Barb (14m 5s):

Like we're all imbued with biases and that's where a machine that's blind can really discover you know, what I would say is undiscovered talent. And that's the beauty of using technology to do your interviews for you.


Chad (14m 18s):

Okay. So a job hopper, let's say 2, 3, 4 years ago, that would have been a problem, but that's standard for today, right? So how does a company actually keep, like yours, actually keep up with changes in the market because you're actually using past behavior. And as we just saw over the last year, well, there's little the same than it was a year plus ago, 18 months ago. Right. So how does an algorithm that is trained on past behavior keep up with the world of today?


Barb (14m 51s):

Because it's not, it's not based on historical data. That is a key difference between us and every other AI solution. You know, if you're using, I won't mention any names, a lot of other AI solutions, they survey your employee group, you know, the whole Amazon story, which I'm sure you're familiar with, where you're using CVs from people that were hired. We don't use historical data to define the profile. What we have that's really unique to us is this incredible data set of words. It's about 300 million words. Now it's doubling every six months. And the signs of being able to discover your profile, as in your personality, your communication fluency from words comes from that data set. It does not come from tracking whether or not you were hired and whether or not you were successful in the job.


Barb (15m 36s):

That's how we start. And we start with, they're called rule-based models. As we get data from your company about who was hired and who left in any other truly obje