How TA Leaders Should Think About AI With Albrey Brown of Joonko

William Tincup: [00:00:00] This is William Tinkup, and you are listening to the Recruiting Daily podcast. Today we have Aubrey on from Junco, and our topic is how TA leaders should think about AI. Aubrey, would you do us a favor? Hey, check my pronunciation of your name. That’d be one thing. To introduce yourself

Albrey Brown: and Junco. Yes, awesome.

My name is Al Bray. Thank you for asking. I’m a long time listener, first time guest, so I appreciate you bringing me [00:01:00] on to give you a little bit of context about myself. I’m the VP of strategy at Jumeco. We are recruiting sourcing software specifically geared towards underrepresented candidates.

And before Junko, I was head of diversity at three companies, VMware, DocuSign, and Airtable, which…

William Tincup: Small companies, no one’s really known or heard of. Yeah,

Albrey Brown: got it. Exactly, especially DocuSign. I don’t think anyone has ever heard of that one. But I’ve been in the diversity and inclusion space for a long time.

My first foray was actually 10 years ago. I started a coding bootcamp specifically geared towards women and people of color. So I’ve always worked at the intersection of recruiting and diversity. And my latest my latest role is VP of strategy at Junko, which you can encapsulate in me being the subject matter expert for both our technology, the market.

How acquisition folks and how they think about diversity recruiting and bringing those three things together. So

William Tincup: we, we could go into how you got here and why [00:02:00] you want to work for a software company in this way, but let’s dig into some of this, some of the AI things, especially when it comes to marginalized groups.

And again, there’s. We, you and we talked about a pre show, there’s a lot of positives and there’s also some potential pitfalls to AI and it’s all evolving right in front of us. Totally get it. But what should they be thinking right now? I know you just recently came back from a show with where you were around a bunch of other TA leaders, but how are they currently thinking about a, the intersection of AI and DEI, if we want to look at it like that, and and how should they, if they’re not already thinking those thoughts, how should they be looking at it? Great

Albrey Brown: question. And. Junco is a technology just to give some context about why we’re in this space.

We use AI natural language processing, machine learning, and automation in order to help recruiters make decisions on who, who to interview. So we’ve been in this space for a long time and I think an [00:03:00] interesting place to start is that eight months ago, let’s call it November or October. We actually had to debate whether we wanted to talk about the AI components of our technology, because when we spoke about it, recruiting leaders were immediately turned off.

This was just October of last year. And come January chat, c p t comes out and continues to be head a headline, and folks then start to reach out to us. To to talk about how they could leverage our artificial intelligence to source underrepresented candidates. So I say that to say that over just the last 12 to 8 months, we’ve seen this complete change of how recruiting leaders are starting to think about it.

First, they were very skeptical. About how I can help them and or, the cost of benefits. And now, just, a couple months later with breakthrough technology, there’s a there’s an ubiquitous embrace [00:04:00] of artificial intelligence. And I think that’s ultimately a really good thing. The leaders that I was talking to.

About two weeks ago, there’s about 85 of them all had the same exact. I saw the same exact benefit. How can we automate some of the things that take our recruiters and our sourcers a long time to do, but have a low impact on interviewing candidates, which is ultimately The job. So first and foremost, things like communicating and scheduling, right?

That seems like a pretty right ob obvious way to automate out some of the chat or automate out some of the repetitive tasks or note taking, right? Taking notes and putting them directly into your a t s. This is something that can easily be automated with a recording software.

So that was the one, one category of things that was an obvious benefit. Then you get to this middle category of things that have both this cost benefit analysis. So there are technologies that help you surface the best candidates[00:05:00] for your sources and recruiters without you having to, without you having to do anything.

For example, There’s a technology out there that will search your ATS and understand which candidates, a profile of which candidates have made it through the process, look and then surface candidates that are very similar to those to those candidates who you haven’t talked to just yet. Now, on the whole, That sounds like a great thing.

Why wouldn’t I want to go and look for folks who successfully made it through my process and automatically get more folks like that? But the cost of that might be perpetuating a similar homogenous group of Candidates that are very similar to the people that you’ve already hired.

Is that a good or a bad thing? It depends on what recruiting it depends on what you as a recruiting person think, but as someone who really focuses on diversity it’s an obvious thing. Negative for creating candidate diversity, and then the last piece I think here and what is using AI [00:06:00] to help evaluate candidates in real time.

So there are tons of AI tools that are assessments or use video and audio to understand whether a candidate is good or not in place of an assessment or a written assessment. And that’s where the costs and benefits become very obvious obvious, obviously challenging. Where? How is this technology looking at a person and looking at a video and not making a biased decision as to whether they could be a good salesperson, marketing person or engineer at your company?

I think those are the kind of the three categories of things that folks are starting to weigh the costs and benefits for.

William Tincup: It’s interesting again, the bias and de biasing of these technologies, you’ve got a products now that can simulate you, yourself, like you, you spend some time with it you speak, it takes pictures of you the whole bit, and then it could do the interview, without you, and oh, by the way, it doesn’t have to be you, it could be an [00:07:00] avatar, it could be someone else so again, video interviewing there are a bunch of unintended kind of consequences of video interviewing is there can be bias in video interviewing, just like at the base of can you do a video interview someone that’s deaf or someone that’s blind or someone that’s, in a rural area or doesn’t have wifi can’t do it or it doesn’t present them in a great way.

Exactly. That’s bias. So so what are the questions or what do you think that they should be thinking? Again, all this is evolving and I know that, we’re just in the moment on this journey. So what should they be thinking now when they’re purchasing different applications about, okay, so these are the questions I should be asking now of on the, how this impacts you.

My recruiters are my hiring managers, my sorcerers, candidates, the executives. What are the questions that they should be asking themselves?

Albrey Brown: And this is a, the key [00:08:00] question that we’ve been working with customers on, and this is not the key thing that we solve, but of course, as an AI, a product that uses artificial intelligence, we want to make sure that our customers are ahead of the curve.

And being intentional about how and where they use artificial intelligence in their recruiting process. And, not to be too markety, but we call it an intentional AI strategy. How do you use AI with intention so that you’re not exacerbating discrimination, but you’re mitigating discrimination.

And there are, three kind of key pillars that I would lay out, or that we lay out for our TA leaders that we work with. The first is is about where you where you apply artificial intelligence to your process. So we talked about those three different situations. The first is applying AI to make, in a way that is unbiased.

So when we’re talking about communicating with a candidate. That is [00:09:00] a, has a low bias rating when you’re communicating with a candidate to set up an interview. You’ve already chosen that candidate. You’ve already, you haven’t assessed that candidate. You’re basically just going through the motions to make sure that you can meet with that candidate.

So that’s a great place to. apply AI in a way that has a low probability to be to bias against or for that candidate specifically. The second place is the second thing that we talk about is investing in vendors that are creating equity using AI to create equity for a candidate. So for example, there’s a product that Uses AI to it to review an interviewer and interviewers language and interviewers tone during an interview and give that interviewer a score as to whether they conducted a good interview or not.

Now, that’s very different than a technology that. That reviews the interviewee because now you’re actually evaluating [00:10:00] the person who has the power so that the candidate and future candidates can have a better candidate experience and that’s where you’re being intentional about where you’re using AI.

So we always challenge talent acquisition folks to understand where they’re applying artificial intelligence. And making sure trying to make sure as much as possible that the artificial intelligence they’re using is being applied to create a great candidate experience or an equitable candidate experience.

The second piece is about understanding. Artificial, the A. I. And the algorithms they’re using. So we really challenge T. A. Folks to challenge vendors to teach them about how the algorithms they’re using. A. I. Is a very big group of technologies. You have large language models. You have machine learning.

You have computer vision or speech to text. These are all Categories that we believe that talent acquisition leaders should at least be familiar with and know how [00:11:00] vendors are using this technology to get what they to provide value and how that technology is affecting different demographics.

So you can see you can have your vendor audit. How does your. The product that you’re using to evaluate whether someone’s 50 15 second video makes them a good salesperson. How does that, those results differ between women, people of color and white males? And can you give me a report that tells me that where bias is coming into the process before I sign on with you as a customer?

And while it’s not up to the TA leader to, to build the technology or fully understand the technology, I think it’s up to the TA leader to truly understand how that technology might impact the candidates that are applying. to the roles that they have up on their job site. So that’s about that, the transparency and really understanding the technologies up to a point before you buy them.

And then the third piece is just [00:12:00] communication. Communicating are you ready to enable your recruiters to communicate to both hiring managers and candidates what technologies you’re using and how that technology might impact their candidate experience. Earlier you said that we were talking about how candidates are now using ChatGPT to tailor their resumes to certain roles.

The reason why that exists is because ATSs look at certain keywords. And prioritize candidates based on those keywords. That’s something that hadn’t been communicated to candidates until two or three years ago. In order to level the playing field, now that has been communicated to candidates, they’re using the same technology to make sure that they get past that ATS.

That they get past that first round of being cut out of the process, because maybe they didn’t use the right keywords. And I think it’s important that talent acquisition leaders, as they invest in this technology, also invest in ways to make sure that candidates are up to speed on why a company is using this technology, how it’s using this technology, so that [00:13:00] they can level the, take the time to level the playing field on their side and prepare themselves to be evaluated by this technology.

So those are the three things that we typically educate TA leaders on.

William Tincup: I love that. Okay. So two questions that I want to get your take on. One is I’ve been thinking a lot about AI in the digital divide in the sense of, okay, we can look at it in candidates. We’ll just say white versus marginalized.

Keep it real simple. Are white candidates Gonna have a better understanding of how to use AI to their betterment, betterment than marginalized like that divide, do we see that growing? Do we see anything that we can do to, if that is growing, is there anything we can do to stop that or to mitigate that risk?

And the other is now from a company perspective, we’re all recruiting the same talent Is there something that we, again, that’s just the war for talent. If somebody, if one company and one recruiting TA team is better at leveraging AI than [00:14:00] another, but it also creates another divide.

And I don’t know if we care as much about that divide. It’s just that’s life. But I wanted to get your take on that first.

Albrey Brown: Yes, so those are two really great questions, and I don’t think they’re mutually exclusive I think the first one about the digital divide I’d say there’s three questions I ask myself that can help me define how the digital divide is going to affect different ethnicities, different backgrounds, and I think it comes down to when I think about the folks who are displaced by technology or left behind or folks who don’t have the means and don’t have the don’t have the means and don’t have the enablement to or incentive to take on a new technology.

So when we think about means, I think about folks who don’t have internet computers, the basic infrastructure that you need to leverage AI in your day to day, and that definitely affects underrepresented [00:15:00] folks more than over, over represented folks, and I’d put, white men into that category of over represented.

I also think it’s a, it comes down to a socioeconomic problem, which are tied to race, ethnicity, and gender, right? And age. I’ll give you an example. My aunt… Didn’t know what chat GPT was until about a month ago. And that’s not anyone’s fault. It’s just by Her either the higher probability given her age that she’s not going to she’s not going to figure out how to leverage that technology that said it is a really big there is a really, there’s going to create an even larger gap between those who those who understand and use this technology and not, and to me, it comes down to the fact that, if you technology at the end of the day is the magic that makes things more efficient for those who use it.

And if you don’t have the means or the infrastructure to use it, then someone is always going to be faster and better [00:16:00] than you at learning, executing, getting results, and then putting those into practice. And I think, I just as a personal example I’d be my first job in tech was as a software engineer.

And I went to a coding bootcamp. I learned how to code. And when I learned how to code, I didn’t just learn how to build build applications. I learned how to use the internet. I learned how to take a really novel problem and turned it into one sentence that will allow me to find what I need on the internet to solve that problem.

And it’s crazy to me how much that skill has impacted my ability to be successful when I haven’t built an app since. But I can, if you just ask me, Hey how can we find the how can we make sure that we have all the dimensions in this, in the room that we’re in so that we can get all the right so we can get all the right furniture I can distill because of software engineering, I can distill that problem into very simple questions that will allow [00:17:00] the internet to solve that for me.

We extrapolate that to a job. And we extrapolate that to to the problem solving in the real world. That one skill has changed the trajectory of my career more than anything else. So I think when we talk about AI and it’s accelerating the digital divide, I definitely think it’s going to have a, an exacerbated effect on inequality moving forward if we don’t take steps to make sure everyone has accessibility and enablement.

And then leading to your second question, which I’ll answer really quickly, is teams that the, although I think that there are major downsides to can there are potential major downsides to using AI and recruiting, especially for underrepresented folks, I think that the opportunity to use it and have a recruiting team use it.

And how much better that recruiting team will be leveraging AI is so great that I would never tell a talent acquisition person not to invest in it [00:18:00] because the difference between in five years, the difference between a team that has mastered how to use artificial intelligence and a team that hasn’t.

It’s going to be ridiculous. It’s going to be exponential. What the team who can that has enabled use uses AI is going to be able to accomplish with less people, less time, less money cause that’s what these tools are for.

William Tincup: Love it. The last thing, and we could elaborate on everything you’ve said, but I do want to get this in is how should practitioners think about asking questions of vendors around two things, ethical AI and audited AI.

So what should be their question base to ask questions? And again, we’re early in all these journeys. So no one’s got this nailed down and I get it, but if you were scripting this for them, what are the questions that you would like to field? Thank you.

Albrey Brown: Awesome. And you’ll notice that I like to talk in threes.

It’s the marketer in me. Yeah. Yeah. No, there are three questions I would ask.[00:19:00] When it, then first is how does your algorithm work? Can you explain to me in very simple terms, what types of artificial intelligence you use, how it makes decisions and and how it makes decisions just to get a basic understanding.

And the more that you do that with multiple vendors, the more that you’ll understand that they’re only. 10 different techniques and five different technologies that can be used that are applied to recruiting, and you’ll get a basic understanding of those

William Tincup: Go through those again, real quick over the questions, how the algorithm works, how is your algorithm,

Albrey Brown: how does your algorithm work?

Can you give me, how does it make decisions? And how does it how does it make decisions? And I even say how does it record those decisions? How are they presented to the recruiter or my teams?

William Tincup: And what I’d say is, and you probably get this too is show me an example of that. Exactly. Okay.

Albrey Brown: All right.

Good. My, my second one, to your point, great segue is.[00:20:00] Can you provide me with a report on how those results different differ based off of the different types of data that are put in, meaning the different types of candidates, right? Give me an understanding that there’s a great story about Amazon that in 2016 built an algorithm.

You probably have heard this story a ton, right? Immediately started to downgrade women engineers that identified as women and because it thought that sororities and women’s groups, et cetera, et cetera, because they were less represented, we thought that was made them a worst engineer. So show me.

The data because you should have a large set of data. If you’re talking about using an algorithm and how and the results that data has created in the past, the results of that data that’s created in the past based off of the different subsets. And if a company says we don’t have that’s a red flag because they should at least be auditing.

The data that the, or the [00:21:00] algorithm that they’re using against the different types of data that they’re inputting. And then the third piece is what does diversity look like on your team?

William Tincup: How is your team… Define diversity.

Albrey Brown: And they should define it. Yeah, if they should. And even if diversity is diversity of thought, which I mean, honestly, I hate that phrase. I know me too. But even if diverse, there should be that really challenges the vendor to they have thought about how they define diversity, how they’ve created diversity and how that diversity has been applied to the out technology that they’ve built.

William Tincup: Yeah. The only thing I’d add to that, and it’d be additive is the and how they’ve thought about bias. So how they’ve thought about diversity, I want to get their take, just their overall take okay, how do you do it internally? Let’s just let’s stop talking about tech for a second.

How do you put these things in a place in your own organization? How do you think of it theoretically and in practice? Great. And then I would [00:22:00] flip the conversation. Okay, now let’s talk about bias. Like how much have you thought about bias? What is your current thought about bias? Like where, because again, the more we peel that onion, where we find other biases that we weren’t even aware of before.

Okay,

Albrey Brown: cool. 100%. I think I’m adding that to my talk track right now, because your point, it’s not about catching a company. Yeah. In an aha moment. Or in a, it’s more if you’re thinking about this, I can trust that if I come to you with feedback. Or.

William Tincup: You’re open to it.

Exactly. Yeah. Yeah. This has been fantastic. I could talk to you forever. Thanks so much. And again, we’ve already got five other podcast topics already nailed down. So Aubrey, I got it the second time. Yeah. Thank you so much for coming on the podcast.

Albrey Brown: Thank you. Thank you. Again I’ve been listening to you for a long time and learning a lot from you.

It’s you and my

William Tincup: mom.

I get questions from my mother on the weekends. Okay, so what was going on here? I don’t understand

Albrey Brown: that. Yep. Let’s get her on [00:23:00] the podcast. The next one we can talk about parenting the entire thing.

William Tincup: All right, my friend take care. And thanks for everyone listening until next time.