Johnathan Kestenbaum (00:00):
All right. We’re going to get started. Everybody, thank you for coming today for this webinar, where we’re going to talk about an AI-first approach to recruitment. We’re really excited to be able to do this in partnership with Eightfold. For those of you who don’t know who Talent Tech Labs is or who Eightfold is, we are going to give a quick two second overview at the beginning of the presentation. And then we’re going to dive into the content. Caroline, you can move to the next slide. Caroline, one more slide. Thank you.
Johnathan Kestenbaum (00:42):
So as I said, just a quick introduction. My name is John Kestenbaum, I’m here on the left side of the screen. I’m the co founder and managing director of Talent Tech Labs. I’m here with David Francis, our director of research and he’s going to give a briefing on the report that we wrote in partnership with Eightfold in just a bit. And we also have the privilege of being with the president of Eightfold, Kamal Ahluwalia, who also is going to contribute to today’s webinar.
Johnathan Kestenbaum (01:16):
And we have two special guests, Britt Thomas and Matt Hill. You guys will get to meet them later today. We’re going to run a panel and get to ask them questions about one, how they’re leveraging Eightfold, but also what they’re seeing in the space. So really excited to be here today. I just want to say we want this to be a collaborative session. So if you do have questions and you want to use the Q&A window on the bottom of the screen, feel free to do that. We will be looking at it. If you want to use the chat as well, you can use that as well. So we’re going to be looking at your questions as they come in. And we will definitely have an opportunity at the end of the presentation, assuming time allows to have some open questions. So without further ado, let’s dive in.
Johnathan Kestenbaum (02:11):
Quick overview on Talent Tech Labs. For those of you who don’t know us, we’re on a mission to elevate the state-of-the-art recruitment technology. We do that by advising heads of talent and CEOs of staffing firms, what technology they should leverage as they go through a digital transformation. I just want to give you a little overview about how the machine works. Caroline, you can go to the next slide.
Johnathan Kestenbaum (02:36):
We’re out in the market, we’re tracking as many talent acquisition technology companies as we can get our hands on. To date, that’s north of 2500 solutions. And we have a team of analysts that are meeting with those solution providers, understanding their features, their functionality, the revenue models and really how they’re working in market. And frankly, our team is one level of analysis. We also put these technologies in front of heads of talents, CEOs of staffing firms, and we really understand how they’re being leveraged, what features and functionality are being paid for, where they’re generating their revenue from. We’re able to understand competitive positioning, and really help buyers of talent acquisition technology understand how they can leverage these tools.
Johnathan Kestenbaum (03:18):
The way we deliver our intelligence in the market is through a research membership, and we operationalize all of our intelligence into research reports. Today, we’re going to be sharing with you a free research report we wrote in partnership with Eightfold, about a topic that’s near and dear to my heart, near and dear to every TA leader’s heart, which is AI, and its impact on the recruitment function. But we also have a number of other research reports that we have behind our paywall. You can go to the next slide.
Johnathan Kestenbaum (03:49):
I want to just show you one piece of research that we call the talent acquisition taxonomy. This is really the glue that ties Talent Tech Labs together. It’s the translation, a paradigm that we use to help buyers and builders communicate with each other. So this is one of our many pieces of research. Today, we’re going to likely call out some of these sub verticals, the bubbles in our presentation. So if you ever want to understand what these bubbles actually stand for or what they do, you could always go to our website. But we’ll do our best to call them out as we leverage them. So as David gives his presentation, we’ll do that.
Johnathan Kestenbaum (04:35):
Just to give you an idea of the structure of the day, first I’m going to hand it over to Kamal to introduce Eightfold. Then we’re going to dive into about a 15 minute overview of the research report. So give you guys the intelligence, down and dirty in video form. And then we’re going to go into a panel where we’re going to get to talk to practitioners leveraging technology in the market. And we’re going to really understand what’s working for them and what’s not. So without further ado, let me hand it over to Kamal to introduce Eightfold.
Kamal Ahluwalia (05:10):
Thank you, Jonathan. It’s been fantastic partnering with you and David on this. And I’m actually looking forward to hearing Britt and Matt share their perspectives. So, a quick thing about us. We are an AI company. We have built a talent intelligence platform that covers everything from hiring internal mobility, diversity, and soon adding the contingent workforce to this thing as well. And then the use case that we need to address for the CHROs and HR leaders.
Kamal Ahluwalia (05:43):
We’ve raised about 85 million so far. Foundation, IVP, Lightspeed and Capital One are some of the key investors. More than that, it’s the adoption amongst enterprises. Today, we have customers in four continents, 25 countries. We support 50 languages and we have users in 110 countries. So the growth and the interest has been very fast and essentially global.
Kamal Ahluwalia (06:13):
And second thing that I’ll share is essentially, what we have done with our AI platform. One, when COVID hit, we did work with McKinsey to actually build an adapter technology to solve for the companies and the individuals impacted by COVID, and all the businesses that were impacted. That has actually gone pretty well, and a lot of the folks are finding opportunities so that we can continue on the employment side. So our commitment to solving employment with a right career for everyone is actually, we keep sticking to our word and the mission.
Kamal Ahluwalia (06:51):
The second part is something that we’re very proud of is we did participate in a challenge run by Department of Labor to help veterans transition into the civilian workforce. We actually competed with 50 other vendors and we won the whole thing. Again, it’s a testament to there’s a well-trained workforce that has both the hard skills and the soft skills, especially around leadership, working in ambiguity, as well as actually being as well organized as you can imagine. So things that a lot of the employers look for, all our veterans have. So we again, repurposed that technology to make sure that we could provide them the best career opportunities. So we’re working with them as well.
Kamal Ahluwalia (07:38):
And all of this is simply to the point that I think as employers, we all have opportunities to do a much better job than we have done in the past. And we want to bring our AI chops from companies like Google, Facebook and YouTube to actually really help the individuals, whether they are employees or they are outside candidates, all of it. So with that, I think it’s been fantastic to partner with you and David, and would love to get into the discussion on how do we actually make it real for everyone.
Johnathan Kestenbaum (08:07):
Thanks, Kamal. So without further ado, I’m going to hand it over to David Francis, our head of research at Talent Tech Labs to dive into the report that you guys will all get a copy of. We actually had a question. Will we share the recording? Sure, happy to share the recording and the white paper with everyone on call. Take it from here, David. You’re muted, though, David.
David Francis (08:30):
Ah, thank you. Appreciate that. Thank you so much, Jonathan and thank you, Kamal. Caroline, you can move us forward. I’m going to spend a few minutes talking through a report that we just released, as Jonathan mentioned, on AI, first recruiting. This is publicly available. We’re going to send a copy of the full report to you for attending.
David Francis (08:47):
Really, my goal here today is I want to give you a broad lay of the land in a sense of the capabilities that exist around AI today, and also where things are headed. So ideally, by the time that we get to the panel discussion, you’ve got a general sense and a common language to be able to speak to you. And then we can get into specific use cases and nuances with our panelists. You can move us forward, Caroline.
David Francis (09:14):
I want to start from 30,000 feet up by just defining what AI is and what it’s capable of. And the reason for this, I think arguably, AI or artificial intelligence is the most overused term in HR technology. And there’s not really a common understanding of what the capabilities are. So if somebody says we’re in HR, or we’re an AI vendor, what does that actually mean, practically? So we’re going to do a quick crash course in AI, so we’ve all got a common understanding. Then we’ll look at some common use cases in recruiting and how the space is evolving. You can move us forward.
David Francis (10:00):
In terms of what AI is and how it operates in the market today, there’s three broad buckets that solutions fall in. And this isn’t just specific to recruiting. So we’re going to start from, again, 30,000 feet, the most generic, and then we’ll start whittling down into recruiting solutions. The first bucket is what we call narrow AI. And what we mean by narrow AI, it’s essentially a bespoke model or a bespoke algorithm that’s been trained typically on one data set, which maybe a very big data set, but typically one data set to solve one specific problem.
David Francis (10:32):
An example of this might be, say, natural language processing, or the ability for a computer to understand human text. A recruitment chat bot, for example, uses natural language processing, but it can really only understand conversations in a job context. If you ask it about the weather or the politics, or who I should vote for, you would break the system essentially. Now, these systems, they’re unique in that they can adapt and learn based on your data. And the vendor or the programmer can essentially adjust the underlying algorithm in order to align it towards desired outcomes. So basically, these can self-adapt and they can also be fine tuned for lack of a better word. But again, this is everything from recommendation engines on Netflix and YouTube, to a recruiting solution that, say, unearths candidate potential.
David Francis (11:30):
Now, the future is, I think what most people conceptualize as AI, which is this general sentient system, which is essentially in technical terms, one data model but multiple domains. So instead of having to train an algorithm say, with millions of data points, you could just say, show a system one or two examples and it would be able to understand and learn, and be broadly applied to multiple domains, to recruiting and to conversation or to image recognition, or all these different things. In practice, this doesn’t exist yet today.
David Francis (12:03):
There’s no practical examples in the market. There’s a lot of great work being done here, actually some incredible work being done here. And I think probably within the next 10 to 15 years, these systems are going to come to market. And what that will most likely look like is the big tech players, like maybe a Google or a Facebook or an Amazon, or an open AI, or maybe an Eightfold is going to have AI as a service, where essentially solutions are going to be built on this generic AI. It’s happening, but we’re ways off from that state.
David Francis (12:33):
What is happening, and what’s the most advanced state right now is what we call an AI platform. And essentially, what this is, is a system architecture where you’re applying many bespoke point solutions into a platform architecture. So you’re more or less emulating a general AI, but the way you’re getting there is through many bespoke point solutions that have been put together. So it’s an AI-first approach to systems implementation, and it’s a metaphor for what this might look like.
David Francis (13:06):
Google recently launched a … not recently, a few years ago, it launched its own cell phone. And it has basically, a very cheaper version of an iPhone. It’s about a third of the cost, but it competes toe to toe in many of the same features that you might get from a top end phone, by using AI. So for example, it’s got a weaker camera, it’s got a weaker processor, but the pictures that it takes are just about as good as a much higher end camera. The experience that happens throughout, it’s supported basically by AI. So it can get by with the weaker system, but at a significantly cheaper cost because of this AI-first platform architecture. This is not a commercial for Google or anything, but I just wanted to give a non recruiting example of what that might look like in practice. You can move this forward, Caroline.
David Francis (13:57):
Now, we can get to recruiting. So the smoke and mirrors versus actual AI. What are the applications in recruiting in our business in hiring people? It turns out we did a survey a couple of years ago of talent acquisition technology vendors, and more than 90% of companies told us that they’re either using AI as part of their solution, and the companies that weren’t doing it, had it on their roadmap. So if every company is doing this, how do you separate the wheat from the chaff, or the noise from the reality? You can move us forward.
David Francis (14:33):
What we found is that AI manifests itself in a number of bespoke areas throughout the recruiting process. So getting back to this idea of narrow AI or bespoke point solutions, these are the areas where we see AI making an impact in recruiting solutions. So things like AI-based job matching, which typically uses a learning model in order to stack rank candidates based on their experience, their potential, other factors against open jobs. Things like sentiment and textual analysis, companies like say, a CAP recruiter or a Textio that can analyze job descriptions and make specific recommendations around how you can approve say, diverse hires using those tools.
David Francis (15:18):
Or things like conversational AI which can emulate the experience that you might have interacting with a human source or recruiter. So on the one hand, when a company says we’re an AI-based solution, the response you probably should have is, “Well, that’s great. But actually, in practice, what do you actually do?” But on the other hand, we get to this concept of AI-first recruiting. Can you move us forward? So what is that? Move us forward one more time.
David Francis (15:50):
And the answer is really basically, applying multiple of these domains or solutions together into a platform. So the idea is that this is starting to happen today. There’s examples of this and we’re going to hear from a company that’s actually starting to do this. But the idea is that you have one platform, many solutions. And it’s important to caveat here that there’s no one vendor that’s going to basically do everything for you. And any company that claims to do that right now is either lying to you or ignorant. So this is more a future state that the industry is evolving towards, than something you can just go buy right now and implement, and have it solve all of your problems. But this is the way that the industry is moving towards. Move us forward.
David Francis (16:38):
So another, I guess in terms of, how do you determine an AI-first recruiting platform versus one that’s not? It might be considered a matter of semantics, or you might think of it as a matter of semantics. But we think this actually manifests itself in practice in a couple of ways. The number one thing is just in terms of how the product is developed and actually built. An AI-first solution is going to tend to have a lot more technical people on their teams, a lot more data scientists, software engineers and the like. We’ve seen some talent acquisition technology companies that are 70% to 80% sales people. Whereas a solution that’s AI-first, it’s going to tend to have more technical folks.
David Francis (17:21):
It also manifests itself in the product roadmap. Particularly for newer companies, this is arguably more important than what product they actually have today is, what products are they going to have in the next year or two? So the solution approach for AI-first, companies tend to incorporate AI as opposed to just throwing working units at the problem. And then obviously, you’re going to want to look at the number of domains or specific business problems that are addressed by a particular solution. Move us forward.
David Francis (17:51):
And the report itself, I’m going to save some of this for the panelists to discuss here. But in terms of what are the benefits of these solutions, the baseline metrics you’re going to want to track are similar to the metrics you’re going to track for any technology implementation you introduce. Things like the improving your time to hire, the time to interview, your cost per hire, drop off rates and candidate quality, which you might measure in terms of retention, post hire retention and maybe hiring manager NPS.
David Francis (18:22):
And the report itself, we break out some benchmark baselines you can use to track each of these. But we’ve seen improvements from companies that have gone from either no system at all, or maybe a legacy system to an AI-first solution. There’s been dramatic increases in performance across the board, from a recruiter efficiency to sourcing efficiency, career site improvements through increased candidate convergence, and even being able to move the needle on diversity hiring. Move us forward.
David Francis (18:54):
The last point or comment I’ll make here is around AI ethics, and this isn’t something you’re thinking about today. It’s something whether by choice or necessity, you’re going to have to think about. So I’d encourage you to start considering these things now before you’re forced to. And really, what I think this is, it’s basically, what output should an algorithm be designed to optimize? And what means or inputs are we as a hiring manager or a recruiter, or as a company, or as a society are we willing to accept in order to achieve those outcomes? So it’s important to note that AI in and of itself is essentially just a software program. It’s zeros and one, and so it’s not inherently good, it’s not inherently bad.
David Francis (19:42):
But if it’s not thoughtfully implemented, it’s got the potential to potentially amplify biases, or it’s got the potential to decrease biases. So in practical terms, you’re most likely not going to be able to go in and actually recommend code to your vendors, but what you can do is know the right questions to ask when you’re evaluating companies in order to make sure that they’ve got the processes in place to align with your company’s overall strategy. So a handful of things you want to look out for. One is algorithm transparency or justification. And essentially, what this refers to is, there’s this black box problem sometimes with the way that inputs and outputs are disassociated from each other. So you want to make sure that there’s a relatively clear line between the inputs that go in to a system, and then the outputs that you get out. So there’s some explainability or justifiability between those two things.
David Francis (20:44):
There’s a whole area of equal opportunity algorithms which are coming up, which would essentially, remove characteristics that are protected from a legal standpoint, from decision making. And so to the extent that you can do those, you want to remove those as well. So with that, I think we can tee up our panelists now, and I’ll hand it back to Kamal to introduce Britt and Matt.
Johnathan Kestenbaum (21:09):
I’ll take over, actually. Thank you, David. So we’re going to go a bit deeper. We’re going to dive into a panel, we have a set of questions. I see some folks have been asking us questions with the chat function. I encourage you to if you want to go deeper. And like I said, at the end, we’ll have the opportunity for more of an open floor.
Johnathan Kestenbaum (21:31):
But without further ado, I do want to introduce our celebrity guests that we have with us today. And so let’s start with Britt Thomas. Britt is the global director of talent brand and technology innovation at Micron. Britt, why don’t you just introduce yourself for a second?
Britt Thomas, Micron (21:48):
Hi, Britt Thomas, Micron. We have 38,000 global employees, a very large global organization. And I’m happy to be with you today. I’ve been part of the Talent Tech Labs for about six or seven years now. So thank you for having me on the panel, and thank you for letting me join, Kamal, in Eightfold.
Johnathan Kestenbaum (22:06):
Awesome. And next up, we have Matt Hill, who’s the director of talent acquisition at Dexcom.
Matt HIll, Dexcom (22:12):
We’re not on the Fortune 500, like Micron, but we are the fastest growing stock in the S&P 500 this year. And I’m the director of talent acquisition at Dexcom. So we’ve been growing extremely fast for the last five years. And so really, there’s high volume activity across all the recruiting segments.
Matt HIll, Dexcom (22:27):
We’ve really accelerated through COVID and have kept up the volume and even increased it. Really, we’ve had very few ways of looking into a single talent network and we really wanted to shift some of that investment into active sourcing tools. So excited to be here to talk about what we’ve seen.
Johnathan Kestenbaum (22:42):
Thanks, Matt. Thanks, Britt. So let’s dive in. I really want to understand what AI tools and technologies that you’ve leveraged. And really what’s had the biggest impact on your recruitment function? Why don’t we start with you, Britt?
Britt Thomas, Micron (22:56):
Sure. At Micron, we do use Eightfold. We’re proud to say we use the Eightfold AI talent intelligence platform. And when we say platform, it really is a platform because we use it in many different ways. We do use the external personalized career site, which also has an automated bot as well for candidate questions. We do use Textio as well for gender neutral job descriptions. We feel like that fits nicely with the very beginning part of our hiring process.
Britt Thomas, Micron (23:28):
But then we use the AI tools and the calibration and the functionality offered by Eightfold to really get a better perspective on not only the applicants that have come in, so it’s fully integrated with our applicant tracking system, but also to reevaluate some of the prospects, people that haven’t actually applied, but we can still calibrate them. And we can still look at the skills and the qualifications mapping, and put those then in front of our hiring managers. We can also do those in a way that’s anomized so that we are showing hiring managers actual qualified and skilled applicants and prospects without some of the bias being included as well.
Johnathan Kestenbaum (24:12):
Thanks, Britt. What about you, Matt?
Matt HIll, Dexcom (24:15):
I’m a big science fiction buff, so I’m reading about AI almost every day. But for me, in terms of my practical use of AI, it actually started about six years ago. I was working at Qualcomm, we had an internal R&D team that was playing around with machine classifiers and machine learning. And they’d built a tool called sorting hat for you Potter fans out there. So it was actually a manual machine training tool where we could actually take data and we could train it up to say, this is good, this is not good.
Matt HIll, Dexcom (24:41):
And we actually used it to look at whether we could train a machine classifier for experienced software engineers and also for new grad software engineers. And so that was a really interesting experience. Very manual though, very time intensive. Fast forward to today, we’re basically using from an Eightfold perspective, the personalized career site as well as the talent acquisition tools, and then some of the talent experience tools for internal mobility as well.
Johnathan Kestenbaum (25:05):
Thanks, Matt. Quick question for you, Britt, we got from the audience is, what ATS is Micron using?
Matt HIll, Dexcom (25:12):
Currently, Micron uses SuccessFactors as our applicant tracking system, and it is fully integrated with our Eightfold platform. And we will be moving towards Workday in 2021, so we are working through how those integrations will work.
Johnathan Kestenbaum (25:33):
Thanks, Britt. There was a couple of questions here. I just want to highlight that we’ve received them. So Hannah Waters asked about diversity dashboards. So Hannah, we will touch upon that. And I would say specifically, I personally think that’s super important from what we’ve seen at Talent Tech Labs to some of our clients.
Johnathan Kestenbaum (25:51):
It’s really about social accountability, and diversity dashboards help you with social accountability, particularly because it’s really hard to get somebody to … You can’t force somebody to hire diverse talent. And if you could hold them socially accountable through a dashboard, it helps go a long way. So we will spend some time talking about that.
Johnathan Kestenbaum (26:12):
And separately, there was a question on roadmap evaluation criteria for selecting AI. That was Julie French. We have that on our series of questions here, so we will dive into that. So we can go to the next question. Kamal, this is really for you. How do you get the best insight for the candidates in hiring manager so everyone’s on the same page?
Kamal Ahluwalia (26:39):
I think there are a couple of things, Jonathan. One is that the job description itself, it all starts from there. Has that been calibrated properly? That becomes a very useful function for the TA person or the recruiter, or the HR business partner to work with a hiring manager and get the job description right. Also, identify ideal candidates because that’s what we all look at when we’re hiring. I want two more Davids on the team because he’s doing such a great job, stuff like that.
Kamal Ahluwalia (27:09):
Second element is transparency in the AI recommendations. Whenever people are applying, for example, let’s just take the personalized career site that we talked about, when people upload their resume, they are presented with jobs that are a strong fit. And it explains why. That builds confidence for the candidates so they apply to the right job. Now, the same information is also available to the hiring manager. They’re also seeing exact same things that this person is a good fit. And we have these four buckets. One is validated skills, likely skills, missing skills and skills to validate.
Kamal Ahluwalia (27:51):
So the philosophy here is trust, but validate. And we have scoured over a billion profiles to understand better and better insights into who’s capable of doing what. And that’s what we apply to show to the hiring manager or the recruiter that, “Hey, here are the areas where we know that we have validated their claims on their resume with externally available data, here are the areas that they didn’t specify in their resume, but they’re likely to have those, and here’s stuff that you wanted, but they don’t have and you should be digging into that if it’s important, and here’s stuff that they are claiming on their resumes but we need to verify that in the interview process.”
Kamal Ahluwalia (28:36):
And that, if you look at it, a resume or even a LinkedIn profile is a sales-attested document. And without AI and underlying data sets and models, that’s what we are able to do with transparency to both sides. That’s what’s driving the numbers.
Johnathan Kestenbaum (28:50):
Kamal, while I have you on the line here, we did get a question, if Eightfold has ATS capabilities.
Kamal Ahluwalia (28:59):
For religious reasons, we stayed out of the ATS and HRIS stuff. It’s like if you have what’s working, keep going. Let’s solve for the people and the hiring and the talent piece. So specifically, no. The experience for the recruiters is about 80% to 90% of the jobs and the day is in Eightfold. So we are providing the AI layer that sits on top of ATS and HRIS.
Johnathan Kestenbaum (29:24):
Thanks, Kamal. Britt, anything you want to add about the insights that you’re providing to hiring managers or candidates throughout the process?
Britt Thomas, Micron (29:33):
I’ll add on to what Kamal shared that when we do calibrate, and we do look at the skills that align or that may be missing, we use that in a way that we hadn’t been able to before. When we would look at a job description, we know that most individuals don’t meet every qualification on a job description, and hiring managers are very familiar with that scenario. But now, we can tell them and show them where they don’t align and where they may need to focus. If they choose that candidate for their role, where they need to focus on developing that individual in the role as a new hire as well.
Britt Thomas, Micron (30:12):
So it’s not just matching the skills before they join the organization. It’s also understanding what skills they may need to add after they’ve already joined an organization as well. So there are insights on skills and qualifications. I know somebody in the chat asked about the diversity dashboard, we use that as well. Again, we feel the diversity dashboard offering that we receive from Eightfold is far superior to what we were getting from our ATS, and possibly from our future ATS as well.
Johnathan Kestenbaum (30:44):
Thanks.
Matt HIll, Dexcom (30:46):
Just to ride on that, I do think calibration is a really key piece of that insight. And not only that, in terms of improving the matching, but really it improves the procedural justice of the AI. Everyone wonders, what’s in the algorithm? What’s actually going on? So if you can bring a hiring manager and a recruiter together to really look across, and really help to train the machine on that individual basis as well, that’s really important. And I think it also adds confidence in the recruiting team for the hiring managers to see that we have these tools at our disposal.
Johnathan Kestenbaum (31:17):
Thanks, Matt. Actually, we have a great question from my pal, Kevin Wheeler. Kevin, hey, how are you? I hope you’re doing great. Excited to have you here. For Kamal, on how long does it take to calibrate the algorithm to deliver the accurate results?
Kamal Ahluwalia (31:32):
Out of the box, we are about 90% to 91% accurate, and it takes a couple of weeks. So once we pull in your data and integrate with ATS and whatever other legacy data sources are there, then it’s a couple of weeks to actually work with some of your experts to say, “Hey, this is what we’re seeing. Are we getting it right?” But within a couple of weeks, we’re off to the races.
Johnathan Kestenbaum (31:55):
Thanks. Let’s go to the next question. I’m sure everyone has still the remote workforce on their mind. And Kamal, I know you had mentioned at the beginning that you guys are thinking about the contingent labor. Are you building any functionality to service this category of talent?
Kamal Ahluwalia (32:17):
Yes, there are a number of things here. Remote work right now, I think is here to stay. I think more and more companies are watching but are not the ones who wanted more people in our office or different offices, because I think it’s better for the culture. But unfortunately, we are all having to adjust and still be productive.
Kamal Ahluwalia (32:37):
So what we are doing is this. If you look at the entire spectrum, if we are continuing to hire like these two companies have tried through COVID, because they’re both running essential businesses, how do we actually do all the events remotely? So even the campus recruiting, to interview scheduling, all of that we can now handle with our platform. Then even the interview feedback because you can’t just walk down the corridor and get people’s input, now the interesting part is because everything is remote and there are no onsite interviews being scheduled, things are actually moving faster now.
Kamal Ahluwalia (33:13):
So on each of those things, we are adding more. So scheduling, event recruiting, all of these things are there. Second part is back to the calibration, looks like we can now hire people anywhere in the country. And in their cases, anywhere in the globe. So the aperture is much wider because you’re not limited by your geographic offices. That means there are more talented people available, and you can use it to get through that increased volume very quickly. Third element is if you’re really serious about DNI, this opens up tremendous opportunity to find additional talent that you wouldn’t have looked at earlier.
Kamal Ahluwalia (33:52):
Either community colleges, other black colleges like Howard University, a lot of our customers are actually starting to make it easy because they’re not limited by how many universities they can go to, they can look at everyone. And also, how to actually identify and move them through the process and master profile so you’re really giving everybody a fair shot? So I think remote work has to some extent been a catalyst on the transformation that TA leaders were looking for anyways. So some of those things that we are actually addressing, even hiring based on zip code and the time zone, not necessarily based on the geographic location.
Johnathan Kestenbaum (34:30):
One question while I have you, and then I’ll dig a bit deeper with Britt and Matt. Do you guys have a chat bot?
Kamal Ahluwalia (34:37):
We do. And it’s live nationwide and in some of the other places, yes.
Johnathan Kestenbaum (34:42):
Awesome. That was a question from the audience.
Britt Thomas, Micron (34:44):
How about Micron?
Johnathan Kestenbaum (34:47):
Britt, do you want to add around the remote workforce?
Britt Thomas, Micron (34:54):
I do think we’re going to continue to see a rise in remote work. I think the technology is helping us recruit faster. Although companies may have slowed down, they are still hiring for critical roles, and we’re doing that faster than ever. There’s not a delay in scheduling interviews and running back and forth to the offices right now, so that’s a wonderful thing for our talent.
Britt Thomas, Micron (35:16):
It also ensures we can interview individuals who are currently employed. That’s always been a struggle in the past when someone’s actively employed, exploring another opportunity and concerned about missing work or giving the impression that they are interviewing.
Britt Thomas, Micron (35:31):
So we use every technology that’s available to speed up the hiring and ensuring that we can meet with virtually, the talent that we need. Again, I think everybody knows talent is everywhere, but opportunity still is not. And this helps us enable the continued communication with that talent.
Johnathan Kestenbaum (35:52):
Thanks, Britt. We’ll go to the next question. One of the big issues that come up more and more these days is how you’re leveraging technology to improve diversity in the hiring process. And now, I want to just give an insight here, about what we’re seeing in the market. As I’m sure you guys can understand, and most of this functionality is actually built within the Eightfold platform, there are certain pieces of technology that you can leverage.
Johnathan Kestenbaum (36:27):
One would be hiding gaps in employment, hiding things in the resume. Another would be being more thoughtful about how you manage the interview process. There are social search tools that allow you to search for diverse talent. But what we found through our conversations with clients is things usually fall short when you get to the hiring manager. And you can’t really incentivize them, and I hinted at this at the beginning to hire diverse talent.
Johnathan Kestenbaum (36:53):
And in that regard, we had a question about diversity dashboards and holding people accountable. In that regard, we find that social accountability is really what makes this work, diversity technology work. Because you use the technology to make sure you have a diversity of candidates put in front of your hiring manager, but then you use social accountability to take it across the finish line. And so I just wanted to share that intelligence. But I do want to dive in specifically, to let’s start with Matt, here, Matt and Britt and find out really how you’re leveraging technology to improve diversity.
Matt HIll, Dexcom (37:32):
Sure. In terms of just the additional context that’s overlaid on your ATS data, the visibility into different aspects like gender diversity, whether someone has studied in HBCU or a Hispanic-serving college, these are all things that are basically brought to the forefront within Eightfold.
Matt HIll, Dexcom (37:55):
So just in general, we have tried to improve the overall visibility in terms of both corporate goals, individual AAP goals and also just general diversity as well in certain areas. So visibility is key, I think. What was it, the social … was it social? I like to call it norming, but I like your social …
Johnathan Kestenbaum (38:14):
Social accountability.
Matt HIll, Dexcom (38:15):
Exactly, social accountability. But really, people want to do the right thing. So giving them the data in order for them to do that right thing is really the key.
Johnathan Kestenbaum (38:24):
Britt, anything you want to add?
Britt Thomas, Micron (38:26):
Yeah, I’ll add. One of the things we are looking at with Eightfold is another feature that we find really valuable is again, being able to anomize resumes, anonymize them, hiding information in particular that is cause for conversation. We’ve had some conversations around, why would you ever hide a hobby or volunteer work?
Britt Thomas, Micron (38:48):
Well, the reality is many of those things that people put on their resume are religious affiliated, or affiliated to their background or their communities. And they do create bias. It’s interesting that we’re having these conversations now, but we do realize you can still look at someone’s skills and qualifications and identify if they’re a fit for the role without knowing their full first name, their full last name and their address and their zip code, knowing how far they would have to commute for the job.
Britt Thomas, Micron (39:18):
So as an organization, it’s creating a lot of really dynamic conversation about some of the things on our resumes, which people have used to form opinions. And although we know we won’t eliminate bias, we hope to reduce it by inserting the technology and turning on that feature. So our hiring managers at the very first screening stage, when they first look at prospects and applicants, they’re looking at a fair representation of the right fits for these roles that they’re hiring for.
Johnathan Kestenbaum (39:53):
Thanks, Britt. Leave it to the audience to ask the tough questions. Amy Kahn just asked, she said some experts suggest promote diversity, don’t hide it. You can be more proactive. And it’s really counter to anonymizing candidate data. I would argue that it’s not necessarily black or white there, because you could do both. But is there any thoughts on a trade-off there? Either of you guys want to address that? I see, Britt. You want to go first?
Britt Thomas, Micron (40:25):
Yeah, if I can jump in. It’s a great question. And it’s something that we hear a lot from TA because for years, talent acquisition teams have been driving that diversity for the organization. And really pushing diverse initiatives, and how to push more diverse talent into the funnel. We did hear that quite a bit.
Britt Thomas, Micron (40:47):
And even with anonymizing resumes for hiring managers, we are opting through the AI technology offered by Eightfold to leave it visible for our talent acquisition team, which we know will still push those individuals into the funnel to make sure there is representation. I know there’s a lot of technology out there, some will anonymize everything across the board. Recruiters won’t know, hiring managers won’t know, and it’s really up to the organization to figure out what solution will work for your organization.
Britt Thomas, Micron (41:20):
So again, we did think through all of those things. Our recruiters will have access. Of course, the information if somebody applied and did self-declare, it is already in your system of record, which is your applicant tracking system. However, again, since our systems are fully integrated, when our recruiters send those resumes to their hiring manager using the Eightfold tool, we will anonymize some of those items I just discussed.
Johnathan Kestenbaum (41:49):
Thanks, Britt. I want to push forward to the next question only because-
Kamal Ahluwalia (41:52):
Jonathan, just one thing, because on this one, I think we can add the diversity dashboard also. What we’ll share as you go through it, and Micron already has it is everything about your funnel, top of the funnel, every stage in your process, down to every recruiter and every hiring manager. And in any case, if there is deviation from norm, it is very easy to identify who is bringing bias decision into the hiring process. And whatever is inspected, improves. That’s one.
Kamal Ahluwalia (42:24):
Second part is we do provide you with the capability, as Amy was asking of actually taking the affirmative action also. If you do want to hire more women, you should be able to find the best women in your funnel, and actually move forward on that. Same for veterans. Same for any other ethnicity that you do want to bring in. And all of those characteristics are available, because it needs to be a single holistic thing. And the same thing needs to apply to employees because that’s what Dexcom also liked is, how do I unlock all of this, all the opportunities first to my employees, or give them as good a shot if not a preferred shot, moving within the organization?
Kamal Ahluwalia (43:05):
So all of that is built in. It’s not sprinkled on top, it is part of that. And nothing about an individual, age, sex, any of those things are taken into the match score. That’s what gives you what people are capable of doing, and it is all data driven. And then you marry the company policy with what you’re trying to solve. And now, you have something that speeds everything up.
Johnathan Kestenbaum (43:30):
Thanks, Kamal. So a few things. I want to address some of the questions that were asked, and then I want to pump through a few more because I want to be sensitive that we get through all the questions in time. First is Heidi. You asked about visual reference to diversity dashboards. What do they look like? I’m sure Kamal can share that with you offline.
Kamal Ahluwalia (43:48):
Yeah.
Johnathan Kestenbaum (43:48):
But on a separate note, we have a report on diversity that we recently wrote, where we dive into specifically, all these different technologies that you can leverage to solve some of your diversity problems. And we did a work group on it. So you can download that actually, on our website, the information on that.
Johnathan Kestenbaum (44:08):
Let’s go to the next question. Another topic of high interest is candidate experience, and I’m sure it’s something that everyone is thinking about. Matt, you want to dive into how you’re enhancing candidate experience in the hiring process?
Matt HIll, Dexcom (44:25):
Sure. It all comes down to a one to one comparison; searching versus matching. So in terms of the candidate experience, instead of going into a website and searching for random terms that you think might be relevant to your background, or just literally just running through everything until you find something, you literally go in, upload your resume and then you get a stack ranked set of jobs that actually match your background.
Matt HIll, Dexcom (44:51):
So from a candidate experience perspective, it’s mind-blowing. Literally, it’s a couple of clicks and you have not only a set of jobs that look like they fit your background, but also descriptions why, down to the skill level. Other candidates have been hired from these similar companies with these similar titles. So it just erases a lot of the information asynchronicity in that process, which is something that the candidates find it so hard to overcome. So for me, that’s the key thing.
Matt HIll, Dexcom (45:17):
Actually, I try to stay close to our candidates. So I remember getting an email from someone saying, “Hey, I just want to get some guidance on how to apply for the right jobs.” And I said, “Well, just use our system. It’s going to tell you.” So we had an email exchange and I said, “Well, what do you think about the experience, because it’s fairly new?” And he said, “I specifically remember thinking that your career sites matching technology was very unique and novel when I did my first search. It was the reason I was able to identify so many positions that I was a fit for so quickly. I was also extremely grateful that your career site does not require filling out all of your resume work experience into the application.”
Matt HIll, Dexcom (45:51):
This is an engineer who basically was able to quickly find the jobs that were relevant to him and apply. So things like that are why we’ve literally seen a 40% conversion of all website visitors into unique applicants. It’s a no-brainer.
Johnathan Kestenbaum (46:09):
Thanks, Matt. There … go ahead, Britt.
Britt Thomas, Micron (46:11):
I would echo Matt. I think one of the things we realized being a very large global organization is we have too many applicants. And when you apply, you do have hope of gainful employment. So being able to match your skill set with jobs that you are actually skilled and qualified for is obviously an added benefit to the Eightfold technology.
Britt Thomas, Micron (46:31):
The other piece of that is it ensures that the applicants were getting in and our recruiters are screening, and our hiring managers are calibrating actually individuals we would entertain hiring as well. So again, it cuts down on all of those horrible disposition emails that applicants receive that, really 90% were not a fit for the role to begin with. But we don’t have technology that doesn’t allow anyone to apply for a role at this point in time.
Britt Thomas, Micron (46:58):
So again, when they go to Eightfold, they can not only apply to 10 jobs, but they can do it within less than a minute. So before or even looking at existing ATS career sites, you have to click on one job at a time. You have to apply for one job at a time. Often you have to fill out the 10 minute application one at a time. It was an easy sell for us for that personalized career site functionality.
Johnathan Kestenbaum (47:23):
Thanks, Britt. I just want to address some of the questions that have been asked in the QA panel. First of all, some of them are more tactical questions we can address offline. So we’re going to save the Q&A and we’ll follow up with you, integration stuff, language considerations.
Johnathan Kestenbaum (47:39):
Specifically, there was a question, Kamal, for you about Eightfold’s integrations with Workday. I’ve heard that a few times, so I figured I’ll just pause here for a second just to see if you could address that.
Kamal Ahluwalia (47:50):
Sure.
Johnathan Kestenbaum (47:50):
And then I’ll dive into the next question.
Kamal Ahluwalia (47:53):
Sure. Two-way integration with Workday, we’ve got a number of customers who are actually using us for that. [AirAsia 00:48:00] is a very good example, very global and very large volume coming in, as well as other customers like Prudential and Capital One, et cetera.
Kamal Ahluwalia (48:12):
A lot of our customers have Workday for HRIS and ATS, and career site, full talent acquisition, because the recruiting needs a little bit of work. So all of that is there. We have people on the team who have actually built Workday’s internal integration. So we know all the ways to actually integrate, and that’s actually fairly straightforward. The good part is Workday has good APIs, so that allows us to actually execute extremely well, and very quickly.
Johnathan Kestenbaum (48:41):
Thanks, Kamal. Let’s go to the next question, but I’m going to start the next question with a question that was asked a few times. I’m going to direct it at Matt. Matt, everyone wants to know what internal mobility looks like at Dexcom. We’ve got this question a few times.
Matt HIll, Dexcom (48:57):
It’s a common refrain. Like I said before, Dexcom has been growing at about 40% revenue per year for a while now, and a little bit less than that from a headcount perspective. We’re trying not to overcome that. But in terms of that, the internal mobility is huge for us. From a process perspective, we’ve tried to make it easy for employees to at least reach out and apply to jobs and see if they’re qualified. And we’ve tried to remove some of those HR and manager notifications that get built into the process, really to unlock conversations that people need to have.
Matt HIll, Dexcom (49:29):
So we’re also really getting close to implementing the internal mobility module from Eightfold, which is actually going to take, again, our ATS experience and turn it into something where an employee can go in, upload the resume and actually see what internal jobs match their background. It’s not career pathing, but it’s like quick and easy career pathing for employees to give them a sense of what jobs across the board they might match to. Not just in their specific business unit or department, but really where their skills apply across the company. And you couldn’t do that before, you really couldn’t.
Johnathan Kestenbaum (50:01):
That’s great. Thank you, Matt. Well, let me dive into this question because we’re down to the wire here in terms of time. We’ve got really an engaged community. Thank you. Great turnout today. How do you ensure that AI-based processes and algorithms aren’t biased? I’m going to point this one to you, Kamal, because you’re building the AI platform.
Kamal Ahluwalia (50:22):
Absolutely. Number one is the underlying data set. We have over a billion profiles that are global. So our models are trained on essentially, most of the workforce that’s out there. And it is not limited to a company’s employees and the transactions or the history within a company. Because when you do that, yes, you will not have enough data set to train the models adequately. And that’s why essentially, Amazon failed two years ago when it became public.
Kamal Ahluwalia (50:53):
Second element is we do have separate training data. All the recommendations that we are making, they are compared against the training data on a very regular basis. And then there are manual audits that are done on the matching to ensure that we are not mixing the decision one way or the other for a man versus a woman, all of that. And we actually have a paper on that. So if anybody is interested, happy to share. We take that very seriously, and a number of companies have run audits on our stuff to ensure that this is done right.
Kamal Ahluwalia (51:29):
But the key part is the features that you’re using, what is driving that recommendation issue? And the transparency can be explained on every single decision why we recommended this one as a three star versus a four star versus five star. That transparency is what allows us to actually get it right and be unbiased.
Johnathan Kestenbaum (51:48):
Thanks, Kamal. One more question for you and then I’m actually going to go to the next question. This is from the audience. They want to understand OFCCP compliance, how it works with matching resumes. Anything you can share on that?
Kamal Ahluwalia (52:00):
Yeah. We are compliant with that also, and we are happy to actually get into the details. And along with ADA compliance, so the user experience for people with disability, all that stuff. All of that, because of all the work we’re doing with a lot of large public companies and public sector is something that is table stakes. Happy to cover that as well.
Johnathan Kestenbaum (52:19):
Thanks, Kamal. Let’s go to the next question, and I’m pointing this one at Britt. If every vendor is saying that they’re leveraging AI in their product, how can TA teams find out what’s real and what’s hype?
Britt Thomas, Micron (52:33):
I think one of the most important things is to talk to other TA teams and to share knowledge across talent organizations and talent professionals across all levels, especially at the levels where the teams and the individuals are hands-on in the systems. They know what works. I will say look for technology that is ready now to implement, not ready in one to two years or two year to three years. We have problems today that we need technology that is ready now to solve for.
Britt Thomas, Micron (53:03):
I’ll share too, we do have the personalized career site. We have the Eightfold bot. And with the technology platform you choose, ensure you have a really strong customer success manager who you can go back to and say, “All right, now we want to take a look at the bot, we want to better the bot. Now we want to take a look at how we’re going to anonymize resumes, and we want to do that a little differently than what we’re hearing other companies do.”
Britt Thomas, Micron (53:27):
I would say look at the technology, talk to other people. Find out how they’re using it. Also, find out what it took to implement the technology as well. Again, the implementation resources are critical to whether the product is going to work for your organization.
Johnathan Kestenbaum (53:44):
Thanks, Britt. I would also argue that becoming a member of Talent Tech Labs will help you as you navigate through. It’s good to know all these. Let’s go to the last and final question, with four minutes left and then we can open the audience here. And Matt, I’m going to address this at you, really. What do you think will be the next big trend in talent acquisition technology?
Matt HIll, Dexcom (54:07):
I think we’ve reached that inflection point where a match makes a lot of sense, in terms of automating that. So I think as we continue to move upstream it’s probably going to look like some sort of automated attraction. So not just sourcing, but the ability to go out and actively attract some prospect candidates based on AI, based on certain parameters that you set yourself.
Matt HIll, Dexcom (54:28):
And then of course, the recruitment process or robotic process automation is something that we’ve been dabbling in. We just released our first bot last month to help with some of our very administrative onboarding processes. So we are the proud parents of a bot in our TA organization today, and I think that’s going to really help in some of those high volume processes.
Johnathan Kestenbaum (54:45):
Thanks, Matt. Britt, what about you?
Britt Thomas, Micron (54:52):
I think the technology trend we’re going to see is the expectation of the platform instead of separate implementations for technologies on their own. Before we would look at one technology for CRM, we’d look at one for anomizing resumes. We’d look at one for matching skills and qualifications, and another one that does internal mobility within an organization.
Britt Thomas, Micron (55:14):
And I think right now, companies are looking at, is there one platform that can do more of what I need and fill more of those gaps, as applicant tracking systems still seem to be more of a legal type system that you use for audit? It’s not as user friendly oftentimes for the recruiters, or the candidates as well.
Britt Thomas, Micron (55:35):
So, again, I think the next big trend in talent technology is going to be that follow me type of behavior that we see on the web today, a more consumer driven approach to, “Hey, Matt is really active. Matt had drives an SUV. I can tell he must have a family, and he’s probably looking for a job close to home.” And surfacing those opportunities directly to him based on his activity in his daily life.
Johnathan Kestenbaum (56:00):
Thanks, Britt. I would second what both Matt and Britt said. I would add, so we issue a trend report every quarter, which is really what we believe the next trend for the next three, four months is. So our next trend report which should be coming out, David might get me in trouble if I announce it too soon, I’d say September, is on recruitment process automation. We’re really excited about that. We think there’s a huge opportunity there. So we’re looking forward to sharing that with everybody. That will be publicly available on our website to download.
Johnathan Kestenbaum (56:42):
First of all, I just want to thank everybody. We’re here with one minute left. I want to thank you, Britt Thomas and Matt Hill for joining us. We say Kamal, thank you, thank you. A really engaging group today. I see some new and old friends on the chat window here. If anybody has any further questions, you can either go to Eightfold, we’ll reach out to you guys with our contact information, Talent Tech Labs, we’re running a survey on the state of talent acquisition.
Johnathan Kestenbaum (57:15):
This is a survey we run annually. So we’d love for you to contribute, both vendors and buyers of talent acquisition technology. I just want to say again, thank you. Thank you for having us. Really appreciate your time, and looking forward to doing this again soon.
Matt HIll, Dexcom (57:32):
Thank you.
Kamal Ahluwalia (57:33):
Thanks a lot, Jonathan. Thanks, Britt, David, Matt.
Britt Thomas, Micron (57:37):
Thanks, everyone.
David Francis (57:38):
Take care, all. Stay safe, healthy. Bye, bye.