How BPOs Can Harness Automation & AI to Drive Hiring Speed and Quality

Auto-generated transcript — lightly edited for readability. Names and product terms have been corrected.

Hello everyone. Welcome to our online event brought to you by Matchboard in partnership with Journeyfront. Our event is entitled How BPOs Can Harness AI and Automation to Drive Hiring Speed and Quality. The event will run for one hour and will be kicking off with a presentation on AI and Automation in Recruiting from Daniel Ash, followed by a brief demonstration of how Journeyfront handles both. Finally, we'll explore a case study from Activus Connect, a division of tech Mahindra, which is a BPO I'm sure most of you are familiar with. Now our sponsor Journeyfront helps activists and many other BPOs around the world.

Higher, smarter and faster with a market leading all-in-one recruitment platform. But more about that later. To conclude the event, we're going to have time for audience questions. So think of your questions as we go along and drop them in the chat room at any point so you don't forget them. Because very important, we're giving away a prize for the best audience question. The prize is actually a copy of a book soon to be released on Amazon, entitled very appropriately for this audience, The BPO Recruitment Playbook, Practical Tactics for High Volume Client-Oriented Recruitment. As most of you know, I'm Sharon Melamed, managing director at Matchboard.

And I'll shortly introduce our first guest, but just some quick housekeeping. We are recording this event. And secondly, if you have internet issues, there's a phone number and passcode in the invite. Finally, could I ask that you stay on mute to avoid outside noise for the speakers? However, you will have a chance to go off mute at question time. So without further ado, it's my great pleasure to introduce Daniel Ash, who is CEO at Journeyfront. Over to you, Dan. Thank you very much. We are thrilled to be here with you all. As Sharon said, I am the CEO and one of the co-founders of Journeyfront.

And today we're going to talk about harnessing automation and AI to drive hiring speed and quality, as Sharon said. So there's a couple things we're going to talk about. What is AI? What is automation? How are AI and automation similar? How are they different? How are AI and automation improving hiring speed and quality across the globe? And how Journeyfront clients accomplish this? Now, I know that you all like we have been bombarded with AI in the last few years. And this is a little bit different of a presentation. I'll just share a brief anecdote. A few years back, I was hosting on LinkedIn around the importance of hiring right.

And I remember there was a comment on my post where someone came and said, training is the only thing that matters when it comes to hiring. Hiring doesn't matter. It's just training. And I thought, wow, that was a very extreme statement. And I thought to myself, without even going to the person's profile, they're probably selling training. Because whenever, you know, often you hear such an extreme statement, it's someone selling something. Sure enough, they were. And that's when I first picked up on this idea that whenever somebody tells you only one thing matters in the world, they're often pitching you something, right?

And we are getting that all day long with AI. AI is the only thing that matters. All old solutions, dumb AI, only the good thing that works here forward, right? This is not that presentation. To be clear, Journeyfront believes in the power of AI. And we're deeply using it both in our work, like many of you, as well as within our technology. But this is going to be a more nuanced presentation on how AI is powerful, as well as the trade-offs, the pros and cons of AI in comparison to automation and in comparison to humans and the power that all three provide. So we want to start by talking about just some myths that you might hear related to AI.

The first is we often hear the term AI used interchangeably with automation. I'll hear someone say, well, so how do you automate or how do you use AI to automate, right? And there's this assumption that AI is the same thing as automation. And it's partially true in some respects, which we'll talk about.

It is very misleading, as well, which we'll talk about. They're not the same thing. The second point is that you might hear this idea that AI is the best form of automation. AI up here, more elevated, more amazing, more effective. Automation down here, less effective. That's also not entirely true. That's false in the sense that it depends on which is more effective, AI or automation, if they're truly different. And then the last, and as BPOs, you'll appreciate this nuance. You hear that AI is the cheapest form of automation. And that is actually also false, or at least it is inconclusive. And for those who have been following the news, this has become a major topic of late.

The fact of the matter is, and probably the largest amount of investment has been pushed into a technology in the last five years in history. And so what you're seeing is that this technology is being heavily subsidized by venture capital money and so on. But it's not at its sustainable rate, right? So if you're following the news at all, even recently from as early as January, all the way up to some of these articles you see here are literally today, they're finding that AI costs are increasing in many respects. And as compute power, excuse me, in the world goes up, but maybe not as much as demand for AI is going up, and that's going to change the equation of how expensive AI is.

AI, the best analogy even the founders of NVIDIA have used is that it's like a utility, and utility has an ongoing cost, and that cost perhaps more than typical utilities like electricity can change. And so AI is not necessarily synonymous with cheap. And in fact, it's very unpredictable where those costs are going to go, and that's something to keep in mind, right? You all appreciate this as BPO is that Gartner just released a report even that found that by 2030, the cost of AI will exceed the cost of offshore human agents per their study that they have done, right? So this is still up in the air in terms of how much AI will cost.

It is not necessarily the cheaper thing. We'll come back to that later. Let's talk about how automation and AI are similar. First, they're both performed by a system, not a human. They're similar in that respect, right? Of course. The second is that they're both faster than humans, right? You wouldn't want to automate something if it's going to be slower than a human. AI is typically associated with speed as well. You at least hope that they're faster than humans. So these are two things that make these two entities or technologies, automation and AI, similar. And that's where it stops. So how are AI and automation different?

Let's talk about that because that's the most important thing here. So first, automation is deterministic. It is rule-based. It executes predefined rules and sequences without deviation, right? So it takes fixed inputs and when it captures those fixed inputs, it leads to fixed outputs. And it is very predictable and consistent in that case. These inputs will lead to these outputs. So in the context of hiring, if you have input A and B and C and D, whether those are scores on a candidate, whether it's behaviors of a candidate, when a candidate completes this step, it will lead to, in this case, the example is a score of 4.5 the first time.

And the second time you do it, that automation, if the automation is around candidate scoring, it will also lead to a score of 4.5, right? So it's predictable. That's what makes it great. It's consistent and always delivers the same outcome. Now, how is that different than automation? Unlike AI, unlike automation that's deterministic, AI is probabilistic. It learns patterns from data and then makes predictions or decisions or takes actions based on those patterns. So it involves, unlike automation that's fixed inputs, fixed outputs, the power of AI is its variable inputs and its variable outputs, right?

So you might have a combination of inputs you're collecting and those inputs might vary from time to time even and it will lead to still score. But when it runs again due to this probabilistic nature, this is dynamic. It's drawing from a large data set that is ever-revolving, which is what makes AI great, but it also can make it less predictable. Now, of course, there's things you can do to try to control that predictability, you know, guardrails, context, all these different things, but it's still, by its nature, much less predictable, albeit more sophisticated in terms of its ability to handle edge cases.

You give an edge case to automation that wasn't designed for, it kind of breaks down. AI can always formulate a judgment no matter how different the inputs might be. It will formulate a judgment or take an action, right? So these are the ways in which automation and AI are different from each other. Now, understanding those differences then helps us know where to use automation and AI as it relates to hiring, right? The pro of automation is, like we talked about, predictable, fixed inputs leads to fixed outputs. It's easier to explain, to audit, to control. So wherever in the hiring process you need explainability, you need auditability, you need control.

Think of that, where do we need control, accountability? Automation is going to be better in those cases because AI, con of AI, is it's harder to explain or audit or control. You've all heard this black box nature of AI, right? There's so much that's going into it. Now, the flip side is, from a con of automation standpoint, it's less flexible. These are fixed inputs and outputs, so you've got to know what the inputs are, and you've got to know that you're going to have those exact inputs, and you've got to specify the output you want delivered, right? Whereas AI is much more flexible, you can have variable inputs and variable outputs.

So wherever your structure breaks down that causes automation to reach its limit, AI can be very powerful. So what this basically means is that for automation, simple fixed tasks, you don't need AI for those things. Maybe it's scoring a multiple choice question, maybe it's sending a candidate reminder. Not only do you not need AI, it's not the better technology to do that thing. Now, I don't like to think that automation just does simple things. It can do very advanced complex things, assessing skills, using assessments, scheduling of interviews, coordinating calendars, and the list can go on.

But whether it's simple or complex, they tend to be more fixed type activities. Versus AI, they're more flexible type activities. So when it comes to, I'm going to create a job description, that's going to vary all the time, what job description I want the AI to help me create. Of course, that's an example of a great use case for AI, writing interview questions, you don't know what question you want, what trade, what use case, et cetera. It's evolving dynamic, great use for AI, even though those are simple tasks. Complex, reviewing and scoring a resume, that variable inputs, variable outputs, you don't know what's going to be on the resume.

So you might be able to use automation to try to grab keywords, but that's not effective because how good are those inputs really, the presence of a keyword, scoring a resume. Whereas AI is much better at that variability you might find in a resume and producing a consistent output, which is a score on that resume. Same thing with interviews. Now AI can mimic a human more than automation can in terms of an interview where can it's going to be answering various questions, you don't know what's coming. So understanding the differences of automation and AI help us learn what they're best for and what they're not best for.

Now what we wanted to show is just an example. If you're looking at across the hiring process where you use all these things, not just automation and AI, but also just where do you deploy humans? It's based on the strengths and weaknesses of each type, right? Our view at Journeyfront is because of this lack of control, because of this variable cost associated with AI, wherever you can automate, automate. If you can get away with just automating something, that's forever going to be low cost if not free completely. Definitely automate wherever you can, but there are some things that AI is best at and that's where we want to focus your energy, right?

This is like a machine. If you were able to use a spring load feature for like a hinge, right? That would be the better approach if you're building some kind of machine than to have electricity running through the whole machine doing everything, right? Same thing, automation can be applied great in certain use cases, but where you need that electricity, where you need that AI, that intelligence, that's where you apply AI. And then of course humans, where they're best, where they're needed, right? That's not just where they're best. Humans are better at some things than automation and even AI, but it's also where they're needed, right?

You might want more accountability or more control over certain things in hiring and that's where you want humans, right? So if you look at the entire hiring process, and this is not intended to be comprehensive overview of like everything that happens in hiring and every type of thing you could automate or you could deploy AI or you could deploy humans, right? This is just to give you some examples of the types of things that can be automated or that can best leverage AI or where humans can uniquely play a role, right? I want to just call out a couple of things. So automation, AI, scoring structured inputs, assessments, multiple choice questions, great for AI, aggregating those scores, ranking and routing based on scores.

You want control over that. Your clients want to audit you and know how you're making those decisions a screening candidate. So it's a great use case for automation. But when it comes to unstructured inputs, things that were very hard to automate before or if you did it was at week at best, resumes, variable input questions, video interview questions, AI is great. Of course humans still need to be in control of making those decisions around what screening plans we build, where we deploy AI, what guidance we give to AI, right? And reviewing the funnel, intervening where needed as automation and or AI are helping us screen candidates, right?

Interviewing, I'll just call out a few things. AI can help, excuse me, automation can help schedule of interviews, but it's a variable input process. So of course humans have done interviews for the ages, but AI is now able to also run that process of performing an interview now, taking those variable outputs or inputs and producing an output similar to how a human might do, making a judgment call on a candidate. Deciding who to hire, that ultimate decision. If you're going to put a human in a loop somewhere, this is a great place to put it. It's at the end of the funnel. After you've let the system do the heavy work, you can act as an accountability checkpoint on candidates.

Certainly you're tying it to targets you have for different hiring classes, et cetera. You don't want to just have the AI or automation send out a thousand offer letters while you're sleeping. That would be a nightmare. But automation can supplement the hiring decision, and Journeyfront does this as you'll see by aggregating scores and giving a human everything they need at their fingertips to make a quick, effective decision. And of course AI can really effectively supplement the hiring decision as well by helping you drive predictions. Tracking and optimizing, when you know what you're looking for, analyzing hiring data is great for automation, but when you don't know what you're looking for and you're trying to discover trends or predictive profiles you didn't know exist, AI is your go-to.

And of course, at the end of the day, the humans can leverage both of those types of technologies to make decisions on how we change the hiring process. So that's just an example of understanding the strengths of automation versus AI versus humans, allows us to know how to build the ideal hiring process by leveraging what one might use versus another versus another. So what we want to do is, I want to turn it over to my colleague, Cody, who's going to show you examples of this at work, right? So he's going to open up the live journey front product and show you examples of how we help clients use automation or AI combined with humans steps and you can configure it as you need to drive a more effective hiring process.

Awesome. Thanks, Daniel. Yeah, so I get the fun part of actually showing you an example in real life how these things work together. Although we're going to be in journey friends platform, my hope is that you'll take away some concepts that you guys can use as well, however, and wherever you're hiring. So what I'm going to show you first is the setup. When to deploy AI, when's a good use of automation, how does it work and how should they interplay with one another? Often when I work with customers, I have a catchphrase, which is AI guided, human decided. And so we're going to show how we can use AI to kind of guide us to make the best decisions, but ultimately humans decide what the AI thinks about, what the automations are, and then ultimately that the hiring decision.

So first, let's talk about deployment of AI and AI tools within a screening plan. So what you're looking at here is a screening plan within journey front. You have the different steps, automation rules, the different components from offer onboarding, manager interview, etc. AI is deployed in a few places in the screening plan. Some that I'll be able to show on the screen and others I'll show here in a minute. The first is using AI to score video interviews, you know, phone screens, right, some of those basic intake interviews. To give you kind of an idea, you know, a really important part of designing and building AI is that you can define the rules of what a good score is and what a bad score is for the AI, as well as provide additional instruction as well.

So that there is still kind of although black box as Daniel mentioned that there will still be variability in how the AI responds and put scores out, you at least have the ability to define what a one should be a five should be to kind of train over time. Another deployment of AI in the screening plan is the use of AI in language assessments. So using an AI language vendor to quickly assess people see for skills are they a B1 B2 C2, as well as providing evidence. Some things that you won't see right off the bat on this screening plan is also kind of two behind the scenes AI tools going on. One is fraud detection and identifying people who are trying to scam you and then Daniel will touch on this more but the use of predictive profiles understanding hey what attributes lead to actual success in the role and applying that across your applicants.

So those are the deployment of AI's in the screen plan and we'll talk about those here in more detail as well. The next thing is automations. Now automations either go really bad for people or really poorly. What I found is that automations that are based on rules and decisions that are backed by data is the best path forward. Let me just give you two examples of that really quickly. First is we have this basic automation that I want to I want to after the pre-screen after they answer the qualification questions, you know, they've passed the basic knockout, you know, the, you know, the years of experience qualifications, resume, etc.

Typing test, that video interview, I want to automatically advance them to the next step if they made disqualifications. So here are two basic rules. Candidate completes the activity, internet speed test, their overall score is greater than three and their predictive score is above of four. I want to auto advance them to the assessments. We also have this other one where it's, you know, I've got multiple clients I'm trying to fill. In this case, I'm using a healthcare example, but I want to take from my general applicant pool and quickly qualify people that are healthcare qualified. So I have a basic qualification that they have healthcare experience. I want to automatically route them to a healthcare assessment and healthcare simulation.

And then there's some other, you know, like I like to call kind of quality of life and applicant experience automations that are really helpful and really powerful and, you know, we'll hear from a little later about how awesome some of these communication automations are, but simple things like a text, if they're incomplete, kind of, hey, we want to see if you're still interested, would you mind completing your application? All the way to quality of life for your recruiters where, hey, if I don't hear a response for them in like 10 days, I'm just going to automatically reject them so my funnel can stay clean and keep moving.

So once I've come in and I've applied AI across the screening plan and automations, this allows now the human side to come in, which we'll talk about here now that we go to kind of the deployment of it. So in theory, when you have the automation set up the AI tools in place, what this leads your recruiting team to do is to only look at people that are qualified and already disposition and organized into categories and clients that they are best fit for. So here's an example, right, where I can quickly sort to anyone that's active after they've either passed or failed, and then quickly sorting, again, an overall score that is derived based on the qualifications that you care about.

And quickly assign people these candidates, quickly move them, and then all the way down to really simple things like getting to the recruiter interview and automatically sending an invite to schedule on their calendars to understand if they've been scheduled, completed, and then again, automate it all the way through. This might seem like too good to be true, but there are so many examples of our clients that are making same day hires, even people that are making hires like three hours after people apply based on these automations. Just to give you a sense, this one applicant started their application at 1111, and 30 minutes later they were automatically advanced to a hiring manager interview based on the qualifications, the AI scores, and the automatic advancing.

So really, really powerful tools when deployed correctly and based on science and data, which is something Daniel will talk about here. Once you have all this automation, now there is this human decide component. And to me, this is one of my favorite parts of Journeyfront is where all of these different activities combine to make a really easy decision. So to kind of break that down in a little more detail, what I have is our predicted scores. So in this case, I have a predicted score of how likely is it they're going to make it 135 days tenure, and then I have a specific client prediction. What's the likelihood that Jaden here makes it 120 days under a centene account, a healthcare account?

And we have these predicted models of, okay, these are the drivers, the attributes that lead to success, and we can see the ideal versus where the candidate fell compared to their peers. Again, I'll touch on this briefly and I'll have Daniel talk about this as well. But being able to understand what leads to success is a really, really key here. On top of that, all this is aggregated into one place. So we have all of our assessments, the pre-screen questions, my interviewer. I even have the ability to, you know, come in and, you know, play the human part of, you know, I still need my recruiters to classify and ask their questions.

So I can quickly come in, review an interview guide and submit, ask the questions, put in the scores that, again, ultimately aggregate into this applicant scorecard. The kind of final thing that I want to show here is the concept of having all these things in one place. So I can see my pre-screen questions that are automatically scored. Again, if they didn't meet the qualifications, they'd just be knocked out. I wouldn't even look at them. But I also have the ability to come and review, you know, AI scores, right, where I can come in. I have a couple of audio questions I have people ask, and the AI comes in, gives a response, and in this case, it's applicant, you know, just gave a terrible answer.

But a really important thing about AI is the ability to kind of understand the reasoning behind it, but still have the ability for me as a recruiter to come in, listen to the application myself, and override a score if I feel like, you know, hey, maybe AI got it wrong, right? Now, it's the interplay of all these tools that really makes a powerful experience. And each one assists a recruiter to quickly optimize and qualify people in an efficient manner. So I'll leave it there. These are just a few examples of how automation and AI come in to play on a square card, and then how the human still needs to come in to make the ultimate decision of hire, not hire, send to onboarding, not onboarding, and then, you know, even review or double down during the interview on some of the skills that you care about.

So with that, I'll turn over to Daniel to kind of wrap things up and talk a little bit more about AI here. Great, thank you, Cody. So obviously, that was just a taste or some examples of using automation and AI and humans combined, building into the ideal workflow that you want to move candidates quickly and effectively through your process and make more accurate hiring decisions. But to be clear, if it wasn't clear already, we believe that the future of hiring involves humans plus automation plus AI working together. That's how you unleash the power of your teams, how you do more with less. That is the key, all three of these tools working together.

As you need to, customize region by region to improve hiring speed and quality. We take this seriously at Journeyfront. As we said at the beginning, we're an intelligent hiring platform purpose built for BPOs. If there's anyone that needs to both balance speed and quality, it's you all. It's with the volume you're doing, with the deadlines that you're hiring on. And Journeyfront helps you streamline that process, whether it's from sourcing to screening, assessments built into the software, communicating, interviewing, offers and onboarding and so on. The big thing is, no matter whether it's humans or automation or AI, it's only as good as its data as the data you have, right?

So we were very deliberate in building this system. So we're capturing critical data along the way, every step of that journey. So you can do really powerful things with it from the intelligence layer, matching candidates to not just an overall role, but various roles, even multiple clients, routing candidates from one path down to other paths. That's what the power of the intelligence layer, when it's built on a fundamental sound data layer, can have. And that is what fuels the automation layer. Cody shared some examples, but you can automate just about anything you need to automate tied to candidate behaviors and scores.

And that's important. Now, Cody mentioned I just shared, before I turn it back over to Sharon, one last piece. Cody showed some predictive scores. At the end of the day, AI, the best thing AI is good at is taking a large, growing, complex data set and learning from it over time and predicting a candidate's chance of success based on all the historical patterns found. And that's what we built at Journeyfront. There's a feedback loop built into the system that as you're collecting all this data from interviews, assessments, resumes, screening questions and using it to make hiring decisions. We have the ability to pull in turnover data, performance data from the data, the systems that you house that and our system and we use AI to analyze that data and identify patterns.

And that's what drives the predictions that Cody mentioned. Automating up front is great, deploying automation with humans and AI and some of that scoring of unstructured data. But where it gets very powerful is when you can, based on all the historical hiring data, not just tied to different roll ties, but even different clients and those performance KPIs or turnover outcomes, you can start to use AI to predict those outcomes of chances of success. Immediately as candidates start to come in and then AI can update its predictions as we gather more data on that candidate. That is truly one of the most unique powers of AI that AI does best and that's how we're using it and that's how you can use it too to unleash power in your own hiring process.

I wanted to leave it at that. I'm going to turn it back over to Sharon and we're blessed to have a friend and fellow talent acquisition leader in the BPO world and Journeyfront client to talk about her experience leveraging some of these principles and this kind of a system. Go ahead Sharon. Yes, and thanks Cody for the great demo. I'm sure everyone's minds are ticking away. Feel free to drop your questions in. As I mentioned earlier, we will have an audience best question prize after the fireside chat, which we're now about to go into. And I know you've probably been looking forward to hearing from a fellow BPO practitioner and recruiter to speak to how everything Dan and Cody have presented actually works.

So I'd like to turn over to Dawn Nash, Director of Recruiting and Inculturation at Activus Connect, which is a Journeyfront client. Dawn will be interviewed very informally though, I should say by Dave Biesinger, who manages partnerships at Journeyfront. So are the two both? Well, thanks so much. Just to introduce Dawn a little bit. So Dawn is the, she's Activus Connect's Director of Recruitment and Inculturation. And in this role she brings people into the team and she shows them what it means to be part of the Activus Connect family. That includes the core values of passion, integrity, respect, authenticity, and fun.

And she has over 22 years of experience in the industry and she's filled every role within this industry. She's inspired by people, she's inspired by their stories, experiences, and above all their passion. And she loves recruiting the nation's brightest minds. So Dawn, thank you so much for joining us. Thanks for having me, Dave. I will say I have not had the role of CEO yet. So if anybody's looking, let me know. That's next. That's coming up next. So Dawn, you've been in this industry for a long time. I'm sure you've seen a lot of changes over the years. How do you think about the pros and cons of both automation and AI?

So in the beginning I think there was a lot of fear. I think we've discussed, Daniel spoke a lot about that, a lot of fear. We're all humans and we all have our counterparts, our teams, and we want to make sure that everybody's busy and that they continue to stay employed, etc. It wasn't until we really started to break it down and as the ages grew that we realized that not only could we keep our teams intact, we could give our teams a better experience along with those that we are bringing into our teams to have new team members. Awesome. That's great. And as you're automating more of the process, I know teams and people are really important to you.

How are you making sure that you preserve that human touch? And has that been a challenge? So I thought it would be, but it's actually not. So when an applicant starts their process, we are all in the same line, really. And so we are all battling for those same applicants. We all know how it goes. They start in one place, they have to do this function, this function, this function, this function. Well, we used to have to touch this function, that function, that function, that function. And then we were all touching those functions to get to an interview to really get to so we can talk to the qualified individual who we believed was a qualified individual.

And my theory is I don't want my team to talk to that individual unless they go to that individual with the belief that they're going to offer them a position. Okay. So, but all of these things happen first. So how can we, and this is what we did, is we created with Journeyfront a way to where we have automation and AI giving us those persons that we want to have those conversations with because we believe that they're qualified based off of how we've asked it to behave and how it's learned to behave through the different things that we have told it, worked with it. And this changes client by client. So that is what that's how we keep an end.

Those interviews are done by humans. So that is when we do the human touch is the most exciting part is the is the interview and and hopefully the job offer. Yeah, um, you know, I actually got a question in the chat that I want to serve to you because I think you could speak to this really well. So at what stage do you see the most dropout? And as as would have been expected with a with a human and not through automation and AI, like, can you give us a sense for dropout before you automated more of the process and drop out now? What's funny is, is I can't give you 100% of the answer that question, because Journeyfront has so much back end reporting, and no one paid me to say that. That is just an organic question that has come up.

Journeyfront has has some robust reporting on the back end that shows us not only where what stage that they're dropping in. So we saw Cody show us like this stage in that stage, you know, one, two, three, four, five, however, but they break down the stages and you can see where people are leaving. So before we had our guests, we had a guest where they were leaving and we guessed they were leaving when we asked them to go, maybe on video, maybe a couple questions and maybe when they realized that it wasn't. Maybe if there was testing, for example, what we learned through Journeyfront was they actually are leaving or were leaving when we asked them a video question. And what we believe drove that was fear of AI.

So what we did was, is we went a little bit further in our, in our question set, we gave them a little bit more information about ourselves, then we asked them to do that video question. Now we still have fallout there, and that's still our biggest spot for fallout. We don't have two thirds fallout. And that was a win. So I love that answer. You know, BPO recruiting is like famously brutal. You deal with like high volumes, you have aggressive client ramps, and historically high attrition. So what was hiring an activist actually like before you built the process? I think you've spoken to this a little bit. It sounds like there was kind of a lack of visibility into things like fall off. And then, you know, sorry for the double barrel question, but what's it like afterward?

Okay, so great question. That's one of my favorite things to talk about where Journeyfront is concerned when I'm asked by people within the business and leadership. Before we had Journeyfront, when you came in, you came in through one area, and you might have filled something out. And then we communicated back to you and we needed to you to do this other thing. And then you communicated back. And then we needed you to do this other thing. And then you communicated back. And this is also we could get the same set of information that we still require. None of our requirements have changed. With automation and with some AI, we've been able to push the successful people forward.

And of course, send our regrets to those that don't make it quite that far. You're 16. I'm sorry. We hire everybody 18 and above. That's an easy one to go off of. And where we really struggled was getting people to that interview, right? Because they had to go through all these, all these lines. Now they can do all these lines almost immediately, one after the other. And then our recruiters notified that there is someone here, they're complete. We see that they have a set or a subset of things that we believe that can make them successful. And by subset, I mean the Journeyfront and its calculated system might let us know that we might want to look at one specific area.

So it lets us know that these persons have the reasonable degree to be successful and that they have either need review or they've scheduled an interview with us. So instead of going here and then we're talking to them and then going here and then we're talking to them and they're going there, they're coming on a straight line and then they're talking to us. And that we find that builds a faster relationship. Clients are happier because they are able to. We're actually not quite to the point where we can tell our clients yet how much time this is reduced because we don't want to overshoot ourselves.

But the data is looking pretty good. We're taking quite a bit of time off of our recruitment time. And we are not leaving applicants hanging. So if you go into the wild and you read reviews, there are so many, even from old activists, I can admit that say we haven't heard back anything. It takes a while to hear things. You won't hear that from us now. We're hearing because because we're able to get to these the same day. And what's really great about automation is that if they enter the system and they've stopped at, I don't know, section four, question three, whatever that might be. We have a little automation that says, excuse me, we'd really like to talk to you and you stopped here. Here's a link.

If you just click on this, we'd really like if you'd complete. So that's one of the ways that we're one of the many ways that we're using automation to push through. But it's one of those ways that pushes people through without us going through and and doing it ourselves. You know, kind of ironic that using a system actually made your process feel more human and enhanced your ability to create relationships. You know, you spoke earlier about not having a lot of visibility into drop off as an example. And I think that there are lots of parts of the hiring process that can sometimes be invisible to us.

And I think for you, fraud was one of those that it was like, you don't know, it's there until you look and then you see it. So talk to us about fraud detection and how automation has helped you detect fraud and act against it. So my other favorite topic is the deep fakes that we're all seeing. And I see a couple of head nods and we are all seeing either someone a complete AI deep fake, we are seeing people from other countries applying countries that so we hire as a part of tech Mahindra, you know, we hire all over the globe, but we still have people high apply for jobs and try to and have gotten through in the past from somewhere else.

And last year, when we had our largest client, and we did some research, our largest client, we hired and went through all of their checking spoke to all of them to classes worth to entire classes worth of fake applicants turned trainees. And so that's that's waste. This is just a waste. It's a it's a waste of funds and all the way around for a for a department that's already overhead, right? There is no money to waste in recruitment. And we all know that. So one of the things that the journey fronts has helped us do, and that we've learned pretty quickly, and we, I like to think that we helped, you know, really give some feedback where this was concerned was let people know one of the things that we needed to see was an IP address, for example, and journey fronts helps us see that IP address. And I'll give you one brief example we watched a gentleman now this was a job we hire in 26 states.

And so this particular job, and we watched this gentleman apply in Georgia and move to Mexico, or we don't hire, and where we can't employ from we just don't, we don't have the credentials there for this particular job. And we were able to track that and journey front was actually able to track that for us and flag it for us. So that and that's one example where we had 60 that got by us last year. We had we caught one in the process and that that's one of many actually that we've caught the process but we're able to catch them much much sooner. Well, thanks, Don. I think I think we're at about time for this session. I feel like I could talk to you all evening, but I know it's late for you. Thank you so much for joining us. It's quite late for Don right now.

So really appreciate it. Thank you, Dave and Dawn. We actually have quite a few questions that come through. As you've been chatting so I might read out just a couple and let's see who wants to jump in and answer so Anthony has asked is journey front set up to handle automate downstream tasks such as offer letters acceptance to draw signatures filing dog management out of the box. Yes. It sounds like a chorus. And I'm sorry, Daniel, I took that from you because we don't just get them to the interview. Well, I should really say that we integrated as far as our HRIS system. So we get them all the way through to the HRIS system.

Yeah, Daniel could have told you that, but I'm pretty proud that we got that. Put better than I would have. There's a great question there. And also just before we move on to someone else, Anthony asked, is there underlying AI being relied on, like Chatchity, Claude and Grog? Maybe Dan could answer that one. Yeah, we are taking advantage of that treasure trove of money that is pouring into some of these large language models as is everyone. So yes, we actually have multiple believe it or not some are better at content generation for screening plan content. Some are better at video at reading, like video responses and, you know, transcribing scoring. So we're using, you know, in certain places of the application and open AI.

In other places, Chatchity, for those that don't know the tie in, we're using Gemini and Claude and others, and we're experimenting with some unique stuff on our side as well. So there's a smorgasbord of kind of sophistication how we're setting it up, but the short answer is yes, we're leveraging some of those technologies. Fantastic. Does that answer your question, Anthony? Or do you have any follow up from that? No, just that I think, I mean, I'm sort of very big learning space on a lot of this, but with regards to some of those, the Grox and the Claude and the Chatchity BTs, they do have some leanings potentially, I don't know, biases and politically and so on, depending on what they draw on that.

You know, I don't want to get too deep, but I'm just curious how that kind of plays out in a field like HR, but anyway, probably more comment than a question. Yeah, we could talk about that for an hour, obviously, Anthony, but we're absolutely right. We talked about, you know, it's harder to control, to audit, you know, the risk of AI is biased. So we're using it very deliberately right now, and we're giving it a lot of guardrails, a lot of context, a lot of calibration over time to control for the kinds of things you're talking about, absolutely. I'm particularly curious about Elaine's question because she beat me to it. It's probably the part that I'm finding the most interesting. So we'll probably pop onto Elaine's question there, and I'm very curious to hear the answer. Thank you.

Actually, Elaine's got a few questions. Thank you for your enthusiasm. So not sure which one Anthony referred to, but one of them was other rejections tailored by the AI and how accurate are they, especially if there are multiple reasons why they're not progressing? Yeah, interestingly, this is where we talk about, you know, deploying AI where it makes sense and automation where it makes sense where you want, you know, Anthony just shared bias, right, that AI can have if you're not careful. Automation, you have a lot more control over how you deploy those things, and it's a lot more transparent. What Canada doesn't want transparency, fairness, consistency of how something is applied, right, whereas, you know, it's a little bit easier to do that automation, it's a lot easier than AI.

So from a rejection standpoint, typically we're relying on AI. Now to be clear, you can reject candidates for all sorts of combinations, the humans in control of what that is, Don just shared, you know, under the age of 18 as an example. It can be if you don't have a see for score below a certain see for score on the AI language test that we have a combination of scores, but you do have control over how you deploy those rejections. Often the AI scoring gives scores and you can rank by those people and you can advance, but we're a little more careful with how we deploy the rejections, I would say, and therefore it's more accurate to and more fair.

Yeah, and I do want to be clear. So AI does not reject anyone are in our platform. Yeah. Okay. Danielle has asked, we have restrictions where the data is held. Are you able to hold data in different locations based on regions? Yes, I've had many conversations with our engineering leadership around this. Yes, we understand that certain time sometimes there's client requirements, etc. But yes. Great. Before I move on to the next question, does anyone want to go off mute and ask a question more easily in person? I'm happy to go off on just from the few questions that I've asked is to get the one that probably I'm guessing was the one that Anthony was referring to Anthony correct me if I'm wrong.

But basically I was just looking at that turnover prediction score. I'm really keen to understand what elements play into that just because it's such a far fetched thing to know, you know, how does the training impact, how does the development impact, the manager impact, everything else around the culture piece, not only the actual fish. So it was keen to find out a bit more about that. Yeah. Yeah, so I'm happy to answer this. We've built the data layer of Journeyfront. If you remember that visual in a very deliberate way for these kinds of analyses. And so as an example, and you brought up a good point, other things can go into turnover, right, training, culture, etc.

The manager. But if we're identifying insights from the same group, who got the same manager, the same training, you know, this in the same job, same pay, etc. It isolates that pattern. And for all those same, you saw different outcomes associated with those people right some stayed two days, some stayed three months, some stayed three years right. We isolate the predictions around identifying patterns in those similar groups that's very important. And then to answer your other question what what inputs are used or what tools are used to predict turnover to be clear the systems bill to use any input that you've collected end to end from the candidate from the moment they answer an application question or upload their resume, all the way to that final interview.

So it can identify and this is where AI is really good any combination or permutation of combinations of inputs that predict longevity within the role. With that said, the most. I could give an example, if Elaine would like one. Sure. So what we do, and we're working on implementing right now. So I understand that our relationship with journey front continues to grow as you don't want to build the plane in the air, right. You know, so we are to the space now to where we're taking existing information as they're graduating or not graduating, or as they're successful or not successful. And we're to this place where we're feet we feed it back in. So it's a full circle where that's how they get that predictive score.

And so we're looking forward to that to seeing more of that. But that is what we are. It's a and we have everybody on board willing they see what's happening on the front end. And so they are very excited to see because they're collecting all of this data. Why are you collecting all of this data to get this number. And so everybody's pretty excited to see it. So sorry. But it's great. It's better to hear it from you. Someone in the trenches do doing it in a practitioner sense. But yeah, automation identifies every single input collected from a candidate and whether it correlates with retention as well as performance.

That's automation, but identifying that unique combination. That's that's where we're leveraging AI to identify if there's a combination or permutation of inputs and scores that predict longevity. And while it can draw from any input, the most common inputs that are that are weighted heavily in that combination is some of the more advanced assessments we have like behavioral predictive behavioral assessments where it measures a candidate's interest, values, personality, competencies combined, our job simulation tool that that simulates for a candidate, what the job is like and test if they can do that job.

That's obviously a good barometer of whether they can do the job is testing if they can do it. And then often we'll find that like there's a specific question or two or trait that's scored in that final hiring manager interview. It's very, very predictive of longevity within the role and that candidates ability to work out in that role and the power of the AI and the algorithm is that it can be it can be a combination of those things I just mentioned altogether, which is more predictive than any one of the inputs alone. Excellent. And question from Stuart. Do your clients maintain a separate candidate database or do your clients use Journeyfront to house the records? If so, is there other specific systems the program works best with?

I'm not sure I understand this question. Cody, chime in or Dave, if you do. Yeah, I think I, yeah, go ahead, Cody. Yeah, are you, are you, are you asking for the, sorry, Dave, are you asking for the extradition of data from Journeyfront into like a data warehouse like, I don't know, Amazon Web Service or Power BI or something? Yeah. I'm not sure if you guys can hear me, but yeah, so my question there was just more so, is this, does Journeyfront become the main capture point for, for all recruitment once you're in place? In other words, right from day one or is it effectively coming into another system and then being shared with Journeyfront?

That was the piece. Yeah, we can be the primary system of record that manages it all end to end. That's right. Now, there are cases where we could be a secondary that integrates with your system too, but mainly usually we're the primary. And the primary then right the way through the candidate lifecycle to capture every data point right the way through. Okay, great. Yes. Yep. I think we have time for just one, possibly two more questions. So I'll move on to the next one. Does your platform have the capability to enrich candidate profiles by conducting external research beyond static data, such as publicly available professional information?

And if so, how is this done and what controls are in place to ensure compliance with privacy and data regulations? Yeah, so first of all, we take compliance seriously. We're SOC2 type two, we're GDPR compliant in Europe, we're Poppy compliant in South Africa or wherever you might be. That's important. Our clients hire across all sorts of countries across the globe. We do not go and enrich data from LinkedIn or some other place. We can integrate with some of these things like an Indeed or whatever and pull data in from those places, but we don't enrich it with some of those third party tools. We do not do that right now.

And then, let's see, what was the other question? Did I capture the questions? Was that the main thing? I think so. Vinit, did you have a follow up to that one? No, that's it, basically. I just wanted to understand. Obviously, you can find a lot of behavioral sort of aspects on the web as well. So that was my question. Yeah, sure. Yeah, it's a good question. We haven't focused there thus far, but it's a good question. That's something that a human does in the process. We'll go and check someone's, I don't know if we do, but I'm just saying you might go and check someone's Facebook page or you might go and check.

I know companies do, so I suppose that's part of the process. But just so, and I'm sorry to go over time, but sort of off the back of Elaine's question there with regards to and your description of the data set that you're interrogating vis-a-vis to get that prediction, which was, it sounded like it's an internal data set. It's looking at the previous, you know, these are the people you've worked with over the last, the thousand people you've worked with over the last year, and this will take towards a prediction. Are you just using an internal data set or are you also going outside and looking at, you know, trends in the world or other studies that have been done to say, you know, these people suit that role?

Or is it only looking at your own private data set? Yeah. We've leveraged a lot of the research out there on other various studies. Our Chief Science Officer is one of our co-founders and there's a lot of good research that's been put in place on hiring effectiveness and turnover drivers and all of that, and that has led to the creation of the tools we've had. In some cases, the underlying algorithms that we first started off with. But in terms of the robust things that train on your own data, we typically help a company use their own data. As being the most accurate, the most accurate data and most precise data to utilize from that standpoint.

So typically it's a company's own data over time. Yeah. All right. There you go. Yes. No, I was just going to say we've often found that those large data sets are far less predictive than your internal data sets that you're creating. So these aggregate models that are out there when applied in Journeyfront do not perform as well as your internal aggregate. Yeah. That's a great question. And in fact, it's time for Dave to announce the winner of the best question prize. Dave, you're on mute. Yes. I'm actually going to cheat. And we are going to award everyone who asked a question with the prize.

Wow. Seriously, would you get there? So Elaine, Anthony, Danielle, Stuart and Vinit. Sorry if I said that name wrong. You will all be receiving a copy of this book. It will be coming out at the end of June and we will ship it to you. You'll get a copy of the book. Wonderful. Thank you so much, Dave. Well, that brings our event to a close and I hope everyone enjoyed it and it was a good use of your time. Thanks so much to our speakers, our sponsor Journeyfront, of course. We will be sending around a short survey to get your thoughts. So take 60 seconds literally to answer that. And Dave will send you a recording because I noticed a lot of people wanted to do their own recordings, but we'll have an official recording that you can share with your colleagues.

That's it. Wishing everyone a wonderful day ahead. Stay safe. Thank you, everyone.