Speaker 2: 0:06

Welcome to Real Talk on Talent, a human resources podcast where we talk about talent acquisition, recruiting and all things hiring. Hey Dina, hi Hilary, welcome back. Thanks, pleasure to be here. So this is actually a really unique podcast. It is For all his reasons this is our first time having a guest and we're super excited about it.

Speaker 3: 0:34

Of course, we got to start out with introductions of stuff at Human People Solutions and I am also our executive sponsor for our AI implementation and strategy projects and programs.

Speaker 2: 0:50

Woo, see fancy, but can you tell us what that actually means as it relates to AI, because that's why you're here. We want to talk about AI and, like all of those elements, dina and I have attempted this discussion before we got here and be very generous. Yeah, yeah, so great title, thank you, but like, what does that mean?

Speaker 3: 1:09

it means that, as human, has uh gotten further and embarked on our ai solutions, and how do we incorporate it into our processes and and how we work both internally and client facing. How do we do that? What are we putting in place? What tools are there going to be? How do we use it? What do we think that the impact is going to be from that? And then, how do we set up all the structures around it so it's used in an appropriate way?

Speaker 4: 1:39

So I, you know, human is unique in that we are recruiting 40,000 individuals every year on behalf of our partners, Almost every industry, almost every job function you could imagine with various volumes. How did you approach where to even start with this?

Speaker 2: 1:56

It just seems like well, not only that like, not only the amount of recruitment and the type of recruitment, but like AI. Is this kind of like ubiquitous topic right? That's like, what is AI? We've talked about this chat, gpt.

Speaker 3: 2:09

it's taking over the world, like, and so walk us through that first conversation when they're like, hey, and we know we need to do this, like we uh it's funny that you say that, because that is something we hear a lot in conversations that we've had with clients and partners around like we want to do AI. And people have said that we have AI. What does that mean Exactly? And we'd like to do it, and what does that mean? And we have tools and it's underpinned, so, like, how are we actually using it?

Speaker 3: 2:41

I think about AI as two different, at least in our context, as two different types of tools. Okay, there's either an AI tool for automation Okay, so ways in which we can make what we do faster, more efficient, better, yep. And then there are AI processing data and insight tools Okay, processing data and insight tools. Okay. So, around, you know, how can we take the data that we have and learn from it, see patterns, get suggestions in a way that we don't today, because we aren't naturally connecting all of those data dots. Yeah, okay, the one that we started with was around automation.

Speaker 2: 3:20

Okay, so was there any reason that you picked automation first?

Speaker 3: 3:23

We were really focused on as our and this comes to like what's your vision, what's your goal? What do you? What's the problem that you're trying to solve?

Speaker 2: 3:30

with AI.

Speaker 3: 3:31

AI. It's fun to have AI and say that you're AI enabled, but, like anything else, it's a tool, and so what's the issue that you're trying to solve?

Speaker 2: 3:39

Yeah, I think that's a really interesting call out, because so often and I say this all the time like a lot of people want to go tactics before strategy. So it's and I've made this criticism of companies where there's this race to AI integration, where it's like, oh, just plug AI in there and sometimes it's the right fit, but sometimes it's also just like so you can say you're AI enabled. That's my interpretation of it, and so I love that idea of saying, okay, we could have gone 17 different ways, but the goal was really thinking about efficiency and how a tool could tie into that process element. Yes, data could be, data analysis could be that piece, but for us today, that was not the business case that we wanted to prioritize.

Speaker 3: 4:17

And, to be clear, data analysis can help with efficiency, and I could argue all day. We could have gone that direction, but what's the lowest hanging fruit? The other thing is like what is what is what we can do today and where we, you know where we want to go, or what's the first thing that we should put in place. And so it is so critical to have a thought of what it is you're trying to solve, because not only does that change the use case that you're thinking about using it for, but it informs then what types of capabilities you are looking for when you go out to market to see like who's my vendor partner and like what is it that we need to put in place? Because there are nuances between how different ones work.

Speaker 3: 5:00

And so that's where we started. We took a look. We were saying, okay, the focus for us is we want to get more efficient. Ok, and so we're going to focus on the automation side of the house. The second piece we did was we said, all right, ai. One of the things that is necessary for automation to be useful from an AI perspective is it needs to be a highly repeatable task with little variation between times, because AI the nice thing about AI and it's different from, like robotics, process automation, or also known as RPA, or just historical like if you think of a phone tree or something there isn't a whole lot of room for variation. Ai does have some room for variation, but you still need a repeatable process.

Speaker 2: 5:43

But, like, when you think about that repeatable piece and the efficiency, the way you described that made me think of the compliance side of the house, because when we think about that's, I think the question we probably get the most, like with our healthcare partners is okay, there are all of these compliance questions concerns huge potential liability there. So when? And I think one of the benefits of the repeatability is not just scalability but control. Yeah, I would assume. Yeah. So it's like, if you know, this is where we want ai to live and it's repeatable and it's scalable, but it's also something that we can navigate the complexity of like decision making.

Speaker 3: 6:22

Exactly to that point. One of the most important things with AI is deciding what kind of well truly with any tool, but also especially true with AI, is what kind of guardrails do you want in place? And you have to. In order to know what guardrails you want in place, how much risk you're willing to assume and how much variation you're willing for the tool to make on your behalf, you need to understand what you want it to do and what that exactly. To that point, like what that is, yeah, Okay. So with the repeatable process, we first looked at what is a repeatable process. We first looked at what is a repeatable process, and the second thing that you want to look at is how much time does that repeatable process take within a day in the life?

Speaker 2: 7:11

yep, the more time we say day in the life, clarify like a day in life of the recruiter, of the recruiter or whomever, though, you're thinking about anything.

Speaker 3: 7:19

How much time does this? Oh see, I'm straight a human's use case right now.

Speaker 2: 7:22

Well, that is more like that's nice. You're thinking about AI, but I want to think about here.

Speaker 3: 7:26

Let's go. We did end up looking at exactly to humans away on a day in the life of the recruiter. Yeah, yeah, but for any automation use case does this take up a lot of time? Because human time, specifically Because if it does, the more time something takes, the more benefit you will get when it's automated, when a human gets their time back, and so we were looking at where can we find a repeatable process that takes up a lot of time.

Speaker 3: 7:57

So then we went to look at humans' day in the life and how were recruiters spending their time, and what we found was screening was, particularly for specific types of roles, just an enormous portion, like between 30 and 50 percent of their time on an average week, and by screening you mean picking up the phone, calling up a candidate, asking them very black and white questions about Everything, from looking at a resume and saying should I pick up the phone and call this candidate, to picking up the phone and calling the candidate, and that distinction is important for us, particularly because when we were looking at what tool and what use case did we want, there's one tool and type of model that looks at something that is static on paper and parses that and understands how to interpret it, and there's a different type, that is conversational AI, that can have this interpersonal reaction in a way that feels human-like.

Speaker 3: 8:51

Two different capabilities, and so when we were looking at how is it that we want to deploy this, we were looking very specifically at both of those together and we found a lot of tools had really great one of them, yeah.

Speaker 2: 9:04

That is something we've talked a little bit about this. When we think about the capability of tools, when you look at kind of what's in the marketplace today, but let's say, the next five years, how much do you think that's going to change? Are we going to like pivot from the vendor that we've chosen? Is that something you think they'll evolve into it? Like we're still very much in the infancy of AI technology. So, like from your exposure, what do you think AI tech will look like in five years?

Speaker 3: 9:32

It is moving really quickly.

Speaker 3: 9:34

I think the easiest thing to say is that I definitely can't comment to what the underlying foundation models will look like in five years, but what they will be able to support will change.

Speaker 3: 9:47

For example, one of the reasons that we also started with screening is from our look at the market. Screening is the tools that support screening are a lot more advanced than, say, sourcing. There are a lot of sourcing tools out there, but as we were looking at like, how high of a level we need this to operate at, uh, sourcing hasn't quite hit the same mark as screening yet, but that's evolving incredibly rapidly, like within. You know, it takes us a while to evaluate a tool and then deploy it. By the time we're done deploying the first one, the second one could be right there with it. So it really is moving incredibly quickly. And I think what I'm seeing is the evolution in how many different scenarios something can work in, like sourcing. What that looks like and the different types of roles or where you're getting them from or how it's finding those candidates is evolving incredibly quickly in terms of what the AI capability can support.

Speaker 2: 10:53

Yeah, that makes sense. Interesting. We talked about the business case. I do want to kind of address the elephant in the room a little bit of like compliance. Well, I don't care about that right now. I do want to talk about compliance, but what I want to talk about is we talked about the human value, which is efficiency, being able to deliver for our clients at a better level. I do want to talk about the client side, which would be compliance, but to me, the like elephant is what the recruiters think 100%. So I want to talk about that recruiter benefit because I mean, we we hear it all the time. I think there was a recent study we're going to have to look this up where like 50 to 60 percent of individuals believe that their jobs are at risk because of AI, and that's industry agnostic. If I'm remembering the data point correctly, we might have to bring that correction section. Oh my God, oh yeah, I think that is a throwback.

Speaker 2: 11:42

Although I'm very confident in what you just said 78% of data points are made up on the spot. Did you not know that? Yeah, I think it's actually 87%. I think you're right. I'm sorry, I apologize. Recruiters.

Speaker 4: 11:54

Yeah, it's the recruiters and the candidates and the candidates Interesting. Okay, and specifically the tech that we're looking at, it's clearly AI that is engaging with the candidate. Yeah, yeah.

Speaker 3: 12:07

When we think about our stakeholders for this, there's clearly human. There's also our partners, and the thing that we find that they are thinking about the most is the compliance side. Then there's our recruiters. We're thinking about how does this impact my day in the life, but also am I in jeopardy? Is this going to replace me? And then our candidates how is this interacting with them? How does it impact their experience? We'll go back to the compliance piece. It's a much larger question. When it comes to the recruiter experience, we firmly believe that this is a tool that can be used by recruiters, not replaced recruiters. There's a lot of discussion and there are different takes on this. Where human stands today is that agentic AI, which means like an AI that can do sort of everything from end to end, standing in the place of a person, isn't actually a direction we want to go from a compliance perspective. Okay.

Speaker 2: 13:07

Interesting, I would say, not even a compliance perspective, but like the service we provide is so catered to. That white human experience, right, literally, like, literally, h U M A N experience, yeah, yeah, so a hundred percent.

Speaker 3: 13:21

True, I bring up the other one, um, because in terms of like, okay, but you could change your mind on that any day For there to be a human involved in the decision making, having oversight over this type of of tool and activity, that we think that the, the administrative work, the taking notes and loading them in from a screening call, the trying to schedule a screening call to happen things move.

Speaker 2: 14:19

Funny you mentioned this like the scheduling of the call, because I knew that was a benefit that with the AI technology we've selected the candidate can pick their own time for their screenings. That removes it from the recruiter. But we, historically, that is in such a pain point for us that we actually, with our client hiring manager, set up blocks of time that are pre-booked. So when we're looking to like schedule calls and stuff, we actually pre-build time to make it easier for recruiters. So this is just an extension of we know this is already a pain. Yeah, let's make that easier.

Speaker 3: 14:51

Well, and there's scheduling in two ways, right. So there's scheduling to have the phone screen with one of our recruiters and then there's, after we've moved a candidate along, scheduling between the candidate and the hiring manager. So we have scheduling sort of in two places. That is a huge pain point and a huge use of time, and so the goal here is that less time spent doing that is more time spent really understanding candidate qualifications and having those relationships with our hiring managers to be able to move forward candidates who we think are the best fit for those roles, and so that really puts time back in the recruiter's day to be able to do those higher value activities and spend less time on the stuff that no one likes to do anyway. Yep, yep, from a candidate perspective. Hil Hilary, you touched on this right out of the gate with scheduling. Yeah, yeah, ai is available 24-7. It doesn't sleep, it doesn't be sure to time it.

Speaker 2: 15:55

Oh, that's why the robots are taking over. Here we go.

Speaker 3: 15:58

And so you know for us where we work a lot in healthcare right, you might be hiring a nurse and they may have just gone off a 7 am shift, or you know they have a day shift and are only available in the evening. A recruiter may or may not be available to take that recruiting call at that time. This means that a candidate can take that screening call whenever is most convenient for them. They can take as long or as short as they want and really fill in everything that they want to say and that goes back to the recruiter for review. And so it's. We think that in a way that the candidate experience we hadn't been able to be flexible or to have as much to think about the candidate experience in that way in the past. We now can, and so it's a totally different way of thinking about candidate experience.

Speaker 2: 16:46

So, dina, I want to ask you this. So, because your world is very different, we're implementing AI on the RPO side right now, correct? Yes, and you live in that direct hire with a very different type of structure? Yes, when you think about this or the conversations you've had, where do you see this changing your conversation with your clients?

Speaker 4: 17:07

Yeah, so gosh great question. So I think, for us, I don't know how this is going to change my conversations with the clients, because the technology we plan on adopting some of the same technology of everything we're doing is under the human brand, and so I think it's probably more in our sales pitch, to be candid, it's telling clients being able to have this tool. At our gig level, we have AI which makes us more efficient, which makes scheduling your interviews easier, so I think, leaning into how it makes us more efficient, able to identify candidates, a little bit.

Speaker 2: 17:42

You think you could hire recruiters more easily because of this. Could I hire recruiters more easily? Like? Is that a part of our employer value proposition? Now, to say less, we remove the burden of like administrative work.

Speaker 4: 17:55

So I will be so interested to see what our internal adoption is of this platform. I'm excited for the launch that we have and kind of the rollout, but I think the internal adoption is really going to tell us how we need to position this for recruiters If I was a recruiter and I didn't have to paper screen all day.

Speaker 2: 18:12

I guess that's a question because we've seen what's the response been of our people and our candidates with our pilots.

Speaker 3: 18:18

So far it's been really positive.

Speaker 3: 18:20

Obviously, with any rollout and new technology, there are always sort of bumps in the road, but we've been able to solve them really quickly and the response has been really great.

Speaker 3: 18:31

We're actually excited. In a couple of weeks we're doing a panel with a couple of our recruiters and the teams that have launched it so far and we'll be able to hear directly from them on what their experience has been to date and on the direct hire side of the house. We're really excited from an AI perspective, because AI to this highly repeatable process piece, one of the things that makes it valuable is around high volumes and in direct hire or for the partners that we typically service in direct hire, their volumes may not be as high, which might make it harder for them to use AI on their own for it to actually be valuable. Now, through human direct hire or through the tool that we can offer like, for example, across a PE portfolio company or, sorry, across a full PE portfolio we can now get to the levels of volume that they might need to actually be able to leverage AI.

Speaker 2: 19:29

Yeah, and that's an interesting pivot into that client side. So will you talk to us a little bit about, like, the value that our current clients are seeing or hoping to see, and that could be efficiency but also beyond. And then those conversations around compliance. I know you've mentioned before that different types of organizations have different compliance concerns, so I'd just love to hear like kind of your insight into the discussions there.

Speaker 3: 19:52

Sure, so I'll start with the values that we think that this brings for our clients. One obvious one is increasing time to fill.

Speaker 2: 20:06

Yeah, that is quite the pitch, emily. We'll make your time to fill larger, crushing your job today. Keep it going, love it, love it ai tools for everybody.

Speaker 3: 20:17

Um, we so making the speed to film a lot faster, yeah, and so, uh. So that's one obvious one. And and how we get there is is also, I think, fairly straightforward. You're able to screen candidates faster, uh, schedule those calls faster. It just gets a candidate through that pipeline a lot quicker.

Speaker 2: 20:37

And one of the things I do want to clarify is in this we're not kicking, we're not like deleting candidates. This is not limiting their talent pool. Like everyone who applies still lives within their system. This is just a way to be able to review and bubble up talent in a more effective way.

Speaker 3: 20:50

Correct and truly it happens faster, like if you think about the time it takes for a human being to look at a resume and scan through it, it's at least a minute. Yeah, truly, the tool can do it in seconds, and so just the order of magnitude is enormous. Same thing with a screening call. Right, if we can do screenings after hours and scheduling automatically, that means that the screenings have already happened when a recruiter comes back to work the next day, and now they can review all of that instead of then conducting all of those calls during the day.

Speaker 4: 21:24

I mean, if you think about it from a recruiter perspective, if I post a customer service job at 5 pm, I'm going to come into work at 9 am the next morning and I'm going to have 100 candidates in there and that is my entire day. Yeah, and I'm lucky if I can get 10 of them on the phone. That's a really great example. So you're getting rid of that entire part. Here it is.

Speaker 3: 21:43

These are these candidates who have been preliminarily screened, vetted, and I mean game changer Instead of coming in posting at 5 pm, coming in at 9 am and having to now go through all of those. Try to schedule, get a couple of them on the phone. I come in at 9 am and I already have 20 of those screens done and now I'm reviewing those instead. The second piece that is a huge benefit to our clients is we think quality of candidate will also increase Because if you think about that speed in reviewing all of those resumes and applications, you have a natural lag where you might go through the first 10.

Speaker 2: 22:27

100%, and we find five and we pass them on. They hit your desk you look at, pick the best ones of those top 10.

Speaker 3: 22:33

Here's what satisfies those, and so let's pass them on to the hiring manager. Now we review all 200 at once.

Speaker 2: 22:39

Yes, you know it's interesting because I hadn't even considered that element of it. And, to your point, you are going to end up with a higher pool of candidates from which you can find quality, but it also opens up a more equitable screening opportunity. But it also opens up a more equitable screening opportunity, and I know that compliance and bias those are things we've talked a lot about as it relates to AI, but I hadn't even considered that. If AI can do a consistent screen of all 200 candidates who hit your pipeline, then you actually are giving all 200 people the fair shake at that assessment, as opposed to who were the first 10 in the door. You get the most opportunity and then the next five and maybe the next three. I hadn't even considered that.

Speaker 4: 23:24

I'll be interested to see candidate responses to this, because ultimately, we are going to be handling addressing more candidates. How many candidates right now complain about ghosting A con of them? Everybody complains about it, so the theory of ghosting is going to be. It'll be gone. However, what that means is you're now engaging with an AI technology as opposed to a recruiter, so there's a trade-off. You're going to need your information, but it may not be with a human.

Speaker 2: 23:50

So I'll be interested, but I do want to say the human, the person, the recruiter still has to decide whether they're going to move forward or reject Correct. So you could get a response but then not know if you're actually being considered or move forward. So there's still the onus on the recruiter to make that decision of yeah, you had the time to take the call and talk to our AI assistant, but you still like, I agree with you, with you, but I think we can't remove the responsibility of the revertor. Yeah, yeah, yeah, no, I didn't think about that either. You get more of a direct response.

Speaker 3: 24:20

Well, and some of what we've heard so far is both we have seen an improvement in time to fill on the quality of candidates. We've seen where we've been able to look at this and where our technology partner has been able to look at this. We've seen an improvement in retention because the quality of the hire is a better fit for the role, based on that ability to look across a larger candidate pool.

Speaker 2: 24:46

Is that a first like 60, 90 days? I would have to look in specific, but it's like we're not looking at like long-term retention yet because we're still early. I'll be on a bit.

Speaker 3: 24:54

Yeah, yeah, on retention, one of the challenging pieces of measuring it is that you have to wait a long period of time. Correct, people who were going to leave have to leave for us to know that somebody stayed instead, but we have some leading indicators that show that this is actually helping with alignment. The other piece that is a potential indicator of quality is the submission to hire ratio. So as we put forward candidates that have been screened for interview, a higher percentage of the candidates who are being put forward then ultimately get hired.

Speaker 4: 25:28

Love that yeah.

Speaker 3: 25:31

On the candidate experience side so far we've had really positive reactions. Same thing with our technology partners. So we have seen like four and five star ratings out of five, some feedback on the candidate experience across different roles, roles and within healthcare, where sometimes people, I think, are particularly nervous about different roles and how they might respond to that experience, and we've gotten some commentary back. I think we had one. I particularly liked this one that was like it was a seven out of five experience, which I particularly enjoyed Very specific rating.

Speaker 4: 26:10

I think that goes to the increased quality of the service that AI is providing, because if you look at the AI chatbots a year ago or not even chatbots, the what do you call it conversational AI where it's actually, if you look at that a year ago, it wasn't great, but the tech that we have is absolutely incredible and so it just continues to evolve.

Speaker 2: 26:31

It's like you're talking from. It's like a year ago I was like talking to Hal from 2001, Space Odyssey, and now it's like an actual. It's like talking to Emily. Emily actually is AI.

Speaker 3: 26:44

She is a robot. It is so fun to play plot twist clips for people.

Speaker 2: 26:49

Oh, it's awesome.

Speaker 4: 26:50

Yes.

Speaker 2: 26:50

But also this is where sorry it is, because like this is where I see the recruiters get the most nervous. Oh, you're 100% correct when they're like oh, it is almost too cool. So that's why I wanted to start with the recruiter saying like yes, it is unnerving, but think about the fact that, like you don't have to have all those phone calls Like this is now.

Speaker 4: 27:08

Yeah, it is completely transforming the day in the life of a recruiter and it is just giving an opportunity to upscale.

Speaker 3: 27:20

And I think the best way to think about it, and the way I encourage recruiters to think about it, is first it's coming regardless, so how do we deal with it? But then, after that, what it means is less that it is replacing my role and more of that the role of a recruiter might evolve, and so what it looks like to be a recruiter might just shift from a little bit of some of the activities we know today to a different set of activities which, in my opinion, are much higher value activity, or a larger percentage are those higher value activities. I agree.

Speaker 4: 27:49

This is if we layer on our favorite TA maturity cycle bottle, because we love the Zena Lloyd maturity cycle thing. I mean, when you look at kind of the pinnacle of a recruiter, it is not that you are a person who is paper screening resumes and sending them over, it is that you are a strategic partner to your hiring manager. How do you become a strategic recruiter?

Speaker 2: 28:09

Recruitment partner Like you are candidates to your hiring managers, to your business. Yeah, you can't do that if you spend all day Exactly, if you're just pushing paper. Yeah.

Speaker 3: 28:18

Well, one of the things that ultimately this AI I love it Just want to say.

Speaker 2: 28:23

Love it. Hurrah to Emily Emily.

Speaker 3: 28:35

One of these things that we'll ultimately be able to do to be a strategic partner to both candidate and client partner is, let's say, you have a candidate come through and they're not a great fit for the role they've applied for. Well, the AI knows the rest of the roles that are open and it knows what screening questions and what sort of criteria you're looking the partner is looking for for that role and it can say I don't think this is going to work out for this role, but we have this opening over here. Would you consider that one? And so this ability to recycle talent into a place that's actually a better fit for them better fit for both candidate and client partner, I think, is another major ability of this and value you know your book of business.

Speaker 4: 29:15

You don't necessarily know what your fellow recruiter is working on and and when you talent on the table.

Speaker 2: 29:20

well, you know your book of business but you know, sometimes you forget what's on your to-do list when you're like knees deep and like doing the work. Um, I know that we've sorry, I want to, I we have to hit compliance, because I know we could do this forever and I'm just putting an eye on time. Compliance, just give us the 101.

Speaker 3: 29:39

It is. I thought this is where you were going earlier with the 100-pound gorilla in the room. It's really a scary thing to a lot of partners and it's one of the reasons why people hesitate on AI. The first reason is like I don't know how to use this or like where to put it or how to deploy it.

Speaker 2: 30:03

The second is oh my God, there's a lot of risk associated with this and I think you should have a healthy like concern about compliance, because you have to take it seriously.

Speaker 3: 30:08

It is totally a valid question and it's something human takes really seriously and that we have been deploying right in lockstep in parallel with our development of the solutions and solutions that we've put in to place. When I think about compliance, there are four pillars I really think about. The first is data. In order for AI to work, it functions off of data, and so what is happening with the data that's ingesting A lot of it, particularly in the talent space, is PII. It's personal information, and there's a lot of regulation around personal information, where that's housed, how candidates have access to it, right to privacy, right to privacy, exactly and so what are you doing with that data and how are you keeping it secure?

Speaker 4: 30:57

Yes.

Speaker 3: 30:57

Okay. The second is around legal, and I think a lot of people have question marks on this, particularly because it's an evolving landscape. Unlike a lot of other places within talent, this is not a particularly mature legal field. We don't know which directions particular cases will go. There's new laws coming out in different jurisdictions all the time, and so how are you monitoring and knowing that the way that you are using the AI because that's part of it as well it's not just what the tool is, but how you are actually using that tool applies in a particular case, and so, for us, we have a partner who's actively monitoring it on a global, national and local jurisdiction level to let us know both how are things evolving, is there a new law that's coming into play, and how is what they know about our use cases relevant for that, and where are there risks associated with something that's changed? And then also, what adjustment do we need to make in order to be able to remain in compliance with how that law is being presented?

Speaker 2: 32:01

So we've actively pursued an expert in this field of AI, legal and has this broad understanding to help guide us interpret and we have that ownership of AI for best case, for our business case. But then we say, look, we can't possibly do this on our own, so we make sure that we have someone to help be that guard in that way.

Speaker 3: 32:23

As anyone in compliance knows and I'm sure a lot of our healthcare partners know this as well it's an ever evolving, and so it's not just about getting it right once. It's about making sure that it's continually right as things evolve and change. The third place that I would think about when setting up governance and compliance is around your internal governance. What policies are in place? How are you evaluating the tool usage? How are you handling any kind of issue that may arise? How are you handling any kind of issue that may arise and how are you?

Speaker 3: 33:00

You know, at what frequency are you testing your AI to make sure that it's operating the way that you expect it to be operating? Everything around that what is your internal policy and governance? And then, how is your team trained to understand? This is our policy, this is our governance, this is how we expect to use the tool, and the final piece is actually on, and this piece, I think, is part of what diverges from traditional data or IT security in particular, is the underlying models and the use case. How are you using the tool and how does the risk that you are taking on change in the way that you're using a tool? For example, if the tool makes a decision, it's a lot higher risk than if the tool presents information that impacts a decision, and so, really, important distinction, I think.

Speaker 3: 33:52

And then, similarly to the underlying model piece, how is it trained? Is it using real data or candidate data? Is the data that you're sorry synthetic data or is the way that you are training it evolving in real time? Or are any changes to the model something that you are like I would call a suggestion and doesn't get deployed unless there's a human intervention to say, yes, deploy that chain Got it.

Speaker 2: 34:19

Yeah, got it. So give us, in short, those four pillars again.

Speaker 3: 34:22

All right, so we've got governance. I'm going to do these not in the same order. That's fine Governance. Legal.

Speaker 4: 34:28

Okay.

Speaker 3: 34:31

Use and underlying model.

Speaker 4: 34:33

Okay.

Speaker 3: 34:34

And data and data, and so there's a lot to think about there, which is why one of the things around AI for I think some of our partners is so scary the nice thing is, when you have a partner like human, is that we've done all of that and it falls in place. You don't have to think about it.

Speaker 2: 34:52

This is all you've essentially done for the past little while.

Speaker 3: 34:56

There's just so many little components on that one. There's a lot to think about there. That's awesome, and it's as small as when you set up the AI. You know you call a candidate, the candidate picks up and it says hello. Just so you know this is AI right, like little things that need to be in place to make sure that you are compliant with a variety of different laws, security, transparency, et cetera. Totally.

Speaker 2: 35:24

I would love to keep rolling on this, but I think that we should probably wrap it up.

Speaker 3: 35:29

My job is not yet automated by AI, so Sure.

Speaker 4: 35:33

I'm just hoping to get there. I mean, I'm just wait, I just need a moment. I'm just so impressed right now Like wow, emily.

Speaker 3: 35:41

We've. Thank you. Wow, we've come a long way no-transcript without how much work that they have put in, and all of them have other responsibilities as well, so they have been just an incredible support to this project, that's awesome, very cool incredible support to this project.

Speaker 2: 36:20

That's awesome, very cool. Final thoughts.

Speaker 3: 36:25

What is the most important thing you want to tell us about AI that we've not asked about? I think it's exciting, it's evolving quickly and the biggest hesitations come from how much mystery is around it. So the thing I would like to say to our client partners is, if you find yourself hesitating, ask the questions that are keeping you from hesitating, because it's probably something that made us hesitate as well.

Speaker 2: 36:49

We've pursued that to understand what the solutions are, or I would even say, and you got me thinking sometimes you don't know what questions to ask, so we can send Emily in to be like here are our pillars, let's talk about them.

Speaker 4: 37:03

Yeah, I think framing up the four pillars super helpful.

Speaker 2: 37:08

We're doing a live correction section Dina Correction section. I have missed that. It has missed me.

Speaker 4: 37:14

Emily actually knows how to sing now.

Speaker 1: 37:24

Yeah, but we're not going to ask her because I like the octoon. Maybe the headbutt nailed it. So, hillary, your guess was pretty close, um, to how many people think their jobs are threatened by ai. Okay, um, it's actually a little lower, though sorry, I went doomsday 60 percent. Uh, what's interesting is, 30% of workers go wide fear that AI might replace their jobs. And what's interesting is, people are pivoting into accepting that personal proficiency in AI is actually critical to their job security.

Speaker 2: 37:52

Oh, interesting, interesting, okay, interesting, wait, wait, wait. Well, say those numbers again, because I think that's first of all. I was way off on the 60%. Thanks for trying to throw me a bone on that one.

Speaker 1: 38:02

So 30% actually fear that and 45% want to become better in AI. Now I'm going to contradict those stats by saying uh, 73.6% of all statistics are made up.

Speaker 2: 38:16

There we go, there we go. That one is official Boom, but I do think it's interesting. The most interesting thing about that, I think, is that we see this decrease in oh, ai is going to take over my job and more AI is a threat to my job if I don't know how to use it.

Speaker 3: 38:35

Yeah, which I mean we could play that over and over again in what we've seen in the last 20 years, right, Computers? Oh, absolutely yeah, I think it's just a continued evolution in the way that we work.

Speaker 2: 38:48

I still use carrier pigeons for my business communication. Emily, thank you for being our guinea pig.

Speaker 1: 38:53

This is awesome First guest. Appreciate it All, right, yeah, have a great day, guys.

Speaker 2: 38:55

First guest. I appreciate it. I appreciate it. All right, yeah, have a great day, guys. Woo, how do we normally end these things?

Speaker 1: 39:01

I have no idea, you just go like. It's been a pleasure to be here. Bye.