← ASU+GSV Summit Sessions
ASU+GSV Summit Sessions
Unlocking Trust, Speed, and Skills in the Increasingly Contingent Economy
00:40:58
Loading summary
Transcript49 lines
- [00:03]
A
This session was recorded live at the 2026 ASU GSV summit in San Diego.
- [00:14]
B
I'm going to go ahead and just kick it off with a really kind of rapid fire question to get us going and to level set on the state that we're in. What do you think has really changed the most in the last say 6 months? Months when it comes to contingent workers, be it around skilling training, you know, retention, what is really different in the landscape today? And I'm going to go down the line, put Ken on the spot.
- [00:41]
A
Sure. Let's get a skip. Oh, that's loud. Yeah, I think a lot has changed. I love that you mentioned like six months too. Like six months is not a particularly long time. And I think the amount of distance that we've seen and skilling has obviously changed a lot. I think two things jump to mind for me. One is the skills themselves. Right. I'm sure everyone's tired of hearing of AI in this conference, but AI has gone from 0 to 100. Not just in coding, but many other different disciplines as well. And what we always are hearing is how do I judge AI? What does that look like? So the types of skills that people value have changed. And then I think really where we specialize is how you assess those skills. So everyone is trying to use these new tools to find ways around traditional assessment and skilling. So for us the cat and mouse game has really emerged of what are the new tools? How are people using those tools to cheat, pass assessments or bypass them? And then how can we use those tools to kind of re level the playing field? But the too long didn't read is a lot.
- [01:48]
B
Nikit, why don't you go ahead next?
- [01:50]
C
Well, I love that. So I'm the assessment person on the panel. So when I love when I hear things like cheese pass the assessment go around assessments. Because I think when people think about assessment, they think traditional assessments. And I think one of the things that has changed, I would say in the last three or four years, but particularly even in the last six months is assessments don't have to mean traditional assessments. Assessments can mean performance based assessments, live practice, authentic workplace assessments. So we throw assessments out there like it still means fixed form, multiple choice assessment. And that's not what assessment means. And I think in the last couple of years and even in the last six months that's changed dramatically. And I think what's also changed is the demand again for fairness, for standards and for ensuring that there's not just authenticity in the way that we're utilizing the innovations in the industry, but also transparency and fairness and, and the capability to really validate that. What you're doing is not just great in essence for showing the skill today, but for ensuring that it's fair, it's valid and it's also being deployed with a sense of integrity.
- [02:55]
B
Tigran, how about you?
- [02:59]
D
It's funny, I like the six months part because usually not much changes in six months, but we live in a world where actually six months is a lot of time and things change. I actually think the biggest change affects not just contingent workers, but the entire workforce and I think it's in the AI model enhancements. There's a lot of talk online about, you know, when are we gonna get to AGI, which is artificial general intelligence, which presumably will be able to do most knowledge work. And I actually think we got AGI in the last six months. And the reason people don't perceive it as AGI is it's just not integrated into the work yet. So you've got intelligence that essentially doesn't have access to tools and doesn't have access to systems and doesn't have access to context to be able to do most knowledge work. But from the perspective of can can, which what it means for all of workforce, that means every job is going to change. The job is going to become more about managing AI rather than doing the job that you used to do yourself. And that's a fundamental shift that I don't think people are paying enough attention to.
- [04:10]
B
Julie?
- [04:12]
E
Yeah, so maybe I've got an alternative view. So I would consider myself the practitioner on the stage here and I'm a chief learning officer for a large business process outsourcer company. So think global contact centers. We hire 40,000 people a year in 22 countries around the world to support clients in every industry you can think of. We start with 3 million people top of funnel that go through our talent acquisition process and we have used for years what I think is fairly mature hiring assessments. And that's not only, you know, assessing, you know, for competencies, but it's also looking at personality profiles and it's also at times has real, you know, live experiences that we ask them to do. So it might be that you record yourself responding to a client call on a particular topic. So how well can you do that? And so my contact center agents aren't using AI in the same way that somebody else in a professional role might use it. We bring a lot of AI driven technology to their work experience, like bringing knowledge directly to their desktop versus they have to go find it in a knowledge base like real time language translation, like a number of things like that. But so my sort of alternative experience and you've got this continuum with AI where you've got things that are moving really fast on one side, but not necessarily for every company. So my practices I've used, we're integrating AI significantly in that hiring funnel and helping us get faster to the candidates that we want to actually interview and consider. But I'd have to say it hasn't changed a whole heck of a lot in six months.
- [06:22]
B
It's interesting what I'm hearing across the board, aside from new tools and assessments. Technology is potentially change, or maybe not of skills or underlying skills. Do you feel like there's a new set of skills which are relevant today for contingent workers or maybe not so much?
- [06:41]
E
Well, so I would say that the profile that we're looking for absolutely has changed because one of the things and Brian and I were just talking about it, is when you think about this particular instance and we've all had to call customer support for something, you're changing a flight, you're changing a beneficiary, you fill in the blank, you have to return an item. Most of that stuff you do in an AI assisted way now. And so you don't actually have to talk to a person, which most people are really happy about, but you do still need to talk to a person for the complex stuff. Right. You didn't get a particular procedure approved with your health insurance, you're going to be making a phone call about that and that you're going to talk to a human in its high emotion and it's high context. And so that contact center agent needs to have the significantly more human skills in being able to navigate that conversation, that text, that email and getting to a successful conclusion. Whereas we used to hire people who could just find stuff in knowledge bases and manage multiple screens at the same time. So it is a different profile, but we have a practice where we are continuously using data across our 40,000 people to identify what is the success profile within this particular client and context. And we refine it on a very regular basis multiple times a year. And that's what we use to go hire.
- [08:11]
D
How long do you think it will take until AI does the complex stuff too?
- [08:18]
E
It's really hard to say because companies and industries are in different places in terms of their willingness to move in this direction. You've got early adopters and we've all read the headlines where they're like 100% replacement with some spectacular fails. And there's other companies who really put value on the human experience. And there's a lot of consumer research that are like, look, I actually want to talk to a person I actually like for some stuff. I want AI, but I don't want it for everything. And so we're of the belief right now that it's not going to fully
- [08:58]
C
automate everything when we were just talking. I'm a chief Assessment officer for Prometric, and we actually verify proficiency and credential people in 180 different countries and have for decades. And so even at this conference, we were having conversations with many people around the way they verify proficiency, how that has changed, how the shelf life of skills is getting shorter and shorter, and how AI is taking over much of that work. And to that very point, what I found interesting yesterday is it was in a financial industry and some of those sort of contingent workers and in positions in general, they said what I was doing sort of in those levels of positions even, you know, a few years ago is not at all what those individuals are doing today. We have 100 people in those positions that come and go quickly. They're not doing anything that I was doing at that time because AI is doing all of it. We, we still have as many of those people as we have today. They're just doing very different things because they're getting fed the information by AI, and now they have to treat that differently. It's all about interpretation, it's all about application, it's all about problem solving. But we don't do any of the same analysis because AI is doing all of that for us. So they don't have less people. They're just now having to adapt those skills, apply those skills differently, and assess for different skills coming into those positions. I found that to be really interesting because they specifically said, we had 100 people doing that job two years ago. We got 100 people doing that job today. They're just doing it very, very differently.
- [10:25]
A
Really quick, too. I think, like, piggybacking off of that. Like, one of the more interesting words on that screen is contingent. Like, you know, I think what we're starting to see a lot of, and I expect a lot more of, is rolling over those hundred people to solve very specific types of issues. You know, when I think about FTE five years ago, I almost. This is a weird analogy, but I imagine like a mountain range where you've got a couple really hard problems and really hard mountains that FTEs are paid to solve. And in the meantime, there's 100 of these minute tasks. You know, as AI starts to pick up more and more of the minute tasks, we're starting to see firms, especially the largest enterprises, really hire people for very limited amounts of time to solve the big peaks. Right. So I think that work is changing, but so too is the shape of the worker. And it's frankly like something we didn't expect, you know, maybe a year ago. We're starting to see in the data.
- [11:22]
B
I'd love to double click to use that acne phrase on the assessment side, which is how people feel comfortable with the new forms of assessment as well as capability around like fidelity of those assessments. Also because you're trying to arguably measure things which were harder to measure in the past. Nikki, do you want to start there?
- [11:46]
C
I can start then. I'm sure lots of people, I can see lots of people have other things to say. I think I've been in the assessment industry for over 30 years and I'll be very frank, there was very little change in that industry till about five years ago. You've all taken tests forever, I'm sure. What were the big three enhancements? We went from paper to computer. Ooh, right. And then we went from fixed form assessments to adaptive assessments or banks assessments. And then we went from computer based to remote. Right. In 30 years, three big innovations. And in any other industry that would seem like the stone Age. Right. And in the last five years we've had massive innovations to our assessments, thanks to, in some part to one of my co workers who's developed really wonderful products for our assessment industry. But it's not just about how we develop the assessments. It's about how we define a job. It's how we recognize skills that are emerging skills, how we define skills that are dropping off that you don't need anymore. It's how you assess those skills, score those skills, deliver those assessments. It's all of it. And it's completely changed the way we think about it and the way we do that work. A lot of our clients are now assessing things differently. Not all of our clients because they again, have some of those considerations about the way they move that. And it's all about the integrity of how you do that, the fairness of how you do that. And I think in some ways they are appropriately cautious because they want to ensure that as we roll out very amazing ways that you can do things now that they don't inadvertently cause an issue because of a lack of measurement fidelity. You inadvertently harm a subgroup of individuals because it looks good when you initially roll it out at an aggregate level, but then as you get more data, you see an adverse impact at subgroup levels, things like that. And they may not have the infrastructure to ensure transparency, especially as you're talking about AI. Right. If you're a lot now we think a lot about continuous sort of evaluation of on the job performance and things like that. If you've got continuous evaluation, you also have to have continuous transparency and you have to have continuous capability for your workers to consent and to be able to audit and to be able to do those things. So some of the infrastructure is not there yet. So I'm excited because we do VR assessments today. We do scenario design, we do live coding, we do all of those things. But it's not as broad and quick as I would like it to be from a purely assessment standpoint when you talk about high stakes assessment. But that's rightfully so as the right measurement and validation and fairness considerations are rolled out. But I'm very happy that it's going much faster than it ever has before because we've got better tools, specifically by people on this stage putting out amazing tools to be able to do and complement things that we're doing today because we have to get better at that single response. Multiple choice items that are traditional knowledge based assessments are not going to get to the kinds of things we need to and cannot adapt as quickly as we need to as new emerging skills come to the table, which they're doing faster than they ever have before.
- [14:44]
D
I mean, interestingly enough, we've always agreed that in principle, simulation is the best way. Can you simulate the job? Right. That's why we have a driving test. Imagine if we just handed over driving licenses to anybody who passed the multiple choice. Even more chaos than we already have on the streets. Now, the problem with how do you create a driving test for every job? Is it's hard to simulate most jobs. And which is why I think the innovation that you're talking about started from coding, because that's a little bit easier. You give people an IDE and it's very clear what's right, what's wrong. And once they perform the task, you can say, okay, you did it right or you did it wrong. Over the last few years, what has happened is we got the option to create simulations without VR because people don't actually have the VR headset. I think that was one of the first ones.
- [15:38]
C
You don't need VR headsets anymore for VR.
- [15:40]
D
Exactly right. So like, I think the first one was like, oh, let's get VR simulations. I'm like, okay, that headset is like $600. But with gen AI now you can create simulations for almost every knowledge work without needing VR. And it could be very high fidelity. So like for example, we do in addition to pretty much every job category assessments, we do sales assessments. And when we create a sales assessment we say first of all, can you define the job? Because that's actually you ask 100 people, it's like, what is the right way to define a job? And to me it's inputs and outputs. So what do you start with and what are you expected to deliver? Salesperson that starts with a product or a service expects to deliver revenue. That's as simple of a, you can define sales as possible, right? An engineer starts with requirements and delivers product features or bug fixes. Once you have the inputs and outputs, then you can say the job is to take those inputs and deliver those outputs. Then you can simulate that. So a sales simulation would look like you start by maybe understanding the product by talking to maybe the head of product of a company you just joined, a hypothetical one. Then you would go talk to a prospect. These are all role played by AI. Like this kind of technology was just sci fi two years ago. And you can also score this automatically, which is something you couldn't do before. If you have a rubric, if you have a very well defined rubric, you can actually go through and score it better than a human assessor, which used to be a job, I mean still a job, but human assessors used to do this for decades and now it can all be done with AI at a very, very large scale.
- [17:27]
E
So we've actually been doing this for years. And I'll say that the contact center agent role is a great one to do it with because they have very well defined KPIs that they need to deliver, whether it's client satisfaction, MPS average, handle time closure rates, things like that. And so we use it in our training for that distributed practice. The beauty of it is you do it one on one. And so it sort of, it replaces the horrible, horrible role play that nobody has ever enjoyed, right? And that also has subjectivity because you're relying on whoever you're role playing to give you meaningful and useful feedback. We do apply a rubric. It does tell them and give them coaching along the way. And then at the end it tells you this is how you scored and this is where you still have opportunity and you can rerun that as many times as you want. And it's actually work that they do on the job. So we do that not only in onboarding, but we've built that into one of their coaching paths. So we use AI to ingest, whether it's transcripts or it's live listening, where
- [18:49]
B
are
- [18:52]
E
and assess what are the skills and behaviors that they're not executing well enough to deliver whatever client KPI outcome that you need. And so part of the coaching plan might be more simulated practice. It's not always the solution, but we've got that really well integrated for this role. And so you could say that it's continuous assessment but for the purpose of ongoing development and performance improvement.
- [19:23]
B
I was going to say a lot of what is being described from a practical standpoint is a potential for continuous assessment. Do we think this is the default like end state? Do you think there's enough trust in industry for it? Is it inevitable or are there things that are going to get in the way of it or what do we need to do so that it becomes an inevitability?
- [19:45]
A
Deep question. I'll tackle at least the start of it. You know, what springs to mind for me is I kind of hate the word assessment, you know, because I think, you know, to your guys point, like assessment of the last 30 years does not equal assessment of the next five. We honestly try to use that new word simulation more and more because I think it opens people's minds to what's possible. Another word I use a lot is interview. Right. You know, today a lot of the world hiring and I think you can make the same argument in like a continuous cycle like after hiring is done by the interviewing of humans. Right. Let's watch how you work in a technical interview. Let's do a mock sales call. Right? Let me go listen in on some of your sales calls while you're on the job to see how you're doing. You know, AI has the can play the role of the human right. You know, evaluation of tests is not about the process. It's not about how you got there. It's all about the end state. So you know, where we see the unlock is if assessments can play the role of the interview and understand not just the end state, not just the pass fail, not just the scoring of a completed result, but how somebody got there, whether that's pre hiring or you know, post hiring. We think that's really the unlock to bring the power of a live interview or a live coach to every candidate. Right. Or every employee. So I think the scalability of the quality of an interview is Kind of what's going to unlock that new paradigm. But it's a deep question.
- [21:26]
C
Yeah, I, you know, I agree with a lot of just what was said there. I do think there it's the methodology and how you get there. It's not the end state of the pass fail, especially if you're doing continuous evaluation. And I think that's a super fair point. I think what people need to remember is when, even when we say AI can do that evaluation, people referenced here rubrics and things like that, and that's incredibly important. Whether you're using AI modeling to do that, you're using subject matter experts to do that. The point is you still have to have validation processes to do that. Right? You need to validate that the, that if you have a human being doing that, there's bias in that process. Right. In an interview process, in any kind of evaluation process with humans, there's potentially bias there. Just like with an AI modeling, there's potentially bias there as well. So if we're considering continuous evaluation, we also have to evaluate the continuous quality control measures in that process, number one. Right. And I do think there's a little bit still of a trust issue, right. If someone's going, if I'm going to be continually monitored, right. First of all, am I opting in to that continuous monitoring? And then what happens if I don't agree with what's happening with my continuous monitoring? I think there's that trust issue, that transparency, right. How do I get to audit or, or raise my hand if I don't think my evaluation is correct? And I do think there's that validity component of how are we monitoring the development and the evaluation of those rubrics we have again, in our simulations and in the work that we do in the wide variety of assessments that we give, there's lots of rubrics evaluated in that and developed in that, not just for the endpoint, but the pathway through the simulations. And we have measurement teams and pilot studies and all the evaluations on that to ensure not just the overall results are appropriate, but the subgroup analyses and the differential functioning and all those things are not disadvantaging any particular groups. And I think that's important from a trust perspective. And I think if we're thinking about continuous evaluation, I don't know if it's every job all the time. I mean, I think it's role potentially it's role specific. Right? Is it a job that's high stakes? Is it a job that is changing frequently? Maybe not high stakes. But a very fast moving, fast forward paced position that really requires continuous monitoring versus a role, maybe that's neither high stakes nor changing frequently. So I think it's probably a little dependent, but I think particularly specifically there could be some that move in that direction. I don't know that everything needs continuous monitoring. And I do think people will feel differently from a trust perspective if it's everything versus if it feels like there's a specific reason that this role is monitored. And, and I also think it should be role specific versus contingent specific. If I'm being monitored continually because I'm a contingent employee versus I'm being monitored continuously because my role demands it based on the task that I'm doing, I think that feels differently from a trust perspective.
- [24:20]
E
Yeah, I completely agree with that. And then just to give a very practical example, again I'm going to go back to this context center role where forever you have been continuously monitored and your performance assessed. But the way that we're using AI within the system that I was describing, to actually understand for their coaches who are, who are supposed to be coaching them and helping them get to improved performance, how effective are those coaches? Well, you would have to have the manager of the coach join a live session and observe and you try and train them in how to do it. But there's so much subjectivity in there. And then you'd have to keep calibrate and all of these sessions around calibrating not only coaching sessions, but coach the, you know, the observation sessions and just rife with variability and tons and tons of hours spent doing that. But now where there are approvals, right? And I think that's really important is allowing AI to listen in on that coaching conversation and then to look across that entire data set for what all of your coaches are doing and identifying, hey, the most successful coaches who are coaching on this particular skill, they do this right? So then you start to have a body of practice and best practice that you can disseminate across to the rest of your human coaches. So that's one way that I think that we're getting more calibrated using AI and in something that's actually shown with data to deliver improved performance.
- [26:03]
D
Just two comments. First on the continuous assessment, I actually think it's putting aside how people feel about it, it's the best form. So like simulation is the second best. Continuous assessment is the best. Because going back to the driving analogy, the person who taught you how to drive knows more about your driving skills than the person who took the test. Because the person who took the test observed you for like what, 20 minutes? And the person who watched you learn how to drive, that was hours and days. Right. So they know every little detail about what you're good at, which you're not
- [26:39]
E
talking about my dad here. Right.
- [26:40]
D
There you go. And why. It's powerful because it leads to better learning. I think when it comes to how do you get people comfortable with this, it has to be contextualized, not in. We're just trying to figure out if we're going to fire you or not. It has to be contextualized. And how can we help you get better? By continuously understanding where your skill set is and how it's evolving. And historically, what's been true is that this only worked for very well defined jobs such as support agents and things that had a lot of structure and changed a lot. I think all jobs are going to start to change a lot, which means this is going to become necessary more and more. And with AI, you can actually do continuous assessment for more complex roles than you could do before. And you can actually create learning opportunities for very complex roles that is highly contextualized. To give you an example, we talked about sales. If you could essentially listen and monitor and analyze using AI, a sales call that somebody had, you can pinpoint places where they do not pitch the product in the right way. And instead of saying like, oh, this person is not performing well, you can actually send them a direct personalized practice where they could role play, handling that type of question better next time around. Now as an employee, I'm thinking, wow, like I'm just getting this personalized coaching without my manager having to go listen to my calls and make me feel uncomfortable. And you're getting a continuous analysis of how well you're adopting to a very fast changing world.
- [28:18]
C
So I was going to, I'm going to ask a couple of questions about that. Number one, what happens when nobody taught me how to drive? So now I don't have anybody who has the sample set of information that maybe other people have. So now I'm at a disadvantage to other people because I don't have the same data set walking into that situation. So I don't have the same sample set of data. So I start at a disadvantage. So again, we're making an assumption from the beginning that people are equitable walking into the situation and they're not. So there's that piece of situation and then there's also in the sales call that assumes that people generally have the same opportunity presented in those calls to get equitable amounts of feedback, which they may not.
- [28:58]
D
Right. But if nobody taught you how to drive, you keep on crushing the car and at least we should find out about it.
- [29:04]
C
No, nobody taught me how to drive. I read a manual and I got in there and I passed my driver's test. I didn't crash.
- [29:13]
D
Usually if you didn't learn from anybody, there is a. You've got gaps that you might not be aware of and those gaps might be getting covered because the data doesn't display it. And it's usually in your best interest for somebody to actually surface those gaps.
- [29:29]
C
But it's not an unusual problem in a data set when you're making assumptions about preconceived characteristics of individuals in your data set. And it's not always going to be true for subsets of your population 100%.
- [29:42]
D
So you don't have to make those pre assumptions to evaluate. Right.
- [29:46]
C
But it's those things that you have to monitor and ensure from a quality assurance perspective that your subsets of your populations aren't being unintentionally disadvantaged.
- [29:54]
D
Right. Assuming that you can define the job properly, at the end of the day, it's not about the individual, it's about what is the job and what does it mean to do a job well, if those two things are clearly defined, you can do it well, if they're not defined, then obviously you're walking in and evaluating people on unequal criteria.
- [30:13]
C
Agree.
- [30:14]
B
One of the interesting things that's surfacing here is it's not just the measurement of things. Right. Like, or kind of move forward, let's say we can all assess wonderfully, it's happening at the appropriate pace and frequency that is suitable for the job. How do we get towards performance quickly? How do we get towards productivity quickly? I think this is a big topic when it particularly comes to contingent workers. And I'm curious what. Maybe we'll start with you, Julie, how you guys have thought about getting people and your like high performing now that you have the data, now you have some good assessment and understanding of what they can and cannot do.
- [30:48]
E
Yeah. A couple of things come to mind. The first is to have what you know about somebody travel with them. And so for a very long time, longer than we should have, it feels like, no, duh, why didn't we do this? But we've got all of this rich data during the interviewing process that then didn't go anywhere. It didn't transfer to their onboarding and their training period, which could be anywhere from a week to months. And so then you're resetting from zero on. The people who are charged with needing to help them develop their proficiency don't know anything. It's like they're a blank slate. So that's just an obvious, have that data travel with them and provide macro and then individual insights to that trainer and to that manager who is responsible for onboarding them. And then that understanding of what they're bringing with them grows. So then you're not being reactive, but you're proactively addressing specifically to them what they need. But then there's just kind of the tried and true, you know, basics of how do you bring somebody into an organization, whether contingent or not, in a way in which they are going to feel valued, they're going to feel like they belong, there's somebody like them there who has been successful and they're going to have an opportunity for growth. Regardless of whatever that time span is. If that isn't in place there, you know, you're going to see it, you're going to see it in disengagement, you're going to see it in not showing up on time. You're going to see it and missed work. You're gonna, you know, you're gonna see visibly them not engaging. And so the first thing that you have to do is you have to, you have to convince them that staying with you is worth their time. And it's going to be a good decision for them. And, you know, so that's why we haven't removed the human experience from this, because it's that human connection that is so important. And so we watch for signals, we record the signals. If we feel like somebody's at risk for leaving, if we feel that somebody is not able to keep up with the pace of what we're training them. And then we proactively, we proactively action that and we monitor how many people did we identify at risk that we kept, that we didn't keep. And then it's that continuous development using AI simulations as part of the solution. But not the only one to get them to proficiency the fastest.
- [33:32]
A
Yeah, I'd piggyback on all of that. I think where we have unique insight is we primarily work with staffing firms, especially really large staffing firms. They have a big problem right now, which is they are having a really hard time placing entry level folks at the Fortune 500. They try and try and try. You know, the analogy we give is like if we had Albert Einstein and we gave them to one of these companies, they'd say, well, Albert only has three years of experience. I want someone with 10. And you know, the reason all these companies demand eight years, 10 years of experience in their contingent population is risk mitigation. Right. In lieu of the data that we've been talking about for the last 40 minutes, if a worker has 10 years of experience, that's a hedge against the fact that they might show up and not have the skills necessary to perform. Where we get really excited is when you go back to that Fortune 500 company and you say, hey, this candidate has three years of experience. But look at the proof, right? Look at the simulation. Look what Albert did on Tuesday. All of a sudden, it doesn't matter whether Albert has two years or ten years of experience. That company is really excited to talk to the individual. And I think contingent is the first area where these people or this shift is happening because it's very related to the dollars. Right. These staffing firms are really incentivized to find ways to get these people employed. But I think it might be a broader opportunity that I think really impacts this conference. If you are a university, what does this mean for you? I think at the end of the day, if we can use assessment data and the fidelity of the assessment data as a new kind of spirit of evaluation, we can start to rely less on proxies like years of experience, where you went to school, et cetera. And while that's really starting in the short term contingent population, I'm hoping that we'll move to the broader workforce over the next several years.
- [35:28]
B
So it's definitely a shift in the last few minutes. We want to make sure everyone gets a bit in and also that our audience can take away something actionable. Right? Because I think the wonderful thing about these summits is we surface a lot of ideas and thoughts, and then it's the like, okay, great, now that we all have to go home, so what? So as you kind of think about, you know, if you were to give advice to a large employer or a platform around, like, a change that they should make in the next 90 days when it comes to their contingent workforce, what would it be? I'm going to reverse order it this time. Always lucky to sit in the middle. Julie, why don't you start?
- [36:10]
E
It's a good question. And I started off with the first question saying, I'm not sure that we've changed a whole heck of a lot. I mean, the role has changed and our success profile is changing. So I guess that's the first starting point is. No, what that profile looks like is It a combination of skills? Is it a particular personality print? What is it that delivers success within a particular role in a particular context? If you don't know that, go figure that out. Because otherwise you're kind of shooting in the dark, you know, years. We all know this, right, that years of experience doesn't correlate to high performance. It might in some individual cases, but you certainly can't use that across the board. So the more you can identify what matters most to deliver the performance that you need, go figure that out.
- [37:04]
D
All right, we're going in order this time. I'd say the biggest thing that you should change is be open to reevaluating your assumptions more frequently than you normally have. Because we're not used to this rate of change. We're used to, you know, making certain assumptions and then assuming that is going to be true for a year or two or three. And the best example that I can give is I still meet people who've tried, I don't know, ChatGPT two years ago, and that's their assumption about what AI can do today. And the reality is that it's changing every three months and accelerating in some sense of what is and isn't possible. And I think for us, it's very, very unnatural to be reevaluating our assumptions so frequently. And we have to really put effort into it to push ourselves to say, you know what, I'm going to have another look and understand where things are today.
- [38:03]
C
It's hard to follow that one because I think that's such a great point. And I think on both ends, which is whether you're contingent workforce hiring or general employee hiring, I think traditionally it's been an assumption that when you define a job, and from a credentialing standpoint, which is again, sort of a different view of it, but you do a job task analysis, right, which you define the knowledge, skills and abilities and job tasks, very formalized process. And people would do that once every three, five, sometimes seven years to the point of like, there's the assumptions about the jobs now even it's maybe, maybe they're doing it every three years. Maybe they're out so out of date by the time you complete those. So I would double down on looking. You've got to reevaluate those assumptions at a minimum, annually. But even more than that, to be looking at what those emerging skills are, what's changing, what's going to be critical to that position in that job today and be open to it not being just the knowledge base and foundation. I think that's important for some jobs to understand sort of your foundational knowledge base, but you need to understand performance based needs and then also emerging skills. And that does mean reevaluating very frequently, which is not traditionally what's been done in individual job evaluations.
- [39:24]
B
Ken, close us off.
- [39:25]
A
Yeah, Tigran stole mine, which I was going to say try, just try the tool, see what's out there, see it strategically. So I'll pivot, I'll go for another one. Maybe mine's surprising, but I'd say know your metrics. You know, I'm in this space 247 and even I get lost at the pace of innovation on this stage and in our field, it's a really hard space to kind of keep a hold of. And what I see a lot in our prospects and customers is people know they want to run with AI, they know they want to do something. But it's easy to see flashy demos. It's very hard to implement something that's tractable. Where we've seen people have success is they come with a very specific problem. And in the contingent economy, a lot of times that's early failure rates, that's time to fill, that's numbers, right? Something very tractable. When you find one of those numbers, I think the hunt for change and some of the tools on stage becomes a lot more fun. It becomes a lot easier to see how a change in one of these platforms or how you assess talent can impact these metrics or vice versa. So start with the metrics. And I think a lot of things kind of clear up and it's a fun time to be an assessment. There's a lot coming.
- [40:39]
B
Well, thank you everybody on the stage for joining us, for sharing your thoughts, for willing to speak up on your opinions, and we hope everyone enjoys the rest of the summit.