Episode Transcript
Speaker 1 00:00:06 Josh Berson. In addition to the regular podcasts we've been doing for the last three years, I'm gonna start a second series of podcasts dedicated to HR technology. And what I'm gonna do is every week or two, I'm gonna give you some insights into the product categories in HR Tech. And there are many, many product categories. The first one I want to talk about is talent intelligence. Now, in some sense, I believe I'm the one that named this category <laugh>. And one of the things that we do as analysts is we, because we talk to so many companies, so many vendors, we, we oftentimes see the way the market is changing and how new categories emerge. And so what talent intelligence is or where, where, where it came from was the original technologies that were used for AI-based sourcing went out. What these things do, and we're talking about eightfold, beamery, phenom, sky hive, tech wolf to some degree, although that's a much more limited product and seek out and others, is they use AI to go out into the internet.
Speaker 1 00:01:12 By the way, LinkedIn does this to a degree and they amass billions of employee profiles. They decompose the profiles into tokens or text, they look at relationships between the text and they determine things like skills, job history, career paths, and much more subtle things than that. Because once the AI starts running on this massive, massive data set, we can build models who look at things like, what is a career for this particular job role? So if we look at a job role like sales manager for example, we can look at all of the people that have that role and how they vary by each other, the algorithms, the AI can do that. And then it can look at all the prior roles that these people had to determine the pathways into that role. And it can also determine the pathways out of that role. And it can look at the skills of the people that came into the role versus the people that are in the role versus the people who went out of the role.
Speaker 1 00:02:13 And it can determine what the skills pathway looks like and how somebody who's a sales person or an S D R A sales development rep could become a sales manager and what the development plan would be for them. And it could also determine where that person could go next and their options. Now, it also does a lot more than that because once you have this massive data set of workers from around the world, including by the way the ones inside of your company, which is a small section of them, then you can start to look at things like, we wanna open a new office in Europe to study genetic engineering. Where are the people that know the various branches of genetic engineering? What cities are they in? What universities did they go to? What ages are they? And you can do labor market analysis to figure out where you might wanna locate a new office, a sales office, a r and d office, manufacturing office and so forth.
Speaker 1 00:03:13 You can use it to see skills that are trending adjacent to your industry. And so let's suppose Chevron or Exxon or whatever company you want, that's in the oil industry where the bulk of the workforce has been trained on exploration, production, refining, marketing, distribution, the things that I did when I first worked at an oil company. Yet we know that the energy industry is moving to solar, to battery power, to electric distribution, to hydrogen to wind to other things. What does the relationship between those roles and skills and how, what does the job family or job clusters look like for these new roles? And how are we positioned versus our competitors? And believe it or not, talent intelligence can do that <laugh>. And there's more, once you have that type of a system, you can create a talent marketplace. So vendors like Gloat and Fuel 50 who started life really analyzing internal mobility and facilitating internal mobility can use talent intelligence to recommend to an individual inside of your company a job opportunity, a position, a gig project, developmental assignment, et cetera.
Speaker 1 00:04:26 So this is incredibly powerful technology. Now, the difference between talent intelligence as I define it and as the market defines it and skills matching or or talent assessment is different. The assessment industry, which has been around a lot longer than I have, was really designed around IO psychology designed tests that were used and still are to a very large degree, to determine if somebody's suitable or trained for a job. So if I'm a, and I remember this example very well, if I'm a jewelry manufacturer and I need people to work in the jewelry store, I need people that are good with customers that can deal with very high priced sales situations, which are stressful for customers, obviously it can select a large number of jewelry that might be useful to this person that can determine their taste and that can be patient with them as they decide what they wanna buy and help them make that decision.
Speaker 1 00:05:23 That's actually not a simple job. There's a lot of complexities to that job. So what jewelry manufacturers did over the years, I forget the name of the company that did this, but I I talked to them quite a bit, is they built a pre-hire assessment based on what they knew about that job, doing a job analysis. And that assessment was actually done in a simulation form where you actually did a simulated experience of a customer coming into a store. That assessment told the company whether you had the skill or how well they thought you had the skills to take that job. That is not talent intelligence. It was the early days of it though, because the assessment companies like Korn Ferry and SS H L and lots and lots of others were generating a lot of data. They didn't really know what to do with it all.
Speaker 1 00:06:09 But little did they realize that the data they were generating was actually gonna be very, very useful for ai. That was before AI was really, you know, popularized at all. So the talent assessment industry, which is much more of a, a tactical oriented solution, is affiliated with the talent intelligence market because sure enough, once you build a talent intelligence system, you're gonna want to use some of these assessments because these assessments are very finely tuned for the jobs and roles and responsibilities in your company. And the data that comes back from one of those assessments is very useful, perhaps could be more useful than the AI generated data. So the talent intelligence vendors have to pull that data in. And what happened to the talent intelligence market as these vendors have grown, and by the way, Workday is, is getting there with the, the Workday skills cloud.
Speaker 1 00:06:56 It really doesn't have this level of intelligence today and nor does it have this much data, but, but they're all trying to go in the same place. Are realizing over the years that as their customers come to them with more and more data, they have to find a place to put it and the talent intelligence data system and a way to analyze it. So imagine you go out and you buy a beamery or Eightfold or Gloat or Sky Hive or seek out and you have all this data about all these employees and you say, Hmm, why don't we stick our employee engagement data in there? So we could look at the relationship between ENG engage and these other factors. Maybe we could run a model that would predict what it is we're doing that is causing people to leave. Maybe we should put our pay data in there.
Speaker 1 00:07:40 And since we're really worried about pay equity and pay competitiveness in the inflation, we could look at the patterns of pay relative to tenure skills, job roles, job mobility, and see if it tells us anything about pay and equities that we don't understand already. By the way, pay equity projects are not done this way. Pay equity projects are statistics projects where you actually take the pay data and you independently look at all of the variables that you have in which ones might be correlated to pay to make sure that gender is not correlated to pay that race is not correlated to pay that age is at least not directly correlated to pay. It may be indirectly correlated to pay because if it is, then you got a problem with bias. But anyway, what the, what the talent intelligence systems have to do, certainly the new ones, the second generation ones is they have to be able to incorporate new data.
Speaker 1 00:08:33 Because once you run these things and you use them initially for sourcing and you then use 'em a little bit more for maybe mobility and skills assessment, you're gonna say, Hey, we have this very special operational, uh, test that we give all our employees who work in this manufacturing plant. We really should stick that in there. So, so these become large AI data sets built on vector databases with probably lots of cloud-based AI services included that have to be easy to ingest new data. By the way, we've talked a lot about that. I'll talk about that in another podcast when I talk more about AI and chatbots. So that's sort of what the talent intelligence market does and how it looks. Now, not very many companies have these systems yet, but they're growing like crazy because they're usually purchased initially for recruiting. And once you use them for recruiting, you realize, well, if we're gonna do skills assessment of candidates, we should do skills assessment of the internal people.
Speaker 1 00:09:29 And by the way, if we're gonna recruit externally, we should recruit internally. And by the way, if we're gonna recruit externally and internally, maybe we should look at alumni, maybe we should look at part-time people. And these things kind of grow like wild wildfire because they really are in some sense the core of the next generation of talent platforms. And I got involved in this market very early with a couple of the vendors initially with Eightfold and a lot of the others. And honestly, I don't think most of the vendors knew how much power and value they were gonna be providing to customers. So it's become a more crowded by name space. Now, just because somebody throws talent intelligence on their product doesn't mean it actually does what I just mentioned the word talent intelligence could be used for a reporting tool. I mean you could say Excel is a talent intelligence tool too, which in some sense it is, but it's just not this category definition.
Speaker 1 00:10:21 Now a couple of the complexities of this space, I want to just mention that I'm sure you're gonna run into when you go to the next couple of months of co conferences. So if you're, uh, SS a P, they announced the Talent Intelligence Hub, Workday Skills Cloud is sometimes referred to in this way, although they don't use those words, Oracle is more than likely to announce something, which one of these things are real and what they do. So lemme talk a little bit about that because you're gonna hear about this technology from Eightfold. You're gonna hear about it from Seek Out, you're gonna hear about it from Gloat, you're gonna hear about it from Phenom, you're gonna hear it about from Fuel 50, and then you're gonna hear it from the ERPs and eventually maybe some of the learning vendors. But I think there's some attempts at Cornerstone to work on this too from, from a different perspective.
Speaker 1 00:11:07 Well, the issue is the use cases, what use cases have the vendors built because just because it has AI algorithms in it and it has a lot of data, that doesn't mean it's designed and optimized with the user experience for the thing you want to do. Because the issue of building a development plan, of doing a skills analysis in detail and analyzing and and aggregating and refining and simplifying your skills architecture, the issue of determining the skills you have in leadership, for example, pay equity analysis, location analysis, and competitive analysis, those are applications of talent intelligence. These systems won't necessarily give you that kind of functionality out of the box unless you're a software engineer and most of them don't want you going after the core APIs. So the real key thing with these vendors is to talk to them about which of these use cases they've successfully built and get access to references of companies doing them.
Speaker 1 00:12:08 Because what I've learned about this space is that the talent intelligence vendors have essentially three big r and d areas they have to invest in. First is they have to get the data. That's not easy. They have to find it, they have to decide where they're gonna locate it from. There are many, many sources of it, they need to clean it up, they aggregate it, they need to put it into a, the appropriate AI based system and they have to keep doing that over time. And we'll talk more about the data market later. Second, they have to build the models in the intelligence system that predict or characterize or define the segments of the database you're looking for. So if I say I want a retention analysis out of my talent intelligence system, we need to build a model around retention. We need to define what high retention looks like, and then teach the system how high retention applies to other people in the system.
Speaker 1 00:13:03 If we wanna build a system for candidate interest. So we wanna know of all these 2 billion people out there, which ones are even interested in our company, not forget about whether they should work here or not, are they even of interest and would they even be successful here? That's another model. If you wanna build a model for how well people could be promoted into leadership, that's another model. If you wanna build a model around how high people perform, you'd have to put performance data in their performance management data. And by the way, these systems also have time series capabilities. So in order for you to really do this kind of analysis, what you'll learn is you're gonna have to slide forward and backward in time. You're gonna have to look at this person's job and role and this this community or group today versus, what was it a year ago?
Speaker 1 00:13:46 What was it five years ago? Now that's extremely useful because once you have, but that's another model. So, so these are very, very high functional, high value application systems that can be used for many, many things. I told you one that I've been working with the most. We use Eightfold for our global workforce intelligence project, who we've been very, very familiar with it and we are quite impressed with the company, but these other companies are great companies too. So that's a little bit about what talent intelligence is all about. Oh, oh, the third level by the way that the application vendors have to build is the front end. They have to build a useful screen or series of tools or diagnostic tools or reporting tools or drill down tools to allow you to see what's going on inside of this database. That's not easy either.
Speaker 1 00:14:32 When you start looking at these intelligence applications, there are so many buttons to push that the design of the user experience is sometimes what sells or doesn't sell the product, not the actual strength of the ai. In fact, a lot of the easy to use user experiences are not very sophisticated under the covers. So they're kind of fun to play with and then you suddenly realize the information you're getting back isn't that useful. So, so that's a little bit about the talent intelligence space. It is growing like crazy once SAP's talent intelligence hub comes out and that will be add more fuel to the fire. And now that we have lots of engineers trained on ai, I think there will be more implications of this. I would expect that the big recruiting vendors like I Sims and some of the other ones, perhaps Avature at some point and others that have access to large amount of recruiting data and have been throwing it into ai, try to figure out what skills people have will, you know, likely go down this path over time.
Speaker 1 00:15:29 But they didn't start there. So I hope that's helpful. As a start, so you understand the tele intelligence market, I will be talking about it in the uh, HR cap conference. One more thing. The issue that a lot of companies have is where does this go? Is this a recruiting tool? Is this a workforce planning tool? Is this a career tool? Is this a tool for, uh, segmentation and other forms of employee analysis? Yes, it's all of those things. And so my hope and prediction is that the people that will do a lot of this talent intelligence will be the group that is currently named People Analytics because the people analytics staff are the ones with the statistical backgrounds and they're learning about AI every day that will probably be asked to use the tool for these kinds of analyses. Now, the recruiters and the recruiting department and the internal mobility and the l and d groups will use it off the shelf for other applications, but they'll probably be the ones that really use the data the most. So we're really high on this. I will send you in the podcast a link to the Talent Intelligence primer, which is a pretty good overview of this. And I think it's important you learn it because it's not exactly generative ai, but it is a part of the AI landscape in HR that's extremely, extremely important in growing very, very fast. Thank you.
