Job Analysis for the AI Era

[00:00:00] So job analysis, it doesn't exactly light people up, but with AI reshaping roles and creating opportunity for new roles, it could become a really important HR skill. And I'm Andrea Adams and today on the HR Hub, I'm here with Edie Goldberg. Edie's the founder of E.L. Goldberg & Associates.

[00:00:23] She holds a PhD in industrial psychology. She's a globally recognized expert in talent management and the future of work and has over 25 years of experience designing HR systems that actually connect what the organization is trying to accomplish with how they operate. Hi, Edie, how are you? Good morning or good afternoon. I'm doing great. Thank you for having me here.

[00:00:50] Yeah, well, thanks for coming on. So we're going to talk about job analysis. And I know, you know, it sounds dry. I mean, I think actually a lot of people are going to be interested in it, but it sounds dry. So remind us why we should care about job analysis. Well, job analysis is really just the investigative process of studying information about a specific job.

[00:01:18] And it's important because it forms the foundation of information of almost anything that you can do in HR from selection or development, training, performance management, and job redesign as we think about AI. So it's really important because it's kind of how we understand the job. Mm-hmm. Okay. So I was going to ask this question later, but I'm going to ask it right now.

[00:01:46] How is job analysis different or the same as the process of creating a job description? Well, job analysis is the process of investigating the job. The job description summarizes the finding from the job analysis. So the job analysis is the research and the job description is the result.

[00:02:13] It does seem, though, like when I think job analysis, I'm thinking a real effort versus sitting down with a client for 30 minutes, which I like literally do. And, you know, tell me about this job and then I extrapolate from what I know about that title and poof, I create a job description. It just doesn't seem like the same process.

[00:02:35] I think a lot of job descriptions are fairly flawed in that a lot of things end up being missing. But more often than not, you know, a job description is created and it's kind of like set in stone. And then the world changes and the job changes like five minutes later and that job description becomes obsolete. So you're right.

[00:03:02] There is much more analytical work that goes into a job analysis than a job description is really just that summary. But a job analysis is really that backbone that, you know, makes you confident that the information you have is correct. Okay.

[00:03:24] I want to ask questions about, like you just talked about the pace of change and, you know, in job analysis being more than, you know, a half hour effort to create a job description. But I will get to that because first one I want to understand is what is the process of job analysis? What does that look like? Like, what are the steps? Okay.

[00:03:45] So I'm going to really talk much more, if you will, from an academic perspective of, like, you know, what are the recommended steps if you're going, you know, I have a background where I've done a lot of work where there has been litigation or concern about litigation. So, you know, the most legally defensible approach is the right approach. And that's where, you know, job analysis comes in.

[00:04:15] So a true job analysis kind of starts with looking at what do we have? What do we know about this job? Right. So what documentation exists? Is there a job description? Are there training documents? Where does this role fit within an org chart? Or, you know, what do we know about that position? Just kind of what's today's state? What do we know? Then you gather information from subject matter experts.

[00:04:44] Now, usually subject matter experts are the people who are doing the job, right? Because they know the most about the job. But it could be a manager who, you know, has the expectations of what people are doing on their job. I think the best approach is doing observation first. Because so often, if you just ask somebody, what are you doing your job?

[00:05:12] You're going to forget a whole bunch of things because they're just second nature. And you're just not thinking about those daily things that you do. So when I can observe you, you know, I had fun earlier in my career. I did all the entry level and promotional exams for the San Francisco police and fire departments. So I got to go on ride-alongs, right? So I got to learn what do firefighters do?

[00:05:41] What do police officers do? And by actually observing them in the field in the moment. Then with that information, by observing, I can identify a series of tasks that people perform or tools that they use or challenges that they come up with. You can also ask people about those things.

[00:06:05] But I think observation really, you know, results in the richest amount of information. I would think so, too. And then just thinking about that, like, police officer example. Like, and I'm imagining things, but because, like, I've never done a ride-along. But I imagine that a police officer job is, like, super exciting and dreadfully boring. Yep.

[00:06:34] And somehow you have to be able to capture both of those. Yeah. So once you've done the observations and or interviews, and I like the combination of both, you know, so I can have my observations and then I can do some interviews. You know, what do you do? Ask some more questions. Then, and this is a really important part of a job analysis.

[00:07:01] You create what's called a job analysis questionnaire that has all the tasks that you've identified through observation or interview, as well as kind of the knowledge that you need, the skills, abilities that you need to do the job. And that questionnaire, you can ask, you know, to your point, how important is this task? Okay.

[00:07:30] How frequently is this task performed? Right. Do you need this knowledge day one on the job or can you be trained to do it? So let's talk about the task importance and frequency. As you were pointing out, there are some things that I might do as a police officer or firefighter every day. That's a pain that happens really frequently, super unimportant to success in the job.

[00:07:56] But there are some things that are done infrequently that are super important to success in the job, like arresting a criminal or putting out a fire. Done really frequently, but the most critical and, if you will, in this case, life-changing aspects of the job.

[00:08:18] So we want to be able to identify those things that are either really frequently done or infrequently done, but super important to success. So, you know, you get a completed job analysis questionnaire by a bunch of incumbents on the job. And then analyze that to say, you know, what are the most critical tasks? What are the required, you know, knowledge, skills, competencies?

[00:08:45] Then you can develop your job description or, you know, an assessment to select people into that role or a development program to, you know, support developing those people. Or you can understand that role so that you can reshape it in the world of AI. Oh, which we'll get to.

[00:09:13] You talked about that a proper job analysis is more legally defensible. It sounds like to me that maybe you've actually been in court over something like this. In which circumstances does a job analysis become a legal document? Well, in every circumstance.

[00:09:33] I mean, if somebody questions how you're hiring people or questions the fairness of the development process, the job analysis is a critical part that would form, you know, the basis of legal defensibility. Okay. It's a piece of it. I mean, obviously how the, you know, assessment is put together, et cetera, is super critical.

[00:10:01] I'm going to be totally honest. Yeah. I don't know that I have ever seen job analysis done except in maybe like almost decades ago in like a union environment where they were trying to put in BFORs and they wanted, and then in that case wanted them to be legally defensible. Yeah.

[00:10:29] But like it hardly ever happens properly. Like lots of job description writing like I described it. Why aren't we doing job analysis more often? Should we be doing job analysis more often properly? Well, I think we should because it obviously leads to better information to help you, you know, create the supporting information that you need to do whatever you're going to do with that job. So it is really critical.

[00:10:58] And I think I was just actually on a webinar the other day with somebody at Salesforce talking about, you know, they're doing more analysis of jobs because they're trying to identify, you know, what parts of the job can be done by AI and then how do you reform. But there's a tradeoff to be made here, right? It takes time and resources to collect that information. You've got a lot of jobs in a company.

[00:11:26] So you do this for every job or obviously when I was talking about police officer and firefighter, these are jobs where a life is at stake. And it's very important to get the selection of these people right. And that also happens to be an environment where there happens to be a lot of lawsuits. So, yeah, the systems that they build is really, really important.

[00:11:52] But you'd think your, you know, average job, you know, across America and there probably is not a lot of job analysis going on because it's a cost benefit issue. It's that I'm going to invest this time, you know, in understanding. I will have a more, you know, so if you're in a litigious environment, you're more likely to do it.

[00:12:17] But I would also say if you're really trying to understand what happens in a job, who we work, you know, who are those other people that we work with? What are the connections? And today, I think there's a lot more interest in that because we're really starting to rethink what's possible at work. So now it's time to, like now it's time to talk about AI.

[00:12:44] Are you seeing a resurgence in the interest in job analysis because of AI? Like you just said that guy at Salesforce was talking about it. Yeah, yeah. And I will admit it's the first time in a long time I've heard somebody talk about job analysis. Really?

[00:13:04] So, so I don't, you know, I haven't seen a survey of, you know, are more people doing job analysis because of AI? However, it is, and it's probably not a complete job analysis. It's probably much more focused on a task analysis and not so much the skills, abilities, et cetera, and competencies needed. Mm-hmm. Successful.

[00:13:33] But trying to understand what are the tasks people are performing and whether or not AI can help with those tasks. So I do think that there's more interest in the task analysis piece. There are a lot of companies out there that now have technology that can, you know, look at the tasks in your job, which maybe come from your job description.

[00:14:01] And, you know, and then make an assessment about how those can be, you know, where AI could help facilitate in that role. So I do think that there is a growing interest in the, at least in the task analysis part of job analysis. Mm-hmm. And when you say task analysis, do you mean the, like, full scope of tasks that an individual does?

[00:14:30] Are you just looking at how one particular task gets done? No, I'm talking about, you know, so, so if you take, you know, financial analyst. Yeah. What are the different things that that person does in their role and where might AI help them in that process? And then, you know, then you can, you know, I've done a lot of this kind of work and job redesign.

[00:14:54] But, and then say, okay, well, if these, you know, five tasks are really better augmented with AI and can reduce the, you know, amount of time you spend in those tasks by 40%, then, you know, where do we shift your time spent so that you, if you will, upskill or upgrade that job to deliver more value?

[00:15:24] And that's the work that we're seeing a lot of today where people want to understand how do we, how do we get rid of the toil and bring more joy to work?

[00:15:36] How do we get rid of the stuff that's just routine and just takes time and allow people to bring their human selves to work and do the work that they are uniquely suited to do, which adds more value? Yeah. That word toil, it was my favorite word for a little while. Had a, had a, I think that I came out of an interview I did with Indeed actually.

[00:16:06] About how they were using AI to augment the platform. And it was about finding, and internally, it was about finding the moments of toil and seeing if AI could help. I'll give credit where credit is due. The first time I heard that was actually from Kathleen Hogan, who's a CHRO at Microsoft. And she talked about removing the toil and bringing more joy. Credit to Kathleen. Well, joy.

[00:16:34] Okay, Indeed didn't talk about joy, but they did talk about toil. You know, and the reason I got interested again in job analysis was because of AI. So you say that you haven't had many questions about it, but maybe I'm your first question because I've been thinking about this. Yeah.

[00:16:59] Is, now is, is, so we're looking mostly, I guess, at the task part of it, but is it impractical when, like, AI is changing so quickly? So, you know, initially in the first, you know, once large language models got, you know, some momentum that, you know, things are changing every year or so. Now I hear it's changing every six months.

[00:17:25] Is it even practical to do job analysis when things are changing so fast? Well, that, that is a really good question. You know, and that, that's often my issue with a job description, you know, they just become out of date so quickly. Okay. You know, the question is, how are you analyzing the job? And the need for speed is very, very important.

[00:17:54] And obviously, you know, I'm going to be looking at jobs where there are a lot of incumbents, where I can really make a substantial change in my company. Mm-hmm. And, you know, new jobs are just being created all the time. So there's a, there's a difference between my job is changing because I'm augmenting what I'm doing with AI. Mm-hmm.

[00:18:23] And because of AI, we are creating entirely new jobs. I mean, you know, who, you know, previously had the job, you know, of the, you know, AI quality, you know, checker. Right, right. Making sure that, you know, it doesn't hallucinate. Mm-hmm. So there are, there are new roles being created as a consequence of AI.

[00:18:53] Mm-hmm. And on those new roles, we can't do job analysis because there's nobody to talk to. That's correct. Well, yeah. So this is going to date me. In 1998, I actually published a journal article on doing future-oriented job analysis. Oh. This is an international journal of selection and assessment.

[00:19:19] But we were already thinking back then of, as work changes and you create new positions, how do you, how do you do a job analysis for a job that doesn't yet exist? Mm-hmm. There are no incumbents to interview, right? Mm-hmm. And so you, you have to do it a little differently.

[00:19:41] You have to, you know, rely on the people who have the vision for the future of what this job is intended to do. Mm-hmm.

[00:20:19] So it's, it's possible. But it's certainly a different process than a traditional job analysis. The world of work is changing. Mm-hmm. And we need, we need to understand how to optimize the work so that humans get to do, you know, kind of what brings their best foot forward, if you will.

[00:20:44] Leverages their skills to the best, you know, of their abilities, you know, for the company's benefit. Mm-hmm. And, and so we really have to understand the roles and where AI can play, where AI shouldn't play. There are a lot of things AI can do, but should it, you know? So there, you know, certainly a lot of, I think your audience is largely a human resources audience. Mm-hmm.

[00:21:13] You know, should, should AI do employee, you know, take employee relations calls? No. Bad idea. It could. That would be funny. I'm sure though, man, I do quite a bit of employee relations. And, and do I ever see a lot of, when you start getting into some of the back and forth, like AI informed advice that various people are getting in the process. So. Yeah.

[00:21:41] I mean, I, I, and certainly with the rise of coaching, right? You know, and AI and coaching is really becoming a big thing. And so you can see the related, like, we could use it for, you know, employee relations issues. But that's just a bad idea because, you know, human empathy is really, really important.

[00:22:04] And, and that doesn't mean AI enabled isn't a good idea because sometimes people don't have the right words and AI can help us, you know, frame something in a way that will, that will help somebody more. So I'm a big fan of AI enabled as opposed to AI done. Mm-hmm.

[00:22:31] So we have to, we have to think wisely about, you know, how we use it. Yeah. If someone wants to learn about how that, to do job analysis, like they can see the demand rising, even if it's just the task part of job analysis. So few people able to do this anymore. How do you suggest we get the skill? Okay. Okay.

[00:22:59] Well, you, you did, in my introduction, you said I have a PhD in industrial organizational psychology. So, you know, it'd be the classic textbook, you know, in job and work analysis. Okay. There's a textbook. It's called Job and Work Analysis, Methods, Research and Applications in Human Resource Management. Mm-hmm. It's by Frederick Morgensen, Michael Brannock and Edward Levine. Ed Levine. Mm-hmm.

[00:23:27] And that's, it's just kind of a classic textbook that takes you through the steps. Mm-hmm. You know, I'm talking about dry academics, but it is, you know, the how to. Yep. How to do it. Yeah. Okay. I'll put a link to the textbook because I think there's so many things in HR that, you know, the textbooks are useful, but only so far because the nuance is so important.

[00:23:52] But I think in job analysis, perhaps that is something we could, you know, find a leaping off point from in a textbook. Um, what, uh, like you're a forward thinking, you know, person thinking a lot about the future. What are you thinking about more broadly about HR these days? And where could people go to find you and follow you? Okay. So what am I thinking about these days? That's always an exciting question, right?

[00:24:21] Uh, obviously in the world of AI, I'm really interested in how work gets done, um, and, and who does what. And, you know, I think the problem and how, the reason we haven't, you know, gotten ROI from AI is because we're just trying to do things faster as opposed to do things differently. And so it kind of, it's very related to the, you know, this topic.

[00:24:51] Um, I think we really need to understand how we truly transform the work that's being done to help people work differently for better outcomes. And that I think is the power of AI.

[00:25:06] So I'm very excited about, you know, looking at changing workflow process, um, how teams are collaborating to get work done, um, and leveraging AI, you know, to augment the tasks that are being done to essentially, you know, put things on fire and help people to do more and to do better work. Um, so I'm, I'm very keen on this for the future.

[00:25:35] I think this is going to be positive. I think we're going to create better work, uh, for people. Where can people find you? You can find me very easily on LinkedIn. Um, follow me there. Uh, you can also just, you know, Edie Goldberg, um, IE Goldberg is a pretty unique name. So you can just Google me and you'll find places I've spoken, articles I've written, podcasts I've done, et cetera. Uh, so I'm pretty easy to find.

[00:26:04] Well, thanks Edie. That was great. I appreciated the, um, both like the way you presented the basics of job analysis, but then connected it to the sophistication and change that we're seeing today. But that's the end of the show today. Thanks for watching out there. We'll see you next time.