AUA: Should HR Rewrite Job Descriptions for AI?
About this episode:
Original Air Date: April 26, 2026Every HR team is getting the same ask right now: rewrite our job descriptions to reflect AI. It sounds reasonable—until you realize you're being asked to update a document that was already a little broken for a world that's changing faster than any static artifact can keep up with. So where do you even start? And is the job description itself actually the right place to begin?
In this mini episode, Rodney and Sam flip the question entirely—arguing that the smarter move is to start with what AI can actually do in your organization, define the human's role in relation to that, and why living, dynamic approaches to role clarity are more essential now than ever.
Mentioned in this episode:
“traditional consulting”: AWWTR Ep. 8
"talent marketplace”: FOHR Ep. 7, BNW Ep. 160 + 161
In this episode:
Rodney Evans
Sam Spurlin
Listen now:
Episode transcript:
Rodney Evans:
Hey y'all, welcome back to At Work With The Ready. I'm Rodney Evans and that guy is Sam Spurlin.
Sam Spurlin:
Hello everyone and hello to you, Rodney Evans.
Rodney Evans:
Hi, Sam. Every other week we tackle one tough thought-provoking listener question and share a few ideas that might help. So let's dive on in. Sam, what do you have for us today?
Rewriting job descriptions should come last, not at the start
Sam Spurlin:
You've talked a few times on the show about how people's day-to-day work is going to look drastically different as AI gets integrated, but I haven't heard much about what that actually looks like in practice. My HR team is getting pressure to start rewriting job descriptions to reflect AI, and I'm not sure where to begin. How do you start that work without feeling like you're building something that's just going to get washed away in 18 months?
Rodney Evans:
God, you guys, it's such a cluster. I've had two different trad consultancies call recently and been like, basically, the TLDR is basically like, "We're aware that our business is drying up. We think we have to do what you know how to do. Do you want to be our partner?" And I'm like, "Get out of here. Fuck no, I don't want to be your partner." Anyway, that's a side note. And the presenting issue is basically this. It's basically like, we used to just tell people what the new org chart was and what all the job descriptions were, and now it seems like our clients aren't going to buy that anymore because they could just get it from ChatGPT. Should we go into your business? And I'm like, I don't care what you do. Anyway, that's a little aside.
Sam Spurlin:
Rodney's therapy session.
What to define instead: authority and human-AI interactions
Rodney Evans:
Yeah, totally. With myself. Here's what I've been saying to clients lately about AI is what's happening now on balance, and this question is another ride on the Ferris wheel, is people are looking at existing work and existing problems and figuring out how AI can solve it. It's like, "How do I jam AI into this to make it better or easier or go away or automated or cheaper or faster?" whatever. The more interesting position to take is what can AI do right now? And what does that mean in terms of this work? And so when you talk about "rewriting job descriptions to reflect AI," I'm like, "Yeah, don't do that. Start with, what can AI currently do in your organization? What problems can it solve?" And then what probably needs to be written down or codified is what's the interaction with the human? Okay, so if AI can create all of the content for your social media, what's the description of the pairing between the social media manager and the AI that you're using so that there's a human in the loop, there are certain guardrails, there's the kind of review that you want.
At this point, I'm defaulting to don't worry about job descriptions. I mean, sort of never worry about job descriptions. That's been a fool's errand for a long time. But even in terms of role clarity, I think right now, really light and really emphasizing the authority piece of it in terms of what can't be held by AI is kind of the only thing you need. I wouldn't spend a lot more time on this pursuit than that. So it's like, if you're talking about engineers and you expect them to be 10xing their output using Claude Code, great. Their accountability is that. What's the decision rate so that they're not like pushing buggy or borked or unsafe code? That's like the level for this liminal period to be working at. And again, I would start with the question of what can AI do and then what constraint does the human need to apply rather than the other way around?
Why articulating human judgement matters more than ever
Sam Spurlin:
Yeah. I agree with that. I think the only little spin that I would offer is I think also articulating outcomes that roles are working toward continues to be a valuable thing because when you're doing that, you're not getting into the weeds about is a human doing it, isn't AI doing it? It's what are we actually trying to accomplish here? And we can figure out as the AI landscape changes, how that actually gets pursued. And I think the newer thing that I haven't necessarily had to do pre AI is actually start to articulate what is the judgment, the human judgment we expect this role to be exercising. And that for obvious reasons is not really a thing we really had to define before-
Rodney Evans:
Right.
Job descriptions should become living documents
Sam Spurlin:
—But now I think there's something interesting to that. So if this role is massively changed because of AI, but we still want a human doing some version of it, well, what's the judgment aspect that the human is bringing to it and can we get a little bit crisper about that? And then the kind of meta point is I would use this opportunity, not just in this job description exercise, but just kind of broadly to see if you can use it to shift toward more living documents that are dynamically changing rather than kind of exercises and creating static artifacts.
Rodney Evans:
Totally.
Sam Spurlin:
And we've always been advocating for creating living documents, but I think it becomes even more essential now when everything is changing as quickly as it is.
An AI-first lens on a talent marketplace
Rodney Evans:
I think that's such a great point, Sam. And the first use case that I had for AI that I waited for like a year for it to be able to do was basically like talent marketplace management. And for like six months, I was deep diving into all of the tools out there that said they could do that and being chronically disappointed by them. And now AI actually does make that possible. It makes the thing you're talking about possible where like work can be fractional, work can be atomized and you can sort of have a network that is continually updating as people's roles change and their partnership with AI changes and AI's capabilities change. And so again, I love what you're saying and like I would really take the AI first lens rather than like, "We document job descriptions in this way. How do we do it taking AI into consideration?" It's like, how should jobs work in your company with AI? Start there and then see what shakes out.
Sam Spurlin:
All right. That is it for this mini. If you've got a question of your own, hit us up at podcast [at] theready [dot] com.
Rodney Evans:
We will see you back next week for a full episode of At Work With The Ready. Thank you for being a listener.