Subscribe
Share
Share
Embed
AI is reshaping leadership, strategy, and the very role of HR. In each episode, the host Barb Bidan explores how AI drives innovation and leadership in HR with actionable insights for the future of work.
The Human Element is brought to you by Harper, Wisq’s always-on AI HR Generalist transforming how work gets done. Powered by deep HR intelligence, Harper delivers instant, accurate, and empathetic support—from policy questions to performance coaching—so your people get answers fast and your teams stay focused.
Learn more at Wisq.com.
Q Hamirani & Barb Bidan | May 15
===
[00:00:00] Welcome to The Human Element, presented by Wisq. I'm your host, Barb Bidan, and in each episode, I sit down with CHROs and senior HR leaders to explore how AI, innovation, and human insight are reshaping the future of HR. We'll explore how technology is reshaping leadership, strategy, and the role of HR by sharing candid stories, practical ideas, and strategic perspectives to help you shape the future.
The Human Element is brought to you by Wisq, the leader in agentic HR and creator of Harper, the world's first AI HR generalist. Learn how Harper can resolve up to eighty percent of routine HR tasks autonomously. Learn more about Wisq at wisq.com
Speaker 4: My guest today is Q Hamirani, the chief people officer at HighLevel. Q is a builder and an experimenter at heart. He spent nearly five years at Airbnb as their first global people operations leader, where [00:01:00] he incubated and designed the live at-- live and work anywhere program. He also guided the company through hypergrowth, a pandemic, workforce restructuring, and an IPO, so just a few really major things.
He went on to serve as the chief people and communications officer at Paper the Series D edtech company backed by SoftBank, SoftBank and IVP. And today, Q leads the people function at, at HighLevel, a two thousand plus person, fully remote AI company that is growing at a remarkable pace. He has also founded numerous companies, is a Forbes contributor, guest faculty at the London School of Business, and a repeat speaker at some of the biggest stages in HR, including Transform where you were just on stage for a panel called Digital Teammates: Where
Speaker 3: Mm-hmm.
Speaker 4: Fit on the Org Chart.
So I think we're gonna dig into
Speaker 2: Yep
Speaker 4: it was a great Transform this year. So that's exactly where we're gonna go today Q. So welcome to The Human Element, and thanks for joining me.
Speaker 2: Yeah, thanks for having me. Excited to be here
Speaker 4: Absolutely. So, I mean, we-- I started your [00:02:00] introduction by talking about all the great things that you've built. So people functions at scale across some really recognizable names, right? In tech, edtech, Airbnb, Paper, HighLevel now. Tell us more about your career arc. What drew you into the work and how has your thinking about people strategy evolved as AI has entered the picture?
Speaker 3: Yeah. First off, thanks for the kind intro. Um, in terms of my career arc, I think I really backed into the people function from an engineering mindset. What I mean from that is, so I studied electrical engineering, uh, way back in, in college, and I've never, um, worked in engineering, but I've applied all the things that I've learnt in terms of dealing with ambiguity, systems thinking, first principles, really questioning the why on everything I do.
And the first seven years in my career, in terms of the arc, was not in the people function. I started off working for CFOs and CEOs doing digital workforce transformations then worked in some global strategy and ops in-house, then ran three of my own [00:03:00] companies. And that's-- At that point, I realized there was only one key to success and one key to failure every single time, and it was the people, right?
And to me, coming from an engineering mindset as well, I found that the hardest thing in an organization is not necessarily the code base or the product, it's the people Right. The people can change the way work gets done, and that can impact your best product or your best process or your best technology.
So I set out mid-career to work in the people function, and I've been doing that for fifteen years now, and it's really been grounded in the arc of how do we-- how-- why do we do what we do, and how do we kind of shift from the mindset of making policies to reducing friction so that the operating system of how work gets done can be more fluid because that's when you see people being their best and working together and helping navigate through that.
So it's been, it's been a fun journey, I think, especially, uh, in the last [00:04:00] six months or even twenty-four months or last few years because technology through AI has brought it full circle for me, right? So when I started in engineering, and then I was-- kind of got into HR technology and analytics maybe fourteen, fifteen years ago, the world was a very different place in terms of the skills and to be able to build and code, right?
I, I did not do engineering through, through my career, so I'd lost touch with how to code. And honestly, uh, one of the, one of the interesting fun, fun facts is I actually did electrical engineering back in college because I did not like to code. I did not like to learn syntax of C++ Java. I just wanted to build stuff, break stuff, try to make it better and do things.
And now with AI, all of a sudden, twenty-five years later, I feel like I can code and I can build without having to have gone down the path of understanding and learning a lot of syntax because of the way vibe coding and all that stuff is now. So it's really interesting because I think us as people leaders, including myself [00:05:00] now, are living in a world where we have the ability to build and prototype.
Production ready at scale is a different kind of ballgame, and there's, there's obviously some variables there. But we have the ability to build, be creative, build different modalities, be it voice to speech, to anything in a way that we could not do before. And what that does to us is it gives us almost superpowers in a way to enable employees and our organization to be their best, and in turn, go back to what I started off with, is to operate in a more fluid, dynamic, and trusting environment so that they can all do their great work.
Because at the end of the day, we're here to just make sure work gets done. Um, and in a world where we are today we can, we can enable that with not only less need for dependencies, be it resources, but also, uh, the extra modalities of being creative and hitting, you know, and everyone in a personalized way 'cause there's no one size fits all as we know in our, in our career, and you [00:06:00] can never please everyone, but you have to do your best to try to resonate and relate with everyone.
So that's a little bit about full circle, I would say, on the technology front, um, for myself at least.
Speaker 4: Ve- very full circle. Sounds like you picked the right type of engineering back in college, right? So you didn't wanna understand syntax, but
Speaker 3: Yeah
Speaker 4: did learn to understand systems, right? In
Speaker 3: Exactly
Speaker 4: degree and like sort of think of the, the whole, the whole system or the system as a whole,
Speaker 2: Yeah
Speaker 4: like what you're, what you're talking about there.
So I
Speaker 2: Yeah.
Speaker 4: this year. You were at
Speaker 2: That's right. Yeah
Speaker 4: You were just on on stage doing the digital teammates where AI agents fit into the org chart. So that framing, I,
Speaker 2: Yeah
Speaker 4: interested in, right? A teammate, not a tool, is doing a lot of the work. Talk to me about what that actually means to give an AI agent a role inside of an organization.
Speaker 3: Yeah, I think more recently in the last 12 months with AI helping us do work in an agentic form, meaning we have agents that can do work beside us I think it's [00:07:00] changed the game of thinking of AI more as a tool that you just kind of implement and you know, install, for lack of better words, and, you know, use, versus more being a teammate that helps you with decision-making along the way, right?
So even when that construct changes or, or is additive, because you still have AI and technology that can just go do work for you, right? There's automations that, you know, no one likes posting a requisition in whichever tool you use, right? If you could automate that, that's great. So there's still the automation part of it, but now there is an additive part of it, which is: How can it be part of the decision-making with you?
So a simple example is if you're doing analytics and you're analyzing data now, you know, you don't need necessarily a whole team dedicated just to doing analytics, absorbing all the data, 'cause you can instantly almost do it, um, with the right tool set. That analytics is gonna help you with almost real-- relatively more real-time decision-making to, to build and iterate and [00:08:00] figure out what resonates and what does not.
So we have to think of now AI or technology in this wave as being both a teammate and doing work for us. So one of the things we, we talk about is how do we-- how can AI do work for you? Example, things like automations and things that you just don't... You wanna, you wanna test it, give it guardrails, but then just let it do it for you.
And then how can AI work with you, right? So when you think of that, the construct of a teammate, the construct of how do you onboard a teammate now that's a digital teammate, right? And I think the part that I've kind of continued to focus on is even if you have AI as a teammate you-- the human is the one that needs to be accountable and responsible for that teammate.
What, what am-- what do I mean by that? When we, when we-- Thinking of AI as a teammate, we need to think of, you know, in a loose form, what is the job description? I'm not saying we need to go write the job description and, you know, do the whole nine yards, but [00:09:00] we have to think of, okay, in this context, what am I expecting AI to help me with?
How am I gonna hold the AI accountable? Because at the end of the day, if the AI hallucinates or does something off, it's m- it's the human-- it should be the human's responsibility to make those critical decisions, whether they're using the, um, data that they're getting or how they're interpreting it or what the context or the heart is to that context when you see data about people.
So I think it's really, uh, a new, uh, difference now or a new wave, I would say, uh, which is: How do we think of how it sits into the orchestration of the team, the work, the responsibility, and how do you onboard them? It's not an installation button anymore. Not that it was, you know, but just to make a point and oversimplify it.
It's now more how is it going to be there and grow? And the, the beauty of this technology is it learns with you It stores memory. It's, it's reminding you of things that you may have asked or done six months ago in a [00:10:00] context that's really, really powerful. So yeah, it's more a digital teammate in addition to being kind of doing the work in a automation form.
This is an added dimension that changes the construct to how we need to look at it as well and leverage and use it
Speaker 4: Completely. I love the idea of like the teammate working with you or for you versus a tool, right? Which in some way is like alongside you, separate and apart in a different way than what you're do- than what I hear you describing
Speaker 2: e-e-exactly. Exactly. It's-- Yeah, totally
Speaker 4: So when you, how do you, when you think about how you define the scope of what a digital teammate owns versus what it can assist with, how are you thinking about differentiating and where, and where maybe to in- insert the, the human operator in that process?
Speaker 3: Yeah. So I think there's a couple things. I mean, my general rule is pretty simple. There always needs to be a human owner of the outcome that's being produced by the mix of technology or the [00:11:00] mix of human in the loop. Meaning you know, you cannot just blame the AI agent, right? So when I think of it, I think of what are we trying to solve for?
Be it you know, may-maybe it's hy-- we're in hypergrowth, and we're solving for hiring as quickly as we can in a very competitive market, and maybe one of the bottlenecks could be We need real-time compensation data, or we need to figure out how to, how to get to market quicker and stay, stay on that leading edge because, um, sometimes you're defining the market data versus getting it on a lag, right?
So when I look at the outcome, I first try to think of what, what do we need help with and what do we not, um, kind of necessarily need help with. And, and this has come from kind of experience on-- I've gone down the path, uh, numerous times, especially when in twenty twenty-three or twenty twenty-four when AI was kind of new and it was growing pretty quick, where I would try to build an AI solution for something only to realize I did not even need it.[00:12:00]
I did not even need a, a solution necessarily if I just had the right judgments put in the loop of the process. Um, so I say, uh, I mean, I would say, like, I really think of it on: How do you, uh, break up the-- Look at the outcome. Look at the outcome, try to see what is inhibiting us from getting to that outcome to improve that fluidity and dynamic nature of operational work, work getting done.
What is, what is slowing us down with getting work done? And how-- And what is-- I, I often tell uh, some of my teammates: "What is the worst part of your day? Like, what do you hate doing?" Right? That's a, that's a good easy win to start off with. Um, either what do you hate about it? Like, maybe it's something you hate doing, or it's something you hate waiting on, uh, because you have a dependency in order to get your work done.
And then I think if you focus on the outcome, you can orchestrate kind of that mix. And I think the biggest thing in, in, in this world that we are [00:13:00] today is, you know, you have to have that agility to pivot and realize when it's, like to my comment earlier, when it's really not even a use case for AI, but it thought-- you, you thought it may be, so there's no harm in trying it.
So really being able to pivot, um, having the agility to say, "Okay, I'm gonna try this." And, and the beauty of it today is I often run into this, which is I f- I feel like I can try things with AI that I would not have even thought of trying before. Because the sky has almost become the limit, which is good and bad.
There's, there's the, the good side of it, where we can kinda push the limits. We can try to do things that we could not do before. The dark side of that is, you know, it is a toll on the human mind and the human body, right? Because you can only do so much. The biggest capacity constraint we have today in a world of AI Is not the technology.
The technology's getting commoditized. Data is getting pretty much infinite, uh, almost like you can get everything at your fingertips. [00:14:00] The biggest capacity constraint is humans. It's our nervous system. It's how much can we take on from a cognitive, uh, perspective to make sure we can create this orchestration.
'Cause my firm belief has always been and will-- is continuing to be, which is that human-- what humans can do in terms of judgment, in terms of making that right orchestration layer to get the right outcome of work, can only be done by humans. So now the biggest bottleneck is us, right? And how do we make sure we balance that both for the outcome and for our own, you know, state of mind and not to race too quick, um, because in a world where the world and the, the ground underneath is moving so quick with new tools every month,
Speaker 4: Mm-hmm.
Speaker 3: modalities, it's tempting to move really quick, and that's what I did for the first year or two.
Now, it's actually what I found is the moat is really trying to keep the calm in that chaos and prioritizing [00:15:00] where I should try something and where I should not. And sometimes it's fifteen minutes of prototyping and scrapping. Sometimes it's, you know, longer. So it's, it's really, um, you know, that builder mindset.
It's almost-- It reminds me of sitting in a lab, you know, in college and trying to build circuits, blowing a resistor, blowing a capacitor, trying it a different way, and then saying, "Gosh, I got, I got the circuit to, to actually work," right? Um, and it's that iterative process, um, within, within bounds.
Speaker 4: Yeah, I mean, that's the fun, the fun within it. And I wanna go back to something you said a minute ago 'cause it really resonated with me. you
Speaker 3: Yeah
Speaker 4: you have tried some things with AI that you just wouldn't have thought to try before. Maybe you didn't think w-were possible or, or what. Can you elaborate on that?
Or,
Speaker 3: Yeah
Speaker 4: something that you've tried with AI that maybe just didn't even cross your mind to do before.
Speaker 3: Yeah. Yeah. I mean, I think one example is I've always wanted to be creative in terms of expressing something in [00:16:00] multi-modalities. And as much as I try to think of it music and stuff, I'm not, I'm not very talented to be artistic in that form. With AI, what we've done is-- So a simple example is, say you have, let's say, an engagement survey, a lot of information that you've gathered, including benchmarking.
You typically have the decks that you create. You work-- walk through it with everyone, you action plan. All that is good, and I'm not saying that needs to be scrapped, but now you have an additive layer that you can use with AI to, to reach people that may digest information in different forms. So an example is you could take all those multiple documents, put it in any tool.
Uh, one we used last year was NotebookLM because it had just come out then. And it created a six-minute podcast of two professionals talking about our company results relative to benchmarking, and you can share it with people, right? You can now spin up a quick website that basically will have the-- not [00:17:00] only the survey data, if I take that same example, but in an interactive format where people can drill down.
You're no longer dependent on your HR tech stack to communicate and drive adoption because we all know in a world where we all work in every day, there are just pain points with the tech stack because they're not built to, to resonate with every single persona and for every single module. They do their core job really well.
It's like the engine of the car. So how do-- You can now spin up apps, you can spin up websites. So multi-modality is what we've done, like just, you know, audio websites, you know, simulations of change management. So change management now, which is really the core part of our job, 'cause if you can drive change management really well, you can come back to what I always come back to, which is have work be done in a more fluid and dynamic way.
And if work is done more seamlessly in an environment where people have the underlying form of trust and understanding of why they're [00:18:00] doing something or why you rolled out a program, that probably won't make sense to them from an... It'll make sense to us as HR professionals, but not to them. I think you can just reach so much more people and drive the adoption.
So yeah, multi-modality is just one from, uh, ways we've tried it. And now I just, you know, kind of go to it. ~I've even made, um, an album, um, about our, um, kind of company, um, uh, and, and there, uh... Let me start this from the scratch.~
Speaker 4: ~Yeah, sure~
Speaker 3: Another good example is also I've created songs that incorporate our company values and who we are, and the entire thing from lyrics to production to post-production is all done with various AI tools.
And yes, I had the idea on how to do things, but in my past, I didn't have the ability to learn and integrate the tools. And this is another way, sound, audio you know, interacting apps. So yeah, it's definitely, uh, the sky's the limit, which is, which is exciting but can be nerve-wracking sometimes too.
Speaker 4: No, I mean, it's so, so fun, right? And it's like even even if you would've had the idea with that, the, some of those creative elements [00:19:00] previously, the time that it would've taken pre-AI, like just what-- it's not, like our teams aren't staffed to be able
Speaker 2: Exactly
Speaker 4: in in those ways. And I can feel your energy when you're talking about the like that level of experimentation.
For those of us who have been, you know, doing this work for like 20 years or so, not
Speaker 2: Yeah.
Speaker 4: need to be that long.
Speaker 2: Yeah
Speaker 4: that we have not been able to do, like that's why I think there's so much excitement, in particular with CHROs who are like in this stage of career. Like all the stuff that you wanted to do but couldn't always get to, like the sky really is the, is the limit.
So I feel you
Speaker 2: Yeah
Speaker 4: so I gotta, I'm gonna shift back to you talked about accountability and like who is accountable, the per- the human being the accountable party for the outcome. But one of the hardest parts of integrating AI into a people function might actually be accountability, right?
Who owns the outcome when an AI agent is involved in hiring, performance screening when it comes to hiring or employee [00:20:00] communication? So how do you set up that human level of accountability that you talked about before so things don't become a black hole where humans have no awareness?
Like I, I-- hiring comes to mind in particular here,
Speaker 3: Yeah. Yeah, and we can take hiring as that example to kind of build on. I think the fundamental blocks of accountability do not change in an AI or non-AI world. Meaning, if we have, let's just say, since we're taking hiring as an example, we have a recruiter that's in charge of a req or their manager, like you have the, the leaders in charge of hiring.
You know, if they use tools like, for scheduling or any, any of the AI tools internally, at the end of the day, when we're giving an offer to someone, they have spoken with hopefully more than one human along the way, and that judgment is on us, right? And hiring is an interesting one because Hiring is never foolproof, right?
Like, you, you gu- it's never... And, you know, I tell leaders that, [00:21:00] like, you can never guarantee it, but you can do your best to get data points along the way. So I think this is an example of even if we didn't have AI, the same person was responsible for it or the same role was responsible for it. Even if you have AI, and the AI may be doing work for that role, like let's say scheduling gets automated with AI tools, or the AI may be helping in the decision-making by doing work with the recruiter, in this example, or the hiring manager to provide more data points, the decision at the end of the day is with the, uh, with the human.
And that's how I look at every single process, right? Be it, you know, be it benefits or be it compensation. Wherever we're using tools, like we've used a good amount of AI to do compensation analysis because that is a very good use case for it. But one of the things I learned earlier when I was building out the model was, you know, the AI tool would just go off and do all the analysis and give me a number at the end of it, and I had to train it to say, [00:22:00] "Do not do that."
Like, "Stop." I had three stops in there to say, "Stop and validate," for example, "the internal leveling. Stop and validate the job code you picked from Radford or from the right database. Stop and validate that this is accurate for what you're looking for." And by building those, you can help the human be responsible.
So it's really that-- to, to your comment on the black, the black box I feel happens, and did happen early on when I was testing things, was because I let it do too much almost end to end. So if you do blocks of it, you make sure you're comfortable with that block in terms of interjecting the human in the loop or setting guardrails.
Like I would tell it, you know, right o- initially when things came up where you know, they could browse the web to give us information. That feels like decades ago, but I think that was like maybe a year or fifteen months ago, 'cause initially when ChatGPT and all these tools came out, they couldn't browse the web, and then they turned on the web.
You know, I learned when I was doing [00:23:00] some of my, uh, testing was that I wanted to turn that off because I did not-- I wanted to only reference my dataset and not the web. So as you build blocks and you refine them with the right guardrails and humans in the loop, then you can almost think of it as a Lego concept of you're building the blocks to a big process, and within those are baked the humans in the loop with the right accountability.
Because at the end of the day, like, AI is not completely a compliance problem. Yes, there's a compliance lens then-- lens to it that I do not wanna minimize AI is really a cultural problem. It's how people are doing work, and if they don't take accountability or want to take-- And if honestly, I'll go a step further to say, if folks are not taking accountability now for a particular process, my guess is they weren't taking accountability prior either.
Speaker 4: Right
Speaker 3: it's, it's really just highlighting the need for accountability because if things go wrong you know it, it has a bigger impact now because just blaming your AI agent is not an [00:24:00] acceptable, uh, form in-- from my lens at least.
Speaker 4: Completely. I think that I asked the question as the black hole, but as you were talking, I was going, it's really talking about having work not be a black box. And then, and then you said it. A bit of a tangent, but I-- as you were talking, my reflection was we've been talking a lot about how we're gonna bring sort of the the newest set of graduates out of college and teach
Speaker 2: Mm-hmm.
Speaker 4: have jobs if AI is doing everything?
And, and I believe that they will, but actually
Speaker 3: Mm-hmm.
Speaker 4: is another great way for this sort of incoming like entrance to the market, the, the newest college grads and so
Speaker 2: Mm-hmm. Yeah
Speaker 4: by not setting things up to be a black box, right? The AI
Speaker 2: Yep
Speaker 4: but you have stops along the way where you learn alongside of it what it was doing to get you to that, that point.
You validate, you step it
Speaker 2: Exactly
Speaker 4: your design if you don't wanna kinda, you know,
Speaker 2: Yeah
Speaker 4: the time along the way. But I think that's also gonna be pretty critical to how we, how we train folks and make sure that people [00:25:00] still know things, right?
Speaker 2: Yeah
Speaker 4: about like HR and the profession.
So,
Speaker 2: Totally.
Speaker 4: completely. So,
Speaker 2: Yeah
Speaker 4: the panel description that that was used at Transform was actually talking about establishing collaboration norms between humans and AI. That mean for you in practice? And how is describing like the interplay between humans and and AI different than just like writing a policy,
Speaker 3: Yeah. I mean, I think it's, it's kind of almost 2.0 of an SOP or a standard operating procedure because when you have, when you have AI doing things and you have the human in the loop one thing that, you know, we've, we've dealt with in the entire working kind of history is how do you deal with when folks transition or when new folks are getting onboarded?
How do they get integrated into learning that, you know, that c- that overall process? So for us it's, it's really having that level of almost RACI, roles and responsibilities and [00:26:00] accountabilities, all kind of built as though it, it's been done with, with collaborative digital and human teammates. And I think you just have to work through and document that so that onboarding becomes clearer.
And then you have to continuously learn and iterate. So, you know, as the tools get stronger, as you expand that box to become-- You know, I love the-- N-now that we're talking, like, I just started doing Legos with my five-year-old son, and it's very, very similar. Like you build, you build a piece of it, you work it, and then Legos come with an instruction manual that you have to like at least use directionally to go, um, in some direction.
So you need that instruction manual which tells you how you're w- orchestrating that job or that process. And then for our world, you know, it is, it is, it is got real repercussions on compliance, uh, be it data privacy or how we're using the data. It's got direct impact on people's lives if it's benefits or pay.
So we have to also make sure that, you know, [00:27:00] we are slowly and cautiously expanding that block in, in domains where we know we have flexibility, like the crea-- the example I talked about of creatively trying to express survey results with multi-modalities. You can be a lot more flexible and try with that versus payroll, for example, right?
But every process can be done if you have the right control. So I would say just kind of going back to the basics of making sure there is a collective process documentation that's being followed. And the beauty of it today is you can tell AI to create the draft. It's seventy percent there, and then you gotta tweak it, and you can tell it to remind you to maintain it or every time there's a change, go and update it.
So we have the tools, but this is kind of what I come back to. The orchestration of it and management of it needs judgment and context, which only humans, especially in our profession, can do best. So we have to make sure we empower the humans, hold them accountable, and give them grace Right? [00:28:00] I made mistakes when I was playing with the...
But it was in domains where I was being cognizant of not impacting people's lives. And you have to give yourself grace. You have to give your team grace. You have to give that experimentation sandbox environment. So be it give everyone the right AI tools that are kind of controlled in your enterprise settings, so you know they can put sensitive information and it'll be contained, so they can play with.
You have to help folks including yourself and your team, and give grace 'cause that is, again, something I think humans will be best at identifying and doing
Speaker 4: Yeah. So that's-- I mean, you've described what getting it right looks like, right? Some
Speaker 2: Mm-hmm.
Speaker 4: like governance around it, having the right infrastructure like allowing for learning within the system, right? Like failure and, and learning from
Speaker 3: Yeah
Speaker 4: What about getting... What does getting it wrong look like?
Like, what are the early warning signs do you think that AI integration might be complicating things within teams rather than enhancing it or, or other ways we might get it wrong?
Speaker 3: Yeah. So there's, there's kind of two thoughts [00:29:00] that come to mind on this. One, if you get it wrong, you lose credibility and trust very quick if you go too quick, right? So you have to be cautious on not just impacting, uh, people's lives if it's benefits or payroll, just as an example, but y- if it goes wrong, you're losing trust and credibility very quick, and then you've got to rebuild it, which takes longer, um, to do in, in general.
So I would say going the cost of going wrong can be very, very, um, dangerous. The, the second point I would say is the part that I think being the reason for why people might get it wrong is people are trying to implement AI without redesigning how the work or the process gets done, right? And it's, it's simple you know, you know, in other words, people are trying to bolt on things onto an existing process that may already be broken, right?
Speaker 4: And
Speaker 3: Or
Speaker 4: doesn't get better, like from
Speaker 3: why-- Yeah.
Speaker 4: not
Speaker 3: W-wondering
Speaker 4: Yeah
Speaker 3: why it does not get better or also [00:30:00] that is exposing the failure points, right? So then things go wrong. To your point, if things go wrong, it's easy to blame AI, but that may not be the root cause. You-- Because the way you've solved with AI is you're fixing a symptom and not a root cause to begin with, right?
So I think that's the biggest trap I, I feel like I fell into early on and I see people doing now is bolting on AI without really re-architecting the process, which means the people, the, the entire orchestration of that work. And then you don't make as much headway, you don't get the ROIs you wanted, and you look at it as a failure, or you reduce trust because you-- the process has just gone wrong.
So I think it, it can go wrong in various ways, but that should not stop us from exploring because if we don't, the flip side of that is we stagnate. And if we stagnate, the world is gonna shift so quickly around us that we will be left with poor judg-- poor inputs to make the [00:31:00] judgment that we are the best at doing as humans.
Speaker 4: Right. So like don't be f-afraid of it.
Speaker 3: Yeah
Speaker 4: to experiment, but a surefire symptom of something going wrong is if you are just finding that things are going wrong but faster, then maybe
Speaker 2: Yeah
Speaker 4: with the work, right? And correcting sort of like for what-- how should the work be before like slapping AI, an AI tool on top of
Speaker 2: Exactly. Exactly
Speaker 4: talked about that a ton in the show, right? Like going, starting with the work not, not like, and is the work happening in the right way? Like, is the process right? Do we do this
Speaker 2: Yeah
Speaker 4: from the beginning?
Speaker 2: ~That goes~
Speaker 4: ~subscribe to your newsletter when, uh, like when we, we need to create like something where we all just share the stuff we're experimenting~
Speaker 2: ~for sure. Yeah~
Speaker 4: ~like, uh, hit me up after and let me know where it exists because,~
Speaker 2: ~Yeah~
Speaker 4: ~I, I find myself like sharing onesie-twosie things with one or two people, and I, I~
Speaker 2: ~Yeah~
Speaker 4: ~thinking like, what if we all were just being like, "Have you thought of this?" Or, "Are you trying this? What's your~
Speaker 2: ~Yeah~
Speaker 4: ~'~ Cause it really is about experimentation. But I'm gonna pull us into the lightning round, so,
Speaker 2: Sure.
Speaker 4: questions
Speaker 2: Yeah
Speaker 4: bring us to a, to a close here. So what is one function within HR or the people department that you would fully hand to an AI agent tomorrow if you could?
Speaker 3: I would say analytics or, or digesting a lot of information. Um, I would hand that to the AI, but the-- again, [00:32:00] that would be the digestion and creating outputs, but then the human would still need to do the judgment call on it
Speaker 4: I love it. I was gonna throw your caveat on there at the end if you didn't, because I made that... That was a hard question
Speaker 3: Yeah
Speaker 4: the answer is nothing would get fully handed to the AI agent completely tomorrow. The human would be there to own the as
Speaker 2: Yeah. Yeah,
Speaker 4: in, in our conversation.
So finish this sentence for me: The risk is not deploying AI agents, it's outsourcing thinking. How would you add on or extend that thought?
Speaker 3: I would say our job as, uh, in our profession is not to implement AI, it's really to figure out where the humans sit in the AI construct. So in other words, our job is to humanize AI at work. That's what I would add on to that
Speaker 4: E-excellent. I'm, So what I endeavor to do at the end of every episode, and I was writing frantically here, as I usually do, is try to recap some of [00:33:00] my key takeaways from our conversation. So just as you are an engineer I think we all could benefit from the takeaway of, like, thinking like an engineer and starting with understanding the why.
I think that forces you back to starting with the work. So I liked sort of that tie out. And I mean, this is-- should be no surprise to people leaders, but always worth a mention the one key to success or failure every time in everything is the people. And so starting also with the human aspect of, of where we're headed here from change management to keeping the human as the out- owner of the outcome, et cetera. And then probably the sort of something that was central to our conversation and came up in a couple of different ways is AI as a teammate versus a tool.
Speaker 2: Mm-hmm.
Speaker 4: a few interesting takeaways here, right? A teammate is someone that you onboard, right? It-- A
Speaker 3: Mm-hmm.
Speaker 4: someone that works with you or for
Speaker 3: Mm-hmm. Mm-hmm.
Speaker 4: be more on the
Speaker 3: Mm-hmm.
Speaker 4: like I like... A tool is something you pick up, right? And
Speaker 2: Yep. Yep.
Speaker 4: you [00:34:00] onboard.
Speaker 2: board. Exactly.
Speaker 4: think about that differently, that, that is
Speaker 2: Hmm
Speaker 4: then of course w- there's something I just recapped in that last bit, is that the human owns the outcome.
The, the outcome is owned by the person. It's like a d- a better way, I think, of like human in the loop. I mean, everyone's
Speaker 2: Yep.
Speaker 4: that phrase, but so like let's... So how about just humans own the outcome, right? AI
Speaker 2: Yeah?
Speaker 4: us to own the
Speaker 2: Yeah.
Speaker 4: better, et cetera. So
Speaker 2: Yeah
Speaker 4: guests though the very last word Q.
So what do you want the HR leader who just listened to us chat today to go and think about on the rest of their drive home today?
Speaker 3: Yeah, I would say just focus on starting small and building the blocks to a h-bigger process or pain point. And just keep in mind the one thing that I just said again, because I feel like that's the most impactful, is your job is not to use and implement AI. It is actually to decide what stays human in the AI world for all of us, and we are the [00:35:00] best at it.
The AI, the AI might digest data, might give us data. AI is never going to be able to, in my view, read the room and make the right critical judgment of what needs to be done next completely autonomously. So we, we have the power. Uh, let's remember that and let's figure out how to kind of orchestrate in a world where we can all coexist and just amplify our, our own power and capability to drive the business forward.
Speaker 4: I love that framing. Thank you so much for joining me
Speaker 3: Yeah. Yep. Thank you for having me. Yeah. Likewise. Thank you
Thank you for joining us on The Human Element, presented by Wisq. If you enjoyed this conversation, be sure to follow The Human Element wherever you get your podcasts. I'm Barb Bidan. Thanks for listening, and we'll see you next time