Building A Company of AI Super Workers - Tracy St. Dic

[upbeat music] Welcome to the Foundation Series, your deep dive into startup compensation with industry leaders from across the startup world. Join me, Matt McFarlane, a people operations leader turned compensation specialist, as we uncover the strategies and practices driving success in tech startups around the world. Hear insights from heads of people, founders, and experts as we explore how to build robust compensation frameworks that not only fuel growth and retention, but do so in the dynamic startup environment that we all know and love.

Let's get into it. [upbeat music] Today I'm joined by Tracy Same Dieck, who is head of talent at Zapier, one of the companies genuinely furthest along the AI transformation curve. We're gonna get into how they rebuilt hiring around AI fluency, whether AI talent should actually be paid more, and Tracy's answer might surprise you, and what a manager's job even looks like now. Let's get into it.

All right. Well, Tracy, welcome to the Foundation Series. I'm so excited to have you here. Thanks for joining me for this conversation.

Thanks for having me. I'm so pumped. I have like 10 bajillion questions that I'm really eager to get into, so I, I am just gonna dive straight in. I like...

I don't know. I like... Zapier to me is one of those companies that is, like, famously leading the way when it comes to this, like, AI revolution in the workforce. I really would love to, like, go back to day one.

Like, take me back to the start, kinda, you know, what, what, what was the, the head space? What was the, the vision, like, thought process, whatever it was at Zapier when, like, AI came on the scene and went like, okay, this is, like, this fundamental new technology, um, a- and kind of from there on, like how it was started to be adopted in the, in the organization. What was day one like? Well, when I think about day one, Matt, I think about when our CEO, Wade Foster, called a literal code red.

It was like code red. It was about six months after ChatGPT-3 had come out, so this was like spring 2023, March 2023. And I think it was very clear at that point that not only was AI gonna fundamentally revolutionize the way that our business and product needed to look, because Zapier's always been an AI automation company, uh, or an automation company. And then now with the power of AI, it would just allow us to fulfill our mission even more, which is to make this technology available to everyone, right?

And so we knew it was gonna revolutionize our product, but to Wade's credit, and also our chief people officer at the time, Brandon Sammut, who is now our AI transformation officer, they both decided this also has to revolutionize the way that we're gonna work. Um, and to be honest, day one, I think that there was all the range of emotions that we've seen from people. There was skepticism, there was fear, there was excitement, there was curiosity. I mean, Zapier was not immune to the whole spectrum of how people were feeling.

But what I think Zapier did that I really admire, especially now in retrospect, is we said, "We're committing. We're going to shut down the company for, you know, a few days, a week, do a hackathon, let people get hands-on on tools." We say hands-on keyboard. Do the reps, start playing with the technology, start experimenting, get curious, get creative.

Um, and I think that started a real domino effect and a real wave of people feeling like, hey, this is something that's not only really important, but it's encouraged, um, it's something we have to get over our fear of, and they're actually... the company coming from leadership is gonna give us the space, the psychological safety, the time, um, and the guardrails and clarity about how we should use this and why we should use it. And so I think to Zapier's credit, we kind of got over...

I don't want to say we completely got over it right away, but early on we got over the wave of AI's gonna take my job, and we really got into the abundance mindset. What else could we do if we were AI-enabled? Not how can we cut costs, but how can we do more? How can we do things that we've never dreamed of before, never would have had the time to do or the opportunity to do?

And so I think we, we tried to make that pivot really quickly. So I think that was a big pivot point for us. We started moving our work with AI into our rituals, our team meetings, our enablement sessions, our employee resource groups, our ERGs. They started hosting training sessions and enablement sessions so people could learn in a community they felt safe with.

Um, and then honestly from there we've had a couple of other big pivot points in the organization to really increase our AI adoption too. Yeah, I love it. It sounds like you had a really, like, clear moment where the business was like, okay, we're, we're kicking off with this. There's no, like, I don't know, no gray area about adoption.

Like, this is the, the, the path we're taking. I'm sure though there was still some, like, sticking points, like some people and that hesitation. Like, what, what from a, like, you know, I guess a cultural perspective were you able to do to try and help circumvent some of those things? Yeah, this feels like a not brilliant answer, but I just...

as a people leader, it's like this is change management 101, right? It's like explaining the why, helping people buy into the vision, helping people understand not just the vision of where we're trying to go, but the direction. What steps are we gonna take to get there? What expectations do we have of you, and how are we gonna help you meet those expectations?

It is, it is about being demanding about how we want to move towards being a more AI first company, but it's also being supportive. Like, you don't have to do this alone. We're not gonna just say figure it out. We are going to provide more enablement, more time, more space to do this.

And so, you know, d- it's not like a, a brilliant tactic. It is literally going back to the roots of incredible change management. But I do think, I think there's some nuances. I think, one, our AI adoption to transformation journey really did come from the top and bottoms up.

We very early on encouraged a lot of citizen development, encouraged people to experiment and try things and share out. We would highlight that, reward it, incentivize it, you know, in all of our systems, performance management, you know, rewards, compensation, promotions, things like that. Um, but it also came from leadership. I mean, Wade and the leadership team really early on was like, "We have to model this.

We have to lead from the front." So like, "I'm gonna get hands on keyboard and build. I'm gonna expect the executive team to come to our meetings and share what they've built recently." Um, and I think that also was really fun to see executives also figure it out, struggle, and say like, "Hey, I built this cool thing.

We should try it," you know? So it kind of put everybody on a, a even playing field. So change management, but also culturally it had to be a we're all in this together, not you all figure it out. And I think that there's a variety of organizations that are doing variations that are not effective because either it's all coming from the top down or they're all come from the bottoms up.

Yeah. It doesn't surprise me. I've always been a big fan of this kind of concept with like, I guess just general leadership. It's this sort of like challenge, but support.

Like, "Here's, here's where we're hoping to get to. We have every faith you can get there." You've given some really incredible examples about how like Zapier's created the environment, the safety, um, you know, but also the guardrails to like experiment and try some things. But it seems like a, a really good combination.

Is that part of the reason why you think the AI transformation lens and the responsibility like sort of landed with, with Brandon? 100%. Yeah, I, I think there was this time maybe in 2025 where organizations were trying to figure out who should lead this AI transformation. Is it the people team?

Is it chief technology officer? Does... Like, who does it come from? I'm really glad that we landed early on that it comes from the people team, which was the impetus for brand transition into now being our chief people and AI transformation officer.

Um, and I think it's because of all the things that you said. I mean, the, the tooling changes, but what's going to actually get a company from, you know, zero to 60 is the culture change, the habits, the rituals, the practices. Working with AI is not an end of its of... in of itself, right?

It is an operating system. It's a way of working. It's an approach to how you, you do your work and how your best performers emerge, right? You wanna empower people with the tools.

It's not unlike how we thought about automation. And so to be fair, I think Zapier had a, maybe a little bit of a leg up, right? [laughs] 'Cause we've always had this culture. One of our values, way before AI, one of our values from founding was build the robot, don't be the robot.

We always had this automation first mindset. Like, do I as a human need to do that or can the computer do that? Is it... Like, is that my unique value or can I like figure out a way to get that automated?

So we've already had as a company that ethos, and so when we moved to AI it wasn't about we have to, you know, completely change everything about the way that we, we do everything. It's like how do we lean into that more and we accelerate that with AI. And so that is a behavioral, you know, like performance task I think. And so that's why it landed with the, the people team.

Yeah. I love it. I've always kind of joked that I'm, uh, that I'm a lazy person because I like don't like to do the same thing twice, but generally sometimes I need to like put a little bit more filter on when I try to automate something 'cause I'm like, actually I do it so infrequently and the cost to automate this is so high that it's just not worth it. But I do, I do typically like to really just be like, is there some way that a robot could do this for me?

So I'm, I'm all for this. I really think it's a kind of almost a mindset and a skill. I've been telling people this too because so much of knowing and understanding how to build workflows, it's not rocket science, but you have to retrain your brain to think about trigger action. This happened and then this and this and this and this and this needs to happen like either sequentially or in this order.

You just have to kind of retrain your brain on how to think that way. But once you do, you unlock a lot and you can get really creative, and you can like start to pattern match, which I think is, um, part of what I call like the magical moments people have either with automation or AI where, where it's like, whoa wait, I didn't have to do that terribly annoying repetitive thing that I do every day? That's amazing. More of that please.

No, I love it. And I guess like as part of this journey around AI fluency, I know again like famously you, you published your AI fluency framework, but I don't remember if I saw version one or what I now understand you're on, which is version two. But in terms of I guess charting the course from like where you were from where you wanted to be, talk me a bit... Like, talk me through a bit about like how that was developed and how you kind of landed on like where you were trying to get everybody, bring the workforce to.

Yeah. So I think it was actually back in like late 2024, maybe even the end of 2023 where I had, had brought up to Brandon and to Wade like, "Hey, from a hiring perspective, should we start measuring or thinking about AI?" And at the time talking to a number of execs it was still so new. People were like, "We don't even know how to define that.

We don't know what AI, use with AI in a good way means. Like what is good? What is, what is enough?" Um, and this technology was moving so fast.

And so we sort of put it on the back burner, but it was always in the back of my mind. And then I noticed, you know, as we went into 2025 we started to want to see more candidates have these AI fluent skills. But couldn't define it, and it was really inconsistent, and we would get people through the final interview and we would say, "We don't think they're AI fluent enough." And then we'd be like, "Well, what does that mean?"

And so as a people leader and someone who, you know, owned the whole front of the funnel, that's when I got even more into conversation. I kind of think Brandon and Wade and I all at the same time came to the conclusion we're like, we need to define this. When time in spring 2025 no one else had really put a stake in the ground and said this is what AI fluency is and this is how we're gonna define it. Um, so that's sort of how V1 came about.

It was really understanding, you know, where- The best performers at Zapier were using AI, how they were using it, and then trying to codify that and say, "Okay, how do we get more people to do that?" We're seeing already this incredible demonstration of people who are utilizing AI in smart, responsible ways are able to start to amplify their impact. So how can we coach that, train that, but also search for that, and what are the qualities that makes that true? We knew very early on it was not about the tooling.

It wasn't like, you know how to use this particular tool. We knew very early on it wasn't tool agnostic. So even our first version, um, moved away from AI fluency as using tool, and AI fluency is a more holistic perspective of how you think about using AI and approach your work with AI. 'Cause again, AI isn't the means in and of itself, it is the way to get to stronger performance.

Um, and so, so we created version one. We monitored, you know, from implementation, like how did those people do? How did they on-ramp? We also changed and redesigned how onboarding works, knowing that we were getting a very different candidate coming in.

So what was cool about it is once we started, um, utilizing the AI fluency rubric, that was our first draft, it then changed everything in the, in the spectrum. It changed onboarding. You know, Brandon was able to work through, like what does this look like for performance management? What does it look like for rewards?

And we were also training up our team, our current team at the same time. So I always like to say, Brandon and I tried to meet in the middle, um, to really transform Zapier. Now, version two has been really fun because we significantly raised our minimum bar for what we call the capable level at Zapier. Um, we, we also added in components that we thought were really important.

So our components before were mindset, strategic acumen, and builder skills, and we've added in accountability as the fourth one, which is an important piece. We've also reapproached how we thought about the slope of how someone learns and works with AI versus just where they are in a point in time. We call that slope slap snapshot, and we also differentiated what this means for managers too, that managers have a role in leading their team through an AI transformation, creating the psychological safety, having the vision.

So we've added a bunch of components, which I'm happy to go into, but, um, I'm really, really happy where it's played out. Um, and it has, has honestly, I think, revolutionized how we think about talent and what we look for, both from hiring, but also how we manage our current team. Yeah. Love it.

It... I guess before, I- I'm keen to dive into, like, some of the applications that you mentioned, but before we do, I, I think one of the sticking points that, that kind of keeps coming up for AI is this sort of like, is it actually, like, is it actually helping our business? Are we more productive or, you know, ROI, things like that. Uh, but what's Zapier's approach to this in- instead of like a, you know, presumably it's not just like a blindly, you know, pursue AI at all costs, right?

Like there's, there's really like a what's it doing for us? What's, what's the approach internally around that? Yeah. So I really like that early on we kind of said if, how are we gonna measure ROI of AI?

Um, and we have three components to that. It is obviously efficiency, which is an obvious one, but also quality, and then employee experience. And I think it's important that we measure all three. So any, what I would consider any, like, successful AI workflow that has strong ROI for the company hits at all three.

It might make us more efficient, but it also might prove the quality of what we do. So in my world, it's like quality of decision-making on candidates, quality of insights about the talent marketplace, um, quality of our interviewer and hiring manager enablement, right? Like, so that they can make stronger decisions. And then it's the employee experience.

So like, how are my recruiters, are my talent folks actually feeling like they can do their best work and, and add unique value? And then in our world it could also include, you know, empl- um, candidate experience. It could also mean we need to improve hiring manager and interviewer experience, so that hiring is just generally more delightful for everybody. Um, so those are the ways that we measure ROI for anything that we're experimenting with.

And I'll be honest, sometimes things that we implement don't make us more efficient, but it makes us have stronger quality in what we're trying to do, and vice versa. But in an ideal world, we're hitting all three. Now, there's like another s- like, second level question that I think has come up recently, which is like, given all of the tokens and the cost of AI, how are we thinking about ROI there? And I think we're, you know, very excited and lucky that at Zapier we have a really open playground and sandbox to play with AI.

We don't have specific hard limits of tokens that we can use, and obviously we're monitoring it from a financial perspective just to make sure we don't get overblown. But we're really thinking about, okay, given the work that our team can do now and how empowered they are with AI, what is that, what is that bringing us on the other side? And how are we comparing that with the cost of what we're using? And so we're just starting to look into that a little bit more.

I think from the people team perspective, one of the things that we saw was like, given the amount of tokens that we're using per month currently, if that were to continue, it would be at the end of the year, maybe the equivalent of one full-time FTE. But then you have 45 people on the people team amplifying their work every single day. So that trade-off feels really worth it, right? Like, we can do so much more with AI, and it's the cost of maybe one FTE that we're not going to hire in the future.

Um, and so but those are the things that I think companies need to really start to think about, and every company will be different. Um, but again, for us it's less about cost-cutting and it's more about abundance mindset of what else could we do that we couldn't do in the future, and how are these roles changing? Yeah. That's fair.

I feel like, um, token optimization is that thing that's gonna be a rude awakening for a lot of, uh, a lot of companies soon, and, um, yeah, being able to hit some of these highs and then think about how they can maybe claw back some of the, the, the token usage is, is possibly a next thing for us all to think about. It's a funny question too, because you, you know, you have somebody... There's this whole thing everyone's talking about token maxing, and at Zapier we say, like, value maxing.

It's like, how do you, how do you- Get maximum value out of what you're doing, and how do you know? And I think there's a piece around this where everyone is having this little reckoning of consciousness. It's like, oh, I don't have to use Opus 4.8 for every single thing I do.

I could actually use a haiku model or sonnet, and that's fine. But, like, can I differentiate that? Or I would even say, like, people reach for AI all the time thinking that's what they're supposed to use, when honestly, what they really need is just a very simple, straightforward automation, which is way cheaper and easier and more predictable, more accurate, more deterministic. That's so important in our work, right?

And so I think that there's almost, like, a nuance and a sophistication about understanding the spectrum of deterministic workflows versus agentic workflows that, um, people are just starting to understand and starting to get more sophisticated, um, and nuanced about in terms of guidance and coaching for teams. We certainly are. We're still figuring it out. Um, but we have people, you know, who use a ton of tokens, and I would probably be in, in, uh, that bucket.

And I'm like, is that good? Is that bad? We don't know, right? [laughs] Like, is that, like, I'm using a lot of AI good, but, like, am I getting maximum value out of it?

I don't know. So we're still kind of in this messy middle of figuring it out. So no token leaderboards. That's not your performance management model of choice at the moment.

No, that's not our performance management model of choice. What we are trying to do though, honestly, is to help people be aware. It's like, hey, we're running all these, like, really expensive models. Do you need to do that?

Or can you, can you, you know, can you do something different that gives you the same output but is more cost effective with a different model? So it's just like training and, and educating people, I think, about what they can do. All right. So we've talked a bit about, like, hiring, onboarding, performance reviews, and, and kind of the reward space.

We've touched on each of them. If you were to sort of like peak one to, like, dive into and just, you know, speak to for the audience, like how it shows up in practice, which, which is, which is your favorite? Which one would you wanna just go into detail on? Oh, gosh.

Well, I mean, I think hiring and onboarding is, is definitely more of my domain, and so happy to go into detail, detail there. Yeah, let's do it. What did... Where did you start?

What, like, were the big overhauls that you made? Yeah. So, um, I... Well, I think a, a couple of high-level things, then we can kind of, you know, take the conversation, Matt.

But, um, I think when we realized that we were going to be using an AI fluency rubric, um, and we implemented it, I think one of the things that I was really happy about that Brandon and I were aligned on from the very beginning was that if we're going to do this, if we're gonna raise the level of expectations for all of our candidates, knowing that we would shrink the candidate pools and shrink the talent pools, um, we, we were very, like, eyes wide open about that. We also need to raise the level of support in which to do that.

So just to give you an example, we measure AI fluency in at least four different pieces in our hiring process. We do it at application and recruit- and then at recruiter screen. We do it in a live skills test, and in version two, we have, like, made all of our skills tests live, so we can see, iterate, and thought partner with AI. That's a really important component of it.

And then we also assess it at our executive interview at the end, which we call our bar raiser interview. And the reason why we do that is because we also want to lean into this idea of slope and not snapshot. We want to see how people learn even over the course of our own hiring process. We wanna understand how people were learning and using AI for the first, you know, six months before they even came to us, but we also wanna understand, even in the four to five weeks you're, you're in our hiring process, are you continuing to push yourself?

That's gonna give us a meta signal on how they learn. But in order to do that, we don't wanna just be like, "Figure it out." So we give them access. You know, even when they first apply, the first email they get is like, "Thanks for your application.

AI fluency is really important to us. Here's, like, five different tools, websites, trainings that are not from Zapier that you can do. Here's a cursor guide. Here's a Claude...

You know, a, an Anthropic guide to AI fluency." Like, from the very beginning, we're being very clear, like, we care about this. We want you to be successful. If you are not feeling as AI fluent as you wanna be in this process, start now and start learning.

Um, and so we, we really pride ourselves on giving that support on our go-to-market side with our, um, account executives process. We noticed that a lot of account executives, um, were using a lot of traditional methods to be successful, and not many were as AI forward. And so we even created an assessment for them that said, hey, like, tell us how you use AI now. We'll let you know sort of where you would've fallen on our rubric, and we'll give you tips to learn more and where you can go to increase your AI fluency so that you're actually ready for this hiring process.

And so we've been playing with a number of things to better equip candidates. And what Brandon and I aligned on really early on was not only do we wanna increase support for our candidates, but the hiring process at Zapier, regardless if you get an offer from us or not, should be a learning experience for you. And as an AI company that is trying to be on the leading edge of AI transformation, we feel like that's also our responsibility to our candidates and anybody who applies to us, that you're learning something through the process.

Um, and that's also, you know, Zapio's ethos of being really transparent, sharing things, open sourcing as much as we can, making sure that people are... We're pulling everybody along with us. Yeah. We're not trying to be the AI company that leaves everybody behind.

That's not- Yeah, you're sort of, like, widening the talent pool again for the future, for someone who might look at Zapier now and go, "Okay, maybe I'm not ready just now, but I could be in three or six months." Yeah, exactly. And people always ask me, like, "Oh, like, it's amazing that you all share your AI fluency rubric." And, and to be clear, we share the external version.

We don't share the internal version 'cause we don't want candidates to, like, you know, see it. But yeah, of course we share it because you all can make it better. We're not trying to gatekeep it. Everybody needs to do this, um, or make a spin on it.

And so we're very happy to share. Um, and so anyway, that's, that's just how we, we try to support candidates through the process. And then I would say some of the very interesting changes that we've made throughout that might be notable is that we do this for all candidates, technical and non-technical. And so technical candidates, yes, they might have more of an AI-enabled coding assessment that we do, but even our non-technical candidates have an assessment where we want them to build live and iterate live and show us how they work with AI, show us how they Um, think about what good looks like, how they can differentiate between slop and, and real work.

And so that has been a really important component of our process that in practice I think has given us so much more information, and it's also supported our ability to, um, understand better what people can do versus just what they can generate with AI and then turn into us. So there's been some things that we're trying to do in the process that also help us combat fraud and embellishment, um, but help us really still continue to lean into the skills we wanna see. Yeah, interesting.

Were there many sticking points moving from, like, the pre kind of AI, uh, rubric through to, like, adopting this approach? Were there... Did you find it, like, impacted candidate experience or, like, you know, volumes or anything like that throughout the, the adoption process? It absolutely impacted our candidate pool, and understanding someone's AI fluency is typically very hard to source for unless they are, like, very, um, vocal about how they're using AI, like, on LinkedIn and, and things like that, right?

And so we... It was tough at the beginning to really help hiring managers understand, like, yeah, you could have an incredible tax accountant, but if they're not AI fluent, they can't work here. And to say, like, you know, people always ask, "Well, what if they have, like, the strongest subject matter expertise that you need?" And we're like, "Yeah, but we've, we've created this bar for a reason."

Because even if they're the best, I'm gonna use tax accountant again, even if they're the best tax accountant you've ever seen, they're not going to be able to thrive here at Zapier if they're not starting to think in an AI forward way. And there's a lot that we can train and teach you, but there's a lot that you have to come in with mindsets. Like, I can't imagine us ever being able to hire someone at Zapier who didn't believe in the future of AI and didn't wanna get curious in using the tools.

Like, that just wouldn't work. That's why it's holistic. It's not about the tools you're using, but it's about how you approach AI, how you think it's gonna change your work, how you're strategically thinking about your work. Um, it's so much more holistic than just, like, I can use an Agent harness and know how to create workflows.

Yeah. Interesting. And then moving to onboarding then. So they've been successful through the process.

They're coming into the business. I would love to know more about what that looks like and how it differs from a, a traditional route. I mean, some of the things that come to mind for me are things like, you know, do people bring their own skills? Do they, you know, do, do they build that through the onboarding process?

Like, how... What does that look like in a, in an AI fluent workspace? Yeah. So a couple of things, and I'll also say this is really iterating, so, uh...

or it's evolving with time. So when we first, um, implemented the AI fluency rubric and we knew, okay, we're gonna get a very different sort of average candidate that is more AI fluent than before, we changed how onboarding looked. So they were actually working with, I think we chose Cursor at the time, Cursor week one. They're already building with...

They're already working with the Agent harness. They're already building workflows. So that layered on top of also our product fluency, learning how to use Zapier, of course, got our, um, new hires building right away in week one. So it kind of started this idea of, like, this is the ethos and culture of Zapier, and it's also a place where we're gonna enable you and support you.

And so right away, that's what was happening, and we're... we encourage people just to play, experiment, do fun things, like create personal workflows, just to continue to experiment and, um, and yeah, and get really curious. I think there's... Zapier really prides itself on understanding that experimentation and play are a big piece of the learning experience.

So that, that piece has changed. As we now have moved forward, there's other pieces of onboarding that have been really important. Um, one is that our company context is much more callable in this world of AI. So we've made a pretty significant move to go from, you know, individual AI, everyone's doing their own thing, to a bit more institutional AI, and part of that is ensuring that we have our company context accessible and callable to everyone.

So think about being a new hire, you know, three years ago, and you're like, "Where do I go to for this? Where's my guide to learn this? Who do I go to?" Whatever.

Now all of that information is at their fingertips, and so there's a lot more self-service that people can do, and there's also a lot, like, faster on-ramp of just, like, general context about the company and their team, decisions that have been made and why. Um, and so it's really been fun to see how that has accelerated and empowered people very early on. Um, and then there's some fun stuff that we're doing too. Like, on the TA team, we hired, uh, three or four new members this past spring, and we decided to recreate on our own onboarding to be all within a skill on Claude Code.

And so the hiring managers used it to create the onboarding package, like, pretty seamlessly. We also created a skill that people can jump into on day one. It, it helps them understand everything about onboarding, who to go to for what. It creates automated calendar invites for the people they need to meet with.

It allows them to ask, you know, all of their basic or silly questions without feeling, uh, vulnerable, and it also checks for understanding along the way. So it doesn't take the place of the live important sessions, but it, it actually is more designed for adult learning, to be self-service, to, you know, go at people's own pace. And so I think there's some really cool accessibility things that are happening with onboarding right now, now that we've been able to make it more personalized, um, and more tailored to individuals, as well as being ultimately consistent and standardized too.

You can actually do both. And so I think it's been really exciting, um, with onboarding, and I'm actually about to bring the onboarding unit, like, under, under my wing in talent. Um, this is gonna open a lot more possibilities too. Yeah.

Exciting. Gosh, I love the idea of a, of a, of a company context, and I think where my head goes to is, like, if you really can truly track, you know, all the decisions that are being made, I mean, to some extent, you know, conversations that are being had and, and, and outcomes and, you know, how things were determined, you really, like, remove this sort of, like- I guess some of the fear of, like, attrition and things like that, that inevitably goes with someone who now kind of like, you know, knows where the bodies are buried, you know, knows why decisions were made that way, and maybe knows how something works that, that nobody else does, so if it breaks, there's no one else who can, who can fix it.

So it presents a completely different paradigm from that perspective. Yeah. I think there is so much more risk mitigation in that. Absolutely.

Absolutely. Um, and to be honest, I think companies that are remote are gonna already have a huge advantage here. I mean, we already have a huge advantage here because so much is written down or documented in some way, even if it's just Slack or email. Um, and companies that are in person have lost a lot of incredible company context to a whiteboard or a meeting where someone didn't take notes, you know?

So I think that's gonna be something really interesting to watch as we've had this huge move back to being in person across the board. Um, I think remote companies have a very distinct advantage in an AI-driven world. I would agree. Yes, having worked for a couple of them that were, like, proper, proper, truly remote.

Um, yeah, I think then there's so many practices... This is a whole other conversation, but there's so many remote work practices that even in person companies should be adopting but unfortunately don't because it's, it's just easier not to. But yeah, I completely agree, definitely helpful from, like, a, a context creation and gathering perspective, by all means, when everything is just naturally, like, you know, documented around. Yeah, definitely.

So anyway, I could also go off on that because I'm a very of remote work, obviously, with Zapier being a global remote company. But, uh, you know, everyone gets to choose their own adventure there. I love it. And, um, one of the things you touched on a little bit earlier was, like, around the reward side.

I would be remiss not to ask because I think this is something that comes up a lot for me, is like companies that I think are still very much either at the early or, like, middle stage of the sort of transition that I think Zapier's already been made, um, which is like, how do we think about comp and reward, and, like, how do we think about it from the perspective of, you know, rewarding and recognizing when people are, are doing great and exceptional things with AI, but also not, I don't know, not out laying huge amounts of cash on what might be a flash in the pan moment, which is this kind of AI, you know, AI, uh, optimization piece.

What was that journey like for, for Zapier? How do you think about it? Because I, I would imagine, like, you've been so such vocal proponents of, of, um, AI adoption. In some respects, I, I would imagine it makes people really wanna try to bring some of Zapier's talent to their own organization.

Like, how, how do you think about that from a reward perspective? It definitely makes other organizations want to bring Zapier's talent to, uh, to their own organization, which is something we are, you know, obviously v- very happy about. Um, you know, there's a, there's a, there's two sides to that coin. But, um, you know, so one thing that's interesting, I got asked pretty early on, like, "Are you gonna start to pay people differently if they are more AI fluent?"

And so far, the answer to that has been no. Now, we, we obviously have different pay ranges for people who are doing niche work, like working in applied AI or ML. That's not what I'm talking about. But I mean, like, if I was hiring a, a recruiter who was more AI fluent versus a recruiter who is not, would I pay them differently?

And the, the answer across the board at Zapier, we've, we've just said no, because our expectation of how people are using AI is going to become more and more normal and hack- happenstance, right? That's even reflected in our AI fluency rubric. In V1, it was okay just to, like, dabble at AI and use it one-off, and now that's not even... That's unacceptable.

Our, our rubric is like you need to be using it every day as a part of how you work. That is going to be so much more normalized and expected. Um, and so I think it's just like, like we don't pay people more because they type faster, that if they didn't, it's just like that might change the, what you're able to produce, right? And so I think that the focus has to continue to be on what are, what are people producing?

What are their outcomes? Whether they use AI or not, we know that people who use AI can produce, be more efficient, have higher quality, have a better experience. I think that's gonna be obvious. AI, again, as a tool, will become a more obvious choice for high performer.

Um, and people who are high performers will know how to use it well, which also includes knowing what slop is and what it isn't, and being accountable for that work. So I think our focus has always been on rewarding top performers, and what we're doing right now is we're watching and monitoring and observing who is using AI and in what ways, and is that something that's leading to stronger performance? Overall, we are seeing yes, like just in the broadest sense, we are seeing in general, people who are using it are stronger performers, but we're not rewarding them necessarily just for their AI usage.

We're rewarding them for their impact, and we're seeing that people who are using AI are significantly outpacing their peers, again, in productivity, in quality of work, and, and also their engagement as well. Um, now I would say like on a micro level though to in- incentivize behaviors, we are trying to, in this transition period, we are trying to, um, really reward and incentivize people who are leaning into experimentation, teaching others, enabling others. Um, so somebody, for example, on my team who has become, you know, more of an AI coach and they're, they're building a lot and they're sharing a lot and they're supporting others to build, like that's somebody that I would recommend for, you know, a spot award or a bonus, right?

Like, so those are the things we're trying to encourage those behaviors in small ways. Um, and it is built into one of the mindsets and orientations in our performance management system. It's in every job description. So it's more of an expectation than it is a goal, if that makes sense.

Um, but with that expectation, we're still looking at overall performance. Yeah. Does that answer your question? I kind of- Makes a lot of sense.

I think, like, yeah, I mean, like, you know, my understanding, a- again, I think, you know, Zapier's always been quite a, like, an open and upfront company. I think my, my understanding from years ago was that you were always targeting high performance talent, and your, your comp and your award was geared towards that. And if I understand correctly, it's just the case that the frame for high performance has shifted to incorporate AI rather than you having to change anything to, to try and lure out talent that, that have that, like, complement, I guess.

Yeah, absolutely. Will say it's a tough talent landscape out there though, because not only are- From Zapier being recruited, of course. Um, but if there is a tax accountant that is really AI focused and AI forward, then that person is going to have... is going to be top talent, and is going to a lot of competitive offers.

And so that is something that we're watching. I mean, I think in general, we still need to compete with other companies who want the same AI forward talent. Um, so we're continuing to keep an eye on it, but I think that, again, it's gonna become a lot more normalized as more pe- more candidates and more companies are creating, you know, a, just a way of working. I'm curious, totally random thought about, like companies like ClickUp who were like fairly famously did a, like a, quite a large layoff when we're talking about this creation of like a million dollar salary, and just things like that.

How, how do you think, you know, I guess for a team that i- is again on the edge of these sorts of things, how did, how did you and how did the team kind of interpret those sorts of things? Like interpret the companies that are- In terms of whether or not it's something you need to respond to or just, again, it's keeping an eye on it and sort of seeing what the market looks like or... Yeah. I think it's really interesting because I think that there are likely companies that are saying, "We're doing a riff and it's because of AI" that that's not necessarily the whole story.

Like we can, we can like argue that, you know, back and forth, the ins and outs of, of companies that are doing that. But I think there again, there are gonna be two camps of companies, ones that say, "We're gonna use AI for cost cutting", which may be what they need to do, right? Especially if you're a public company, that may be what you have to do as a part of your strategy. Um, but I also think there are companies more like us that are saying, "What can we now do?"

So, you know, do we try to do this, do what we were doing before with less people and save money, or do we try to say we have the same people, they're so much more enabled, they're able to do three times the productivity, what could we do instead? And so I, I think there's gonna be a, I don't wanna say like it's binary, but there's certainly a spectrum of where companies fall in terms of philosophy. We are certainly on the side of what more can we do, um, while being, you know, smart in terms of cost, just like we always are.

Um, so I think it's interesting. I do think that what we're seeing in the market is that we're hiring for roles that haven't even been created yet. Like we're trying... When we hire, we're hiring for the future.

We're hiring for people who can do future roles, even if we don't know exactly what that looks like. And so there is going to be for talent, a need to evolve very quickly, regardless of role and regardless of function. How are you changing the way that you work? How are you rethinking how you do your work, whether you're a recruiter, a tax accountant, or an engineer or something else?

Um, and if you don't do that very quickly, then I do think you're gonna be left behind. And so when I see companies who do these layoffs, I'm wondering, are they, are they doing it because they have created such incredible AI ROI that they feel like they're at that point where they can do that? Or are they cutting people who are not willing to use AI? Or are they...

Like there's just so many aspects of that. Um, and so all to say is, I think it's something to monitor and to watch and to like be aware of. I wouldn't say it's affected anything that we do here at Zapier, 'cause we're pretty philosophically clear on, on what we're trying to get out of AI and how we think about this for our people. Yeah.

Seem to be very much leading with your own intention rather than being reactive to, to sort of the environment. Yeah. And we will- You touched on something that is... Sorry.

And we, I say we will take that talent if it's incredible too. So... [laughs] That's the, uh... I'll put the plug in the show notes, and the link to the careers page.

Um, you touched on something that I, I would love to, I guess, explore a little bit before we close things out, which is this concept of like role design and the fact that it is evolving so quickly. I mean, there was, you know, talk about the elimination of pure manager roles, not at Zapier, but at other companies, but the elimination of, of pure manager roles and the introduction of like a player coach. There's, there's so much frothiness, I guess I would say, in terms of like what a role even is anymore, and like how they're evolving.

How are you thinking about that from the perspective of like, what are we even going to market for? What are we even trying to bring into the organization that like when we do talk about roles of the future? Gosh, it's such a good question. I will say a couple of things that, um, so far I felt pretty strongly about, and I say that with the caveat of things are changing all the time, Matt, so in six months I could be eating my words.

But one is, uh, something that we're looking for more now is, um, not that we don't have technical thresholds of, you know, the technical subject matter expertise that people do need to bring in, but I'm looking a lot more for what I'm calling durable skills, um, versus just the technical skills. So it's great that you can do X, Y, Z thing technically, but you know what? In six months, like you might not need to do that, or AI might be able to enable you to do that better. So what are the durable skills that are going to allow you to evolve as this landscape changes?

So, um, just to be concrete, a lot of times those are considered soft skills, but it's like, how are you navigating ambiguity? How are you proactively communicating with stakeholders? Um, how, like how quickly can you learn hard things? Like learning hard things fast- The human skills.

Exactly. Learning hard things fast is one of the most important things I think that we hire for right now. I'm very interested in how candidates are able to pick up something new, dive in, and become a master at it very quickly. Um, to me, then regardless of the next hard thing we're going to have to learn, we will have a workforce that is able to pick that up and evolve and move and not be resistant to the change that's gonna be coming with that.

So I think like, um, hiring for durable skills, which again, hard to source for, but very important to evaluate for in the process, is absolutely critical. I think just other, other concrete things is, you mentioned role redesign. I mean, one, I think that in general, people need to be really aware and thoughtful and articulate how they think their role and function is gonna change. So I wouldn't hire a recruiter right now who couldn't tell me how they thought AI was going to like enable their work, not just now, but in the next couple of years, and have a vision for what that could look like, 'cause that's what I want them to be building for.

And then on the management side, I mean, this is a really interesting one, because I think in some ways we've started to see companies say, we can have fewer managers- that manage a much flatter organization because a lot of the admin work of management can be automated. Not the judgment, not the people stuff, not the stuff that, you know, affects comp or pay or hiring decisions, but like, you know, progress to goals, writing agendas, you know, like all, all of the little things that mana- takes time for managers to do but isn't the highest value.

So I've seen companies move in that direction. I've also seen companies think about we have managers that are more player coaches, and they manage smaller teams because instead of managing a team of 10, you can manage a team of three and do the same work. Um, and as a manager, you could be a coach, but you can also have a strong individual contribution. At Zapier, we are testing both models in different ways.

Uh, and so I don't think we have a conclusion yet, but I do think that it's certainly gonna change the way that we think about management, um, and we think about manager roles because traditionally, manager roles just were like, you know, kind of... It was, it, it was really administrative in a way, even the coaching pieces. It's like, it's interesting to see how it might evolve to include a little bit more of a unique contribution as well as development of others too. So I think the jury's still out, but we're experimenting with both here at Zapier.

Gosh, yeah, my mind's, my mind is firing with all the different, uh, I, I guess, different things that we've covered here. Sorry, we've got... Yeah, I'm, like, even ha- having trouble articulating. Clearly, I'm failing the, uh, learn hard things fast test right now.

No, actually, thanks. I went in a lot of different directions, and you're trying to, like, reel me in, Matt. So I, I get it. No, but I love it.

Like, I love that you've kind of got this sense of like, we're hiring people who are, who are AI fluent. They've kind of got this like, you know, I, I know the word taste comes up a lot, but like, you know, judgment, curiosity, collaboration, those sorts of things. And, and, and I guess maybe it's like, sure, they may have some subject matter expertise in tax accounting or, or recruiting or something like that, but maybe it's just a passion or an interest for it these days, and actually, the other things are more interesting, and, and the rest will come if they actually have those, those other skills.

It's such an interesting change on the approach. And I mean, skills-based recruiting, skills-based compensation, skills-based everything has been such a dominant sort of thing, so it's, it's very much turning that on its head. And, you know, when, in a world when a lot of that can be ingested by AI and, and, and, you know, spat back out, then maybe it's less important. Um, maybe last que- second last question from my mind is like, what role do you think people coming into the career force, into the workspace, and just whatever, coming into their career, what, what role do you think they have in this kind of new, this new world?

That's an interesting question. Well, I think it's, it has never been easier to learn, to learn things, things that are outside of your subject matter expertise. And so I think it's incumbent on all of us, whether you're earlier in your career or not, to continue to utilize that and say like, you know, "This is my role right now, but I can go outside my lane and learn this other thing. I'm a recruiter, but why, why can I not learn about benefit strategy and compensation?"

Like, why not? It's, it's so easy to learn right now. Um, and so I think if you're, if you're earlier in your career, I do think that you have to be curious and utilize your resources to learn as much as you can. It is interesting to, to wonder if people are gonna become more specialists because they'll have this deep expertise that will be absolutely needed for discernment of understanding what's real and what's not, or if people are gonna become more like generalists and just, like, incredible generalists who can learn anything and pick anything up.

It's kind of the difference between, you know, T-shaped talent versus I-shaped talent and all that. Um, and so I don't know if there's a conclusion on what people coming into the workforce should do, but I do think that the career advice I give everybody is, like, dig into AI tools and consistently experiment and try new things. There... It is...

There are so many free things. There are so many opportunities to do this. Um, just getting in and experimenting in and of itself is honing a skill that's gonna be necessary later. Um, and then if something piques your interest, like, really dive into it because it has never been easier to do so.

Um, and you can, and you can learn a lot that way. So that might be kind of high level, but I think that that's, that would be my advice, for sure. Yeah. I would agree.

I, I still think we're so early into this journey, and I feel like it feels like when you scroll LinkedIn or if you scroll, you know, YouTube, and, and you only have to watch a couple videos before all of a sudden your, you know, your whole feed is filled with it. But it looks like everyone's got these billion-dollar solo businesses or side hustles or something. Like, e- everyone else is making out like they've got it made, when actually, if you look at the data, companies are still grappling with this.

Individuals are still grappling. A lot of people still use ChatGPT, the free version. They've never used anything more. Like, we are still at the, like, the start of this journey, and that certainly doesn't mean that I think people should hang back and, like, see what happens.

But, you know, you could, you could get started now and still be ahead of, of most people. And like you said, the resources are endless. If anything, they're overwhelming, but, [laughs] but there's a lot there to get started with. Yeah.

And I really just think that, like, you know, I don't think there's a lot of things where time equals better outcomes [laughs] in general, but I think with AI and learning AI, time and reps does equal better outcomes because it's one of those things where you just get better at it when you start to dig in and play with it. So we always say, like... I always recommend to even leaders of talent, say, like, "You gotta get hands-on. Do not rely on your talent ops team to build everything for you" Which, to be honest, I think a few years ago, I was in that camp.

I was like, "I have a vision for this workflow. My incredible ops team, can you help me build it?" I have gotten more hands-on-keyboard building things myself because it makes me sharper as a leader. It allows me to solve my own pain points, but it also is, is just a skill that everyone is going to need to have.

And so, um, I think there's a lot that just doing the reps is... it's just really important. You build confidence, you build clarity of your own skills but also what AI can do, and there's just no substitute for it. So maybe that's what I would recommend everybody, especially people in early career, to do.

Just get hands-on-keyboard. Just start an hour a day. You'll be surprised at how much you learn. Yeah, I love it.

That's probably the, the closing advice, I think. This has been amazing. Thank you so much. It's been so much food for thought.

I think, um, yeah, you've, you've, you've helped me. I think you're gonna help a lot of people chart a course between wherever it is they are in their AI fluency journey through to a point that is at, is at as advanced as, uh, Zapier. But I think the thing I love the most is that so much of this is still, like, being determined, and it's still in flux. And you...

Like, you don't have the answer. You're still figuring it out for yourselves. You're just maybe a little bit further along than some others. So yeah, this has been a m- been an amazing conversation.

I appreciate your time. Yeah. Thanks so much, Matt. This was super fun, and I hope it was helpful for people or give something, give something for people to think about, so.

It absolutely was. Thanks for joining us on another edition of the Foundation Series. Make sure to head over to our website to subscribe to the Foundation Series for monthly editions featuring future interviews on startup compensation. In each edition, we'll explore the different strategies, trends, and challenges faced by startups around the world.

If you have a burning topic you'd like me to cover or a standout guest that you wanna hear from, drop me an email so I can dive deeper into the compensation topics that matter most to you. Join me next time as we continue to uncover the strategies and practices driving success in global tech startups around the world. Thanks for listening.