Ep 135: Assessing and Evolving Human-Centric AI Readiness with Tracy St.Dic

[00:00:09] Hey everyone, welcome back to Elevate Your AIQ, your go-to source for insightful conversations on human-centric AI readiness, talent transformation, responsible innovation, and the future of work. Today I'm very excited to share my conversation with Tracy St.Dic, VP of Global Talent at Zapier, who spent 15 years in education as a classroom teacher and then leading national recruitment at Teach for America, before bringing that same belief in access and untapped potential to Zapier's talent organization.

[00:00:37] Tracy and I dig into how she and her team built and have since evolved Zapier's AI fluency rubric, what actually separates real AI transformation from surface level adoption, and why hands-on experimentation beats any classroom when it comes to building AI capability. We also talk about how her team is reshaping the recruiter role itself, freeing people to spend more time on the coaching and relationship building that only humans can do.

[00:01:03] It's a genuinely candid look at what it takes to build an AI-ready workforce from a forward-thinking talent leader who's actually done it. Stick around for a truly insightful episode and thank you, as always, for listening to the show. Let's go hear from Tracy. Hey everyone, welcome back to another episode of Elevate Your AIQ. I am your host, Bob Pulver, and I am looking forward to my conversation with Tracy St.Dic. How are you today, Tracy? I'm doing great, Bob. Thanks for having me.

[00:01:29] Absolutely. Thanks so much for spending some time with me. I know you are a very, very busy executive, but we've got a lot to dig into. So why don't we get started with just you just sharing with my listeners a little bit about your background and your current role leading talent at Zapier. So I am the VP of Global Talent at Zapier. I've been here for about four and a half years. And what that entails is I lead talent acquisition, talent intelligence, internal mobility, onboarding, employer partnerships.

[00:01:59] All the things that have the word talent may be attached to that title. But the other thing I get to do, which is really fun, is work on a lot of our company initiatives specifically related to our AI transformation internally. So some of the things that I work on with my team are our AI fluency rubric and ensuring that all of our candidates, our hires and our employees all are moving towards an AI future. And that's pretty inevitable. So really exciting place to be. Zapier is a AI company.

[00:02:26] So we specifically focus on AI orchestration and ensuring that Zapier can be your governed action layer, whether it's automations, agents, working with MCP, all of it. So we have a little bit of wind at our back, too, when we're building with AI. But before that, I was at Teach for America. I spent probably about 15 years in my career starting as a teacher and moving all the way to leading our national recruitment team at Teach for America before I joined Zapier.

[00:02:52] Amazing. So everything that we're going to talk about, at least as it relates to AI fluency and how people are sort of upskilling themselves and sort of absorbing this learning by doing all these different modalities and different ways of learning is basically, you know, near and dear to you from your earlier career.

[00:03:12] Absolutely. I think one of the things that I learned really early on in my career is that access to, you know, any of the things that are going to democratize, whether it's education or technology, is really important. And so when I started as a teacher with Teach for America, I got that lesson very early on. My students were incredibly capable, but I believe that brilliance is distributed everywhere and opportunity is not.

[00:03:37] And so they needed to be shown the tools, how to be successful in the classroom and then set that foundation for the rest of their life. What's really interesting is I think there's a parallel to like why I was drawn to Zapier's mission as well, which started as make automation work for everyone. The mom and pop shops, the big enterprises, everyone. And I think we're at a time right now with AI where that's more important than ever.

[00:04:00] We don't want to create a split of who can use this technology to improve their work, their life and those who can't. And so one of the things I love about Zapier and what I try to do with my team is to ensure that this technology, which truly can be life changing, is accessible to everyone. It is a tool that we can help them reach their potential and do the things that we want to do. And so you might not think that like tech and teaching kindergarten are that connected, but I really see them as sort of the same mission that I've been working through my whole life.

[00:04:29] Yeah, and I see a lot of, I mean, I see parallels in terms of like learning styles that you've got to embrace that people need some level of personalized learning. Some people learn better through sort of course instruction.

[00:04:43] And I mean, I know reaching to the choir here, but it just seems like when it comes to AI, I think people are, if you haven't realized it before now, you're realizing that it really does take, you know, getting your hands dirty, so to speak, and experimentation. Failure is a learning opportunity and just trial and error. And you can't just sit in a classroom through, you know, more traditional learning and development types of activities.

[00:05:12] You've got to really be in it and experiencing it to understand some of the nuance of how you use some of these tools more effectively. Of course, use them responsibly and ethically, which I know is part of your rubric and everything that you, you know, instill in all your Zapier employees. Do you call them Zaps? What do you call them? Our employees? We actually call them Zappians. Yeah. Zappians. Okay. Zappians. Zappians. Yeah, the Zapier automation. So that one was already naked.

[00:05:41] Yeah, that would make more sense. Yes. Got it. So, yeah, so talk to me about the genesis of the rubric. I mean, you guys are almost the poster child for, you know, how an organization can do this at scale across an enterprise, regardless of, you know, function or line of business, et cetera. And so, so just take me through like what went into that and how you, how you developed the initial version.

[00:06:10] I know you came out with an updated sort of version of that rubric recently, but just take me through that and how you sort of correlate it to the types of skills and profile that you look for in future Zappians. So I'll take you a little bit back to the context.

[00:06:28] We realized we needed some sort of AI fluency rubric and standard because as we were hiring people, which is so much of what my talent team does, we realized that we were not calibrated at all about what we expected. We knew as a company, we were moving towards every Zapian using AI more extensively. And even internally, we couldn't define what's good enough, what's okay, where are we trying to get people to. And imagine that in hiring is even harder.

[00:06:58] Is this candidate good enough in their AI fluency? How do we know? And the market and the technology was just changing so rapidly. And so the reason why we felt like we had to do this is we had to say, okay, let's actually define what we mean by AI fluency. And not just in terms of what tools you can use, because it was very clear even at that point in, you know, late spring 2025 that the tools were going to be changing all the time. But what was a more durable way to think about AI fluency for now and the future?

[00:07:27] And so when we first started, we had three pillars and now we have four in our version two that came out in April. And those pillars are mindset, strategy, building skills, and accountability. And those are very holistic. Only one of them has to do with actually using tools. And the rest around it have to do with how are you leaning in.

[00:07:50] So mindset for us is curiosity, leaning in, thoughtful experimentation, willingness to revisit how work gets done. It's sort of like that growth mindset. How do they lean into learning hard things fast, as I like to say. Strategy or strategic acumen is a little bit more function specific. So asking recruiters, what does the future of your function look like with AI? Can they envision that? Can they understand where humans should keep the human things, but computers could do more work?

[00:08:19] How do they start with the outcome and then decide where AI belongs? And so that is all about vision and understanding your work as a subject matter expert, but also understanding the role that AI can play in that in the future. Building or builder skills, that's pretty self-explanatory. We do like to see people use AI in smart ways. So we lean on Anthropics AI Fluency Index for some of those. So when we ask people to build, we're looking at do they iterate? Do they use AI as a thought partner?

[00:08:48] You know, not just are they technically able? Obviously, that changes based on the role, but really like how they build. And then the last one is so important. Accountability, we sometimes call this discernment or taste is, do you know when to use AI and when not? Can you check the quality? Can you stay responsible for the result? Because at Zapier, we say you can delegate the task, but not the accountability. And so that's incredibly important. I think now is it's so easy to build anything that you want.

[00:09:15] How are people staying accountable to the work that's being produced? So mindset, strategy, builder skills, accountability. Those are the four pillars now of our AI fluency rubric. And then we also believe that managers have different accountabilities as well or different responsibilities too. They need to ensure that they are building psychological safety. They are working through all the basic tenets of change management. They are helping people upskill in really intentional ways.

[00:09:42] So when we hire someone who's going to be a manager, when we look at a current employee who is a manager, we expect different manager behaviors that support AI fluency for their team and also for the whole company. So that's kind of the basics of how it works. And then, you know, we have different levels, what we call capable level, adoptive level, and then very, very hard to reach a transformative level. And that is just related to how you do the work. So capable right now for us is, you know,

[00:10:12] you're using AI in really specific ways for, to improve your performance. It's more individual. Adoptive is you're starting to build workflows and orchestrate things for teams and for others. And then transformative is really like, you have redesigned completely how work gets done. So it looks measurably different than how it did, you know, six months ago or two years ago or whatever it is. So I just think back to a couple of years ago

[00:10:39] and where we were when it comes to, you know, AI readiness, AI fluency. I mean, recruiters, like two, two and a half years ago, recruiters, they wouldn't even know what to ask, what to look for. They certainly didn't have any kind of assessment to test whether someone was using AI. And maybe they hadn't thought deeply enough about, you know, the elements that you just described.

[00:11:06] Like you said, I mean, even you've had to update it already based on the pace of change and the changing nature of just expectations. And also the fact that Zapier is leaning so heavily into its own AI future for yourself and for your clients and partners, et cetera. But I mean, I just remember my niece interviewing with a big consultancy a couple of years ago

[00:11:32] and she took the initiative to do the assignment that they asked her to do manually and then used AI and then just showed like a side-by-side. This is a step-by-step, this is what I did. And the recruiter was just like, like they didn't even know what to do except just, you know, drop their jaw and be like, that's just amazing. Like I didn't know you could do these things and whatever. So it's, we've come so long. I mean, I think the last McKinsey report

[00:12:01] I saw on AI fluency said that the demand for AI skills, AI fluent talent has gone up like 14 fold or something in the last two or three years. So it's amazing. And I still think that you guys are, like I said before, you guys are still, you know, sort of the case study of how to do this. But I also know you're pretty, I wouldn't say critical, but you know that there's so much more to do.

[00:12:31] I believe you rated yourself a four out of 10. Is that what you rated yourself? Which is, you know, amazing to think that where other organizations are. I mean, most organizations must be at one or two if you guys are at four. I mean, I think it just speaks to how much further I believe we have to go. I tend to think that there are companies, you know, and there's all sorts of AI maturity frameworks out there. We have one at Zapier too that we think about when we support other companies.

[00:12:58] But basically I think about adoption, AI adoption and AI transformation. Most companies right now, if they're doing anything with AI, are in AI adoption. And that's a fine place to start. We consider AI adoption as the ceiling, I'm sorry, as the floor, and AI transformation as the ceiling. So AI adoption is like people are using AI, they're bolting it on top of workflows they already have. They maybe are making some things faster, maybe you're making things a little bit better,

[00:13:26] but they're just putting AI on top of the work they've already done. And it's making them more productive as a worker. Maybe they feel like their work is getting better and maybe you're measuring it in multiple or in percentages. Like my work is 10% faster, this feels 20% faster, et cetera. AI transformation raises the ceiling. And that is where you're really actually redesigning how work gets done as if it was natively built with AI. I tell my team like, you should not be able to pull the AI out of it.

[00:13:55] If, you know, just like our work right now with the internet, like if the internet goes down, Bob, I'm like, I can't work. I don't know what to do. I cannot do my work without it. If we're truly redesigning for AI, it's going to be really similar. And so I rarely have talked to companies that are doing a lot in the AI transformation space where they are redesigning teams, they're redesigning how work gets done, maybe they're redesigning metrics, and they're thinking about how are we stripping it down to its parts

[00:14:22] and rebuilding something in this age of AI. I think Zapier is sort of leaning towards AI transformation. There are certainly pockets where that work is being redesigned, teams are being redesigned, et cetera. But even we're not there completely 100% of the way where things are much more agentic and focusing on that. And so that's why I say that we're four out of 10 because I see a future where all of that will need to be all of our work, whether it's in talent or accounting or legal, whatever it is,

[00:14:51] it will need to be rebuilt and redesigned. And I don't know if in general, the industry and any industry is kind of thinking that way yet. I think some people are envisioning it, but it's very, very difficult to get a company actually moving in that direction. When I was at IBM, we went through something similar when IBM Watson came out of the labs and Ginni Rometty said, okay, we're now a cognitive computing and cloud company.

[00:15:20] And most of the company, you know, 300,000 people were just like, what does that even mean? Right? So you had to go through this whole exercise of, well, let's take you through, let's explain what it means, what are the foundational technologies that even allowed us to create what you're seeing on Jeopardy or what you're seeing coming out of the labs. But like, this is what it can do and this is what it means. And now, as you learn more about it,

[00:15:50] we're going to go through this exercise of, you know, forming teams. Let's come up with new ideas. Let's put them through crowdfunding platform. Let's put them through a shark tank and all these things. And it really got people energized and it educated at least a third. We didn't get the whole company, but we got at least 120,000, 125,000 people to take the courses, these, you know, massive online courses, we used to call them. And it really was the sort of genesis

[00:16:19] of the next phase of IBM's enterprise transformation. And so these days, it seems even harder to do that because of the speed at which the technology is moving and because of the responsibility, I think, that we all have to build, to solve real problems with the technology. It's easy for people to, I think, latch on and just build something because they can, as opposed to building something that they should.

[00:16:48] So how do you think about that as people are trying to, I guess, move up at least one level in the rubric? That's such a good distinction of building what you can build versus what you should build. And I'd say, I agree. There's a lot that people can build and it's all very fun and exciting. And I honestly think there's a place for that. Experiment, there's such a place for experimentation right now because that's how you start to unlock creativity and build confidence in what you're building. But in terms of what you should build,

[00:17:18] there's a couple of ways to answer this. One, I'll give you a super simple framework that we use at Zapier when we evaluate any sort of AI initiative or experiment. It's asking us, does it improve quality, efficiency, and employee experience? Or sometimes we think of this as stakeholder experience because it's often easy with AI experiments to improve efficiency. I can do this. It used to take me three hours. I can do it faster. But what are the things that you're looking at to know if it's better? Is the product better?

[00:17:47] Is the decision stronger? Are, is the research that leads you to a different action that was a better decision for your company? Did it improve the customer experience, right? So those are some things that we think about with quality. And then the employee experiences, you know, as an employee, did this thing that I did with AI, did it actually make my life and my job better as an employee? Did it allow me to do more of the human stuff and did it automate the toil away, right? So really simple, quality, efficiency,

[00:18:16] and employee experience. And if we can answer, yes, this, what we're doing with AI is actually improving all of those things, then we consider that a good experiment or initiative. Sometimes it can only do one of those and that could be okay. But truly, I think the golden goose is like improving all of those things for, for the people who are doing that work. So that's just one super simple way that we look at it. But I think there's also other ways that we think about, you know, depending on the function,

[00:18:47] okay, what is this improving for certain stakeholders? So in recruiting, is this improving the candidate experience? Is it helping us get stronger candidates? Is it improving our ability to move people quickly through the funnel? Is it improving the signal that we get? Does it decreasing the fraud? You know, all of those, there's like so many questions in there, right? And every function is a little bit different, but that's why I like the simple framework of quality efficiency and employee experience. Cause you can almost run that through any department. Yeah, no, I think that's the right way to look at it. I get pretty critical of folks

[00:19:17] who are just focusing on efficiency. It's, you know, are you doing the wrong things faster? That doesn't sound so good. And, and you're also going to get diminishing returns, efficiency, right? Unless you plan on getting rid of all, all the humans, which I hope most organizations are not thinking about. And that won't be such a good thing, but I also appreciate, you know, everyone basically contributing to, to innovation, right? To responsible innovation.

[00:19:46] And whether that's new revenue generating opportunities, entering new markets, business model innovation, or, you know, new ways to, you know, save costs and, and things like that, you know, those, those all have a place in, in a big organization. And so I just want people to see that there's ways for them to contribute that value, create that value for organizations. Yeah. And I'd say two things, again, going back to this idea of what can you build versus what should you

[00:20:15] build is I think it is the leader's job, the leader of whatever the function is to have a philosophy and a vision for what that work is going to be looking like. And then decide, you know, after they've identified what are the problems they're needing to solve, identify where the AI tools fit into that. So when I think about hiring, I believe it's my responsibility as the leader of talent to really narrate and define our hiring philosophy, the vision for what I want that to feel like, look like, sound like for,

[00:20:45] you know, for all of the people that are involved and then figure out, okay, where does it make sense for AI to be in terms of things that it can do better than humans can do so that then we can reallocate that time for humans to do the things that they're best able to do. Right. And if I don't hold that vision, then of all the citizen development that my team can do, they're just going to be building stuff all over the place. Right. So I'm really a big proponent of you can't fall privy to like, you can't fall to the level of where your tools are.

[00:21:15] You have to really have a vision. And even if the technology isn't there yet, eventually it will be. And then you'll be ready for it. Right. If I want something that does this particular thing in my workflow, and that's just not invented yet, I guarantee you it will be soon. And then we'll be ready for it. And we're building more of our workflows on a philosophy versus just chasing whatever AI tools are there. So I think that's really important in terms of like the should versus can. But then on the citizen development piece, you're right. Once that vision is very clear to your team and they know where they should be building,

[00:21:45] then we are such a huge proponent at Zapier of letting everybody build safely. And I believe like that's a combination of AI fluency and upskilling your team or hiring for it in some cases, but upskilling your team and then having safe governed citizen development. So people feel like they can build safely without messing up a system or overriding something or accidentally using PII and, you know, a workflow they're not supposed to. So you need to have those guardrails.

[00:22:12] And once you have an AI fluent team and you have clear guardrails, then you can kind of let people loose so that they can build an experiment within the confines of your vision. That's where I feel like my team has gotten the most traction because we're all kind of solving, even as we're all building together, we're all trying to solve the same problem. We may come at it at different angles and then we can kind of collaborate and compare, but it's very clear what we're trying to aim for. And so I think there's a responsibility there for the leader.

[00:22:40] I think there's a responsibility there for the governance, the upskilling, and you can, you know, whatever teams are involved in those, but that's, what's going to unlock, I think the next layer of citizen development, which then unlocks the next layer of transformation. You know, context matters so much and you have employees have this tacit knowledge that, that they know better than perhaps, you know, an early tenure, you know, job candidate, you know, how things need to, you know, operate.

[00:23:09] And they might know, might have a better sense of what should stay on a human's plate. And the reasons behind that versus someone new that might come in with, you know, eyes wide open, they're a sponge and they want to learn. And they've already proven that they're on this, the trajectory that you look for in terms of their adaptability, their durable skills, their curiosity, systems thinking, things of that nature. But they don't, they don't yet have that,

[00:23:37] that tacit knowledge. So I'm just curious if the, if the rubric changes somehow or how you evaluate it somehow, do you pair these people up when they do get onboarded and in some type of mentorship? So you've got the, you know, that balance, like, I know that was a lot, but I just wanted to add some context to that. So couple of things. I mean, we don't have a different rubric for our internal workforce. In fact, I would say they're very closely related.

[00:24:07] So while we have a very clear rubric for hiring, all of that is built off of the data and the evidence of what was already happening internally. So we have, you know, when we do performance management, we have what we call impact behaviors and AI fluency and the ways that we use AI fluency are sort of embedded into there. So we don't, you know, evaluate people and say capable, adoptive, transformative and a performance evaluation, but it's actually all very much embedded in our nine different impact behaviors that we have. So it is very much the same thing,

[00:24:36] but also they feed on each other. So the reason why we decided to do a version two of the AI fluency rubric this past spring is because the level of AI fluency internally had risen so significantly. And so to come in as a new hire, we would want you to come in where the average Zapien is or better. And so because the level of what we were able to do internally rose across the company, we had to change the rubric to meet it. And so when new hires are coming in,

[00:25:05] we capable now is operating at the level of the average Zapien, being able to use AI to measurably improve your work, operate at a meaningfully higher level, you know, start to like point to real ROI, real experiments, you know, orchestrating your workflows. That is what Zapiens are doing now. And so it's, it's very parallel, even if we're not using the exact same rating system for each of those different populations. I'm just really appreciating the, the sort of human centric nature of,

[00:25:35] of the transformation that you're driving because I see a lot of organizations and they talk about, you might send out an assessment or try to gauge someone's, you know, maturity or whatever. And they're taking a very top down sort of organizational lens to it. Like, are we like, you are not, your readiness is not your ability to fund a technology investment. Your readiness is not, Oh, we, we deployed, you know,

[00:26:05] co-pilot to the entire organization, you know, job well done, you know, mission accomplished. I mean, these are, these are silly things to think about because I contend that like, if your people aren't ready, all bets are off. And, and your AI investment, your people are part of your, the portfolio of your AI investments. And if you don't see that, if you're just trying to automate as much as possible and remove human,

[00:26:35] humans, human judgment, and all our, all that we bring to organizational decisions, if your goal is to just, you know, keep removing more and more people, I don't know where you think you're going, but I don't want to work there. This is actually a very real live conversation we're having with our talent team, because I think the question that companies get to when they have automated a lot of the busy work and the toil is, okay, well now what do we do with that extra time?

[00:27:04] Like what, okay, what does that go into? What does that change? Right? Otherwise it's just, oh, well now you can just take on more of the same because you can do it all faster. And so I think that that's a really interesting conversation. And so what we're doing within our talent team, for example, is we're trying to start to move our recruiters who are brilliant recruiters, you know, specifically, you know, we have a really tenured team, but moving them to more what we call talent advisors. And that's a really buzzy term, I think in the TA world right now.

[00:27:33] But what we mean by that is they should be pouring more of their time and energy into the human relationships and specifically the coaching aspect of things. So coaching hiring managers, being able to use AI to pull in a ton of talent market insights and be able to say, okay, now let me make sense of that for you and advise you on what we do with that. Right? Before pulling all those talent insights would typically take a while. AI makes that so easy, so fast, almost instantaneous, but then it's up to the recruiter to interpret that, understand the business.

[00:28:03] They have to have strong business acumen about your company. They have to understand the market and then they have to be able to coach and advise a hiring manager. Similarly, coach and advise interviewers on how to get the best signal. Coach and advise candidates on, you know, not only how to prep for an interview, but is this the right job for you? Does this feed on your motivations? Does it address your concerns? Is it aligned with your values? Right? Those are different conversations than, hey, let's make sure you get ready for the interview and like, let's get you scheduled.

[00:28:32] And I think the best recruiters have always leaned into the more human aspect of the work. And what we're trying to do is give them significantly more time to do those things. Those are the reasons they got into the job. And those are the reasons why they're effective and brilliant. Right? So if we can say, you know what? The research can be done for you. The reminders about scorecards can be done for you. The tracking can be done for you. The reporting can be done for you. Those are all like busy work things. They're not best positioned to do. The coaching,

[00:29:01] the investment in people, the advising, the relationship building. Those are the human things. And so when that vision, again, going back to vision, when that vision is really clear to the team, I think it's one, very motivating, but it also then allows us to say, have a clear answer to like, what are we doing with the time that we're getting back? And how is that making us more impactful? I don't want to say, just say more effective, but more impactful as a team on the business and on the goals that we have.

[00:29:28] I've talked to many folks around what that reinvestment should look like. It's sort of a dilemma that some organizations haven't quite figured out, but yeah, how much more can you put on people's plates? There's a purpose to, you know, downtime. There's a purpose to some of the relationships that are behind the scenes of those workflows.

[00:29:53] There's other things that make the organization run and build the culture and build relationships and strengthen all of that. And so to discount that and just replace it with more, you know, cognitive load is just going to backfire in terms of, you know, burnout. And we are not machines. I feel like we're still trying to figure out what does it mean to be human? What does it mean to be, you know,

[00:30:22] machine intelligence and robotics and things like that. Sometimes we act robotic in the things that we do, just as machines try to act human sometimes. And so I think, you know, we're still, it is still very early as we're trying to navigate this, but there's no doubt. I mean, you alluded to it before. I mean, this is, AI is part of the evolution of work itself and modern life. So we've got to embrace it.

[00:30:51] There was something someone said at a conference I was recently at, and I love this sentiment, and I don't remember where they got it from, but they said their company's philosophy is to keep what you kill. So if you are finding more efficiencies in your work and say, you are doing, you know, your 40 hour work week through AI and automation, you've saved 10 hours because you become so efficient. You get those 10 hours. We're not asking you to do more. We're saying that's, that's your time. That's,

[00:31:21] you've earned it. You get to keep it. So I kind of like that philosophy. I don't know how many companies are going to go that direction, but it's a nice thought, right? To think through if we're becoming more efficient and impactful in our work, what do we do with that? And does it give us time to do the more human stuff, both the work human stuff, but the personal human stuff too. So I liked that philosophy. I want to think about that a little bit more. Yeah. I mean, we, I remember when I was at NBC, we went through that and I was working on, you know,

[00:31:48] automation strategy and I was pitching all this savings that we expected. And someone put it in, you know, financial terms. Oh, we can get rid of, you know, 50 headcount. And I was like, well, what are you? No, no, no, no, no. What? That's not the right question. The question should be, what else could those 50 people be doing? Yeah, exactly. What are the new abilities? I think that's such an important piece, Bob, because I think there's, again,

[00:32:17] a lot of companies who are thinking about cost optimization, which to be fair, companies do have to think about like they need to survive. So I understand it, but I think the, the much more advantageous position to be in is when you can think about what does this unlock for us? What does this unlock for our customers? What more possibilities are out there? And so, yeah, I, I'm just, I was just nodding my head vigorously with that because I think that's something that Zapier does well. We didn't think about, Oh, we're, we're such an efficient company. Now we can cut more people, but we thought about what can this unlock for our customers,

[00:32:47] for ourselves, for the work that we do. And what are the new possibilities out there? We haven't even, we might not have ever been able to do without AI. I think that's a much more exciting world to live in. Oh, for sure. I think some of it is like, there's always more to do, right? I mean, we're entering strategic planning season. I know it was, you know, must, you know, the must do's, the should do's, and then the nice to do's and the nice to do's had no,

[00:33:15] no shot at making the budget. So it's the should do's. How deep can you now go into the should do pile of projects and initiatives? How deep can you now go with the capacity that you've, you know, freed up? And so, so I'm not saying you shouldn't bank some of the savings. I mean, these were public fortunate 500 companies. So they've got to show, you know, market, you know, improvement and, and the key metrics,

[00:33:44] but how much are you reinvesting in the business? And then a subset of that should be, you know, the, the people side of, of that equation. And so, and that's just the things that you can name that are already on your, on your list, right? To your point, there's plenty of other things. Like what else could we be doing? And that's once again, where you tap into, you know, the, the collective intelligence of the collective human intelligence, perhaps augmented by AI, but where else could we, these,

[00:34:14] you know, spending our, our money, our time. And, and what is that investment? You know, what does that innovation portfolio now look like? Because the dynamics of, have completely changed, but just to cut people because you can, and not think that that's going to have longer term, you know, implications is myopic at best. I mean, you said it so well, I couldn't agree with you more. And I think we've been hearing for several years now, companies saying, because of AI,

[00:34:44] we're now doing this massive headcount reduction. And I think it's less that, from what I've dug into it, it's less that they are actually creating all this incredible ROI through AI at this point. Because I think a lot of companies are still struggling to do that. It's a very small percentage of companies that are seeing real ROI with AI right now. That's a lot of acronyms. But what they're saying is, is, you know, we want a different type of worker. So companies are, you know, reducing a lot of headcount,

[00:35:12] but then hiring back people that are more AI fluent. They're trying to redesign their companies. But I think to say it's because of AI, we're cutting all this headcount is like really distracting and it's not quite accurate. What's happening is AI is changing the business all around us. And they're realizing, gosh, we need to compete in different ways. We need maybe different talent, maybe different priorities, goals, et cetera. We need to kind of revamp from the inside. But I think to just like call it because of AI is really lazy.

[00:35:41] That's not the only thing. Yeah. I don't know if I would call it, AI is involved, but you know, you're conflating, you know, sort of the direct impact versus indirect. And so actually I think there's some, a different McKinsey report. There's, I think their latest state of AI report talks about that a little bit. Like the, the, the head, the headline grabbing, you know, stats do not reflect reality. It's really, I think it's less than half, maybe a third of the reported,

[00:36:12] you know, workforce impacts, negative workforce impacts from AI, only about a third. There's, there's truly that. And the companies that are investing in AI for, for growth, you know, to augment humans are the ones that are seeing outsized returns. Yeah. And their, their companies are growing too. I mean, so I think there's, you know, just even anecdotally in the market, there's still a lot of recruiters being hired.

[00:36:39] There's still a lot of hiring that's being done specifically in tech and some of the most AI forward companies. And so people are realizing, like, we still need humans to do a lot of this work. They're just doing different jobs than maybe what they were doing before. Well, I think that also ties to, you know, I know what you, what you've talked about, like you've called it slope, slope over scorecard, right? Like you're, you're not just hiring someone for the role as it's described in the job description, literally. I mean, but that job,

[00:37:08] the job you're hiring for could change in months, right? So you've got to, you know, you've got to see the, the skills and that adaptability, and you've got to see all those things that are going to, you know, make them sort of thrive throughout the, the changing nature of, of the work. So we call that slope over snapshot. So very close. And yeah, the philosophy behind it is we're not hiring people for just what they know today.

[00:37:36] We want to hire people for the trajectory at which they climb. And so when we rebuilt the AI fluency rubric, this version two, we, we tried to be very specific about that because the compression story about learning is real. Like people are saying that learning cycles that used to take 18 months are now running closer to three. And I use this phrase that I said earlier is, how do people learn hard things fast? Because we don't know what the next hard thing to learn is. And we know that this technology is continuing to change.

[00:38:03] So what we're hiring for are not only people that can be adaptable and flexible and can evolve, which is what companies have been saying and hiring for, for a long time. But specifically, I want to understand what is their trajectory been using AI as a proxy. So in an interview, we're not asking what tools are you using today and how are you using them? Solely we're saying, okay, how are you using AI six months ago versus today? And what changed in between? And, and what did you do in between those six months? And then even within the hiring process, for example,

[00:38:33] we give people a lot of support, a lot of resources to continue to hone their AI skills, especially if they might not quite meet their bar, our bar at the very beginning of their process. And then we kind of see how they learn and grow even throughout the hiring process, which is a very interesting meta signal of how they're going to learn and grow once they get here because somebody who's used, you know, the same three tools for the past three years tells me less about what they're going to be able to do than somebody who is like very visibly and

[00:39:01] specifically evolving the way that they use AI and they're evolving with the landscape too. So that's a really important piece to us. And I think it, you know, going back to my roots as a teacher, Bob, it's very much aligned with how we think about learning and growth mindset and development and upskilling in the first place is it's less about where you are today, but what is the potential that you have to grow and what is your motivation to do so? And that's really, I think the people that we're looking for in this moment. Yeah, no, that makes total sense.

[00:39:31] So I know you guys have been very open about some of the things that you're building inside your team, like a talent acquisition team. And I was curious if you could just give us like a quick little, you know, snapshot of some of those and some of the things that you've, you've learned over the course of, of building them. Yeah. Well, I can tell you the one I'm currently obsessed with that is in process of being built. So it's not, it's not done yet. It's still on the experiment side. So what we're trying to do,

[00:40:00] the problem that we're trying to solve is that recruiters spend so much of their time looking through many different platforms to get the information that they need, right? They might go to their ATS, then they might like, so we have Ashby, they might then need to look at transcripts from Brighthire. They might need to go back and reference something from Slack, a conversation they have with the manager. Maybe they took notes on granola and they might have, you know, so we have all of these different, we have so much data. We have so many different inputs. And as a remote company, everything is documented somewhere at Zapier, right?

[00:40:29] That's, that's actually a really good thing. I think that's what puts us ahead. The AI curve in some ways, but everything is somewhere. And there's so many things in a recruiter's day and workflow that they have to remember to do or do in a certain order and stuff like that. And so what we're trying to do is to really start to wrangle and control that environment for them so that they have one dedicated place where all their context layer lives, all of the skills and tools that they need to use live, and they can work through their day in one space.

[00:40:58] And so what we started doing, and we started this at Zapier a long time, about six to eight months ago, is we've moved everyone onto an agent harness. So everyone at Zapier uses codex, cursor, or cloud code to do the majority of our work. We also have an internal agent harness as well. The people team has chosen cloud code. And so every one of my recruiters works through cloud code to do a lot of their day. So what we're building on our side is what we're calling the TA workbench, which this is,

[00:41:26] is basically a place in the agent harness for our case cloud, where we have our entire shared company brain context intelligence layer is pulled into this harness. And then we have all of these different specific skills, like markdown skills that we have built for best practices through each part of our, our hiring process. So for example, there's a skill that helps a recruiter create an AI powered hiring plan with all of the information that they have from, you know,

[00:41:56] the job description and Ashby and the headcount plan and like all of that stuff. Right. Then we have another tool that kicks off and talent intelligence and market mapping. So as soon as a recruiter opens a role and they create their workbench, they have all the intelligence from the company. They're kicking off skills immediately to get started. Right. And then they're able to like take all that information and then do the human work of going to a hiring manager and actually starting to work with them on the role. Once they talk to the hiring manager, they're able to take that transcript from that conversation,

[00:42:26] feed it back into their agent harness. And now that's a part of the context for that role. So everything that they're doing is building on top of the intelligence that they're building throughout their process and pulling in intelligence from the company, the market, whatever at the same time. So I hope I explained that well. Basically what it is, is allowing people to have a workbench where they have skills, tools and context to do all of their work in one place, as well as, you know, lean on best practices and collaborate with others on the team too.

[00:42:55] I think that's an ideal sort of structure to make sure that what you're building is secure, it's responsible, it's scalable. So that people are, you know, I imagine you've embedded all your proprietary, you know, policies and all of that so that everyone is, is not going off and building their own, you know, sourcer and their own, you know, whatever, their own individual agents that may not be, you know,

[00:43:24] sort of operating in a consistent way. Right. Yeah. So we've, we've also been able to, because we have this structure, we're able to pull together things that are people are building into best practices. So let's say in our experimentation phase, we had a recruiter who was kind of building a, a hiring plan creator, and then another one who was building a hiring plan creator and another one, right? We've been able to pull those together to say, okay, what's the best practice here? What is, what is the best practice across the team from all the different perspectives of people building?

[00:43:53] And now let's make it a default practice that everyone can use. And yeah, we're going to improve it and iterate it over time. But right now we can say, if I had a brand new recruiter come in, I could actually hand them the skill and say, here's our best practice at this point in time of how we kick off a hiring plan. And then they are like ramped up 10 times faster than they would have been. Right. And they can again, focus on building the relationship with the hiring manager, spending time with candidates. Right. So it's not just using best practices

[00:44:22] and using AI to like make things that you do go faster, but it's also to accelerate your learning and to help wrap you up sooner. Right. To get all of the intelligence in one swoop versus having to go to a bunch of different platforms and try to figure out the answer to a question that you might have about the company. So trying to pull that together is the goal. And again, I'm telling you, Bob, it is not out in the wild yet. We are still building it. We are still experimenting. So everything I'm saying now could change as we iterate and get more into it too. Well, it sounds super valuable.

[00:44:52] I was also sort of mentally tying it back to what you said before about, you know, improving ourselves. Trent Cotton and I talked about this. Like he was really emphatic about why, forget about AI for efficiency and, and, you know, off, offloading your, your, your own sort of judgment. Why aren't we looking at it? You know, so in your example, looking at the bright hire transcripts, how can we, how coach people to say, well, why, when we look at,

[00:45:21] at quality of hire or retention or whatever the metric might be, why aren't we looking back and saying, let's, let's do a, you know, the trace of have the traceability to say, what, what is this person, this interviewer doing differently? That's either better or worse than their peers. Why does everyone, this person interviewed stay much longer than, than everybody else? Like, what are we, what can we learn from this? And that's another way that AI can sort of augment us to,

[00:45:49] to increase our own potential, whether that's for our own performance or to make sure that we, that we are benefiting the organization and the team by hiring better quality people. And I'll just say, that's a perfect example of something that a recruiter would not have had time to do right. And their normal work day is to go through, let me rewatch every interview and compare that against like this other person set of interviews, right? Like those are just the things that we would never have thought that we

[00:46:17] should be able to spend time coaching on, but it's the things that could make a difference that can help your quality of hire and could help your company, you know, attract, evaluate and retain talent. So I love that example because it's so simple, but it's, again, this is something that we would never have been able to do without AI. And now we can utilize that knowledge to do something even better with it for a greater impact. Tracy, I want to be respectful of your time. Any final thoughts, advice for listeners out there about AI fluency and,

[00:46:47] you know, finding the right talent? Yeah. Gosh, there's so much to say and we could have gone on another hour. I'm sure. The advice that I always have outside of leaders need to own the vision for their customers is that something that you said at the very beginning of our conversation, Bob is, you know, it just takes time to learn AI. I think there are very few things that you can say like, oh, if you just do it more and take more time to do it, you'll get better. Maybe swimming is another one of those things. I'm not sure,

[00:47:15] but I really think that AI is one of it because experimentation, playing with it, using the tools, getting hands on keyboards, as we like to say at Zapier, it not only builds your confidence and understanding of the tools, but it can build your creativity as well. And I think there's no substitute for like the time that it takes to like build those things. So I encourage everyone to, if you're not doing it, get on it. And if you are a team leader, really helping your team find the time to do this within their workday is

[00:47:42] critical to actually moving your whole team and your company forward. Agree. Well said. All right, Tracy, thank you again so much for joining me. This has been great. Thank you everyone for listening. We will see you next time.