How Zapier made AI fluency part of the job from rubric to rollout

Welcome to today's webinar. Before we

get started, let's go over a few quick housekeeping items. First, yes, we've had a few people ask, this webinar is being recorded. You can access the

being recorded. You can access the recording at any time after we close by using the same link you used to join today. We'll also send you an email with

today. We'll also send you an email with the recording [music] link once the webinar has ended.

Second, for the best experience, we recommend using Google [music] Chrome.

If you run into any technical difficulties during the webcast, simply refresh your browser. That resolves

[music] most issues.

Third, we'll have time for a live Q&A after we've covered all of today's content. But don't wait until the end.

content. But don't wait until the end.

Feel free to submit your questions throughout the presentation using the questions tab. We'll prioritize the

questions tab. We'll prioritize the questions with the most up votes. So,

please post your questions there rather than in the chat. With so many people joining today, questions in the chat can sometimes get lost.

And [music] finally, be sure to check out the links tab. That's where you'll find all the resources, links, [music] and next steps we mentioned during the session. There are no silly questions

session. There are no silly questions here. We're all here to learn, [music]

here. We're all here to learn, [music] share ideas, and grow together. So,

don't hesitate to engage with your fellow attendees. [music] And now, let's

fellow attendees. [music] And now, let's hand it over to today's presenters.

Awesome. Hey, welcome everyone. Um, fun

fact, that's actually my voice when we record videos. No, it's not. It's an AI

record videos. No, it's not. It's an AI voice. Um, very smooth. Hopefully you

voice. Um, very smooth. Hopefully you

liked it. That's kind of a new thing we're doing here. Um, thanks so much for tuning in today. Really excited about our session. Um, we have Colin and Tracy

our session. Um, we have Colin and Tracy here. We'll do intros in just a sec. Um,

here. We'll do intros in just a sec. Um,

but they have been instrumental in basically building out the scaffolding for how we turned kind of an executive AI mandate um, into real AI fluency that

we could measure and evaluate and act upon across our 800 person workforce here at Zapier. Um, and I think one really fun thing is I we have a really

neat tidy story here um to outline this.

Um, and in truth it was also very messy along the way. There was lots of iteration and evolution and all that.

And so, you know, lots of trying things and failing and finding the successes and blind spots. So, we're going to go through all of it. Um, but I think it's a really really cool thing. So,

hopefully you get a a kick out of of today and learn something new. Um, let's

do a little bit of intros here. I'm

Ryan. Uh, I am on the AI transformation marketing team. I host a bunch of

marketing team. I host a bunch of programs and events just like this. I've

been with Zapier for four years. Um, and

I live in in San Francisco. Colin, you

want to go next?

Hey everyone, I am Colin Monigan. I'm

the director of learning and development. Been here for eight years

development. Been here for eight years and calling in from Port Angeles, Washington.

Uh, hey Tracy.

Hey Colin. Um, hi everyone. I'm Tracy

Sek. I'm based here in Los Angeles and I've been at Zapier for just about four and a half years. I lead our global talent team which includes talent acquisition now onboarding, talent

branding and I focus a lot on how we're working together with the people team to ensure that we are building an an AI native team for the future. So excited

to be here.

So awesome. Um, well, welcome both.

Tracy, I think Colin, I think this is your first event that we've done outside of AI leaders lab. Tracy, you are a vet here. So, welcome back. [snorts]

here. So, welcome back. [snorts]

Yeah, go easy on me, everyone.

[laughter] Uh, nice. Well, listen, we have about an

Uh, nice. Well, listen, we have about an hour here and so just high level, I think the way that we will talk about these things and this is probably really familiar for lots of folks is like maybe

a couple times over the last few years, maybe since chat GBT3 came out. Um there

is like a lot of executive pressure to become an AI first company or AI native company or however you kind of explain it. And then what does the organization

it. And then what does the organization do? How do you kind of meet the urgency,

do? How do you kind of meet the urgency, meet the moment? Um, I think large context here at Zapier is that our talent team really picked up the baton

and built the scaffolding across the three layers you see here, and we'll dive into each one of those, hiring, onboarding, and um employee experience, i.e. like how do we upscale our existing

i.e. like how do we upscale our existing employees? Um, so we're going to go

employees? Um, so we're going to go through those those three areas and then we'll answer questions along the way. So

use the chat um and then the Q&A panel to make sure we definitely answer it and we'll save some time uh at the end. Um

okay. So

um I think this is like like I said this is uh something probably folks are familiar with right like leadership can declare the destination. Um they can

make AI a priority. They can say let's go further. Let's set the bar higher.

go further. Let's set the bar higher.

We've done that a few times here at Zapier. Um there are so many questions

Zapier. Um there are so many questions pro prerogatives priorities under that you know some qu common questions just to the right and so I think this is like

something uh to know I don't know Colin Tracy can you kind of like think through uh like maybe it's our code red or or what have you that kind of makes you

think about uh how how urgency showed up here at Zapier.

Sure. Yeah. I mean so many different stages to this right where there's kind of um organically people talking about it and using AI. Um I think you know for

me it was you know at Zabier I think fairly early on leadership was was pretty on the ball with saying hey there's something here we really need to

pay attention to u and there was this code red event which for us was was like hey let's take this seriously um this is changing the way that we work um for us

it was changing the way our product is positioned we needed to dig into this um and uh that's that's what stuck out to me where it kind of became like, oh, this is this is the real deal. This is

what everyone at the company is starting to talk about.

And just to be clear, not everybody loved the fact that we called Code Red.

[laughter] Let's just name that. Like,

for some people, it was, "All right, let's go. Let's do this. Let's put our

let's go. Let's do this. Let's put our energy here. I'm ready." And for some

energy here. I'm ready." And for some people, understandably, it was, "Okay, what does this mean for me? What does

this mean for my job? What does this mean for my day-to-day?" Okay. And so I actually love the questions you pulled out, Ryan, because we heard all of that.

Naturally, when there's any sort of change in an organization, people ask, okay, what does this mean for me? And so

we had to get really fast at answering those questions as soon as possible so that we didn't allow people to stay in this sort of purgatory of what does this mean,

right? And I I think like um the urgency

right? And I I think like um the urgency has come in a couple different ways.

There's the three buckets that we have here, but there was kind of almost like three obvious waves for us too. There's

the initial code red of like, hey, everyone should be going and experimenting with this like, you know, you've you've got it, go for it. Then

there was like, how do we raise the the fluency of the team? And we'll get into this a little bit later. Um, we've been really early across the org to adopt

agent harnesses. Um, and so it's kind of

agent harnesses. Um, and so it's kind of like those are three pushes. Um, and

y'all have really had to meet the moment to help the organization get there. Not

just the people inside Zapier that are like naturalborn tinkerers and naturalb born I'm going to try a new piece of software early on kind of thing. Um, so

I think this is this is all just building up to like the roll out here.

Um, and maybe you can talk, Colin, about like what the founder and executive urgency like un unlocked at Zapier. Um,

and what it what it didn't unlock for folks.

My my son is 12. He started playing club soccer, so we're at soccer tournaments all weekend. And if you've been to a

all weekend. And if you've been to a soccer match, the ones our the way ours work is you have the parents on one side of the field and then the coaches and the players on the other. And I was observing this dynamic where on the

parent side, which is me, the parents are just like enthusiastic like go GET THE GOAL, GO TO SCORE, like go pressure, pressure. Not always very helpful

pressure. Not always very helpful comments, but like the urgency is there.

Like there is excitement, there is motivation, um, you know, there is yelling, um, typically positive yelling. And I I kind of think of that as like how the

leadership helped at least at Zapier and how I see it working at other organizations where like boy it's great like you have to have leadership buyin.

You have to have leaders setting the tone showing this is important like showing the excitement cheering people on uh role modeling like using AI themselves. Um, but at the same time,

themselves. Um, but at the same time, like in that soccer match, I'm like, boy, if it was just parents yelling at a bunch of kids, but the kids didn't know the rules, they didn't have uniforms,

they didn't know strategy, they didn't have basic skills, like it would just be a mess, right? And so, I think the part that we're going to dig into here is like what are the coaches doing? what

are the what are the players doing at practice every day um to make sure like all that enthusiasm that the that the parents or the leaders are providing turns into like actual tangible like we

are playing the game we are scoring goals so it for us at AI and the work that a lot of Tracy and I were doing was okay what does it actually mean to

support all that enthusiasm and and role modeling that we're getting from leaders how do we set people up for success um and there's tons we can dig into there for for for from my point of view there

was pieces of that where uh a big chunk of it was just about like what I consider infrastructure which is basics of like do people have clear access to tools? Do they know how to use the

tools? Do they know how to use the tools? Um you you know just kind of like

tools? Um you you know just kind of like ba basic um access to things. Um and

then there's also this whole layer of like boy we have a lot to learn on top of that of like how to use this effectively. um you know how how are our

effectively. um you know how how are our peers using it? How does it get integrated into the way that our teams work? Um and there's a big you know

work? Um and there's a big you know Tracy you already kind of mentioned this like there's a big emotional component to that too or just like an identity aspect to this as we're all like learning and um you know I have mixed

feelings about it too. So those are kind of all the pieces where um you know the teams that Tracy and I are a part of have started to dig in to figure out like okay how do we start to kind of make this real throughout the

organization continue to like push push the standard in specific ways but also in a way that people are feeling supported the whole time.

Yeah, I I love those points, Colin, because I think it's kind of like it's so e easy that urgency unlocks permission. And then what it also does

permission. And then what it also does though is potentially push everyone to figure this out on their own. Go like

side of desk, work in, you know, after hours or before hours or whatever. Um,

but you know, is the organization learning or like you said, are we setting a standard? are we helping bring people to that? And so I think one of the first areas um that you know after

we had done some organizationwide experimenting lots of side of desk maybe a hackathon or two I feel like Tracy you really picked up um like let's name this

because we're interviewing lots of folks and bringing them in here um and hiring was like a really good place to start naming these things. So, um, can you

just talk a little bit about one like starting to build out like what our AI fluency rubric was and primarily then like employees can use that, but obviously we're going to measure

candidates against it as well.

Yeah. So, I would say just to give a little bit of context about why we felt like this was so important, we thought that to strengthen the overall fluency of the organization, we really did have

to do two things in parallel. So while

Colin and our chief people officer were very focused on upskilling our current workforce, it became clear to me very quickly that we had to also make sure

that every new hire sort of met or exceeded the level of the average worker at Zapier because that's how you raise AC the bar across the board. You can't

have our workforce kind of on this trajectory and then you're having to like start from scratch and upskill everybody as they join in. So if AI fluency was going to be a part of how

the company expects work to happen, we didn't want to introduce that to someone after they join when it it feels a little late because they may get overwhelmed. You know, you're throwing

overwhelmed. You know, you're throwing them in then into a situation where everyone's operating at a different level. And so that's what kind of pushed

level. And so that's what kind of pushed us to say we have to meet in the middle.

You all I was like callin our chief people officer, you all start with our current Zapians. I will start at the

current Zapians. I will start at the beginning of the employee journey and we're going to meet in the middle and that's how we we overall raise the bar.

Um, and so what you're showing right here, Ryan, is the the artifact itself.

This was I think this is actually our second version of our AI fluency rubric.

But what I really like about this, and I I'll share a little bit more about the components and durable behaviors in a second, but we describe fluency at Zapier at three levels that we've really

defined and we've also transparently shared with candidates. So just to go over them briefly, capable is, you know, I use AI to operate at a meaningfully higher level. I use it in my work. I it

higher level. I use it in my work. I it

makes me a stronger performer. And

here's specific and concrete ways I can explain that and demonstrate that.

Adoptive for us is starting to really orchestrate AI and build systems that elevate how people work. So thinking

about workflows, moving, you know, workflows through multiple tools, potentially building things out for teams that are adopted, and then transformative is um the golden goose. I

will I will say that transformative didn't really change between version one and version two of this rubric because um it is it is really hard to do. It is

really hard to meet and that is specifically re-engineering how work happens. Um so that is how we think

happens. Um so that is how we think about AI fluency. Our bar for being at Zapier is capable. Um and then we also you know have have a little bit more of different ways that we assess it too

which I can get into in a second. But we

also spent time thinking about this for every department and every like very specific type of role. The rubric alone would have stayed abstract. But the

translation layer being able to say, okay, what does unacceptable, capable, adaptive, transformative, what does that mean for marketing? What does that mean for engineering? What does that mean for

for engineering? What does that mean for the people team? um it looks really different in those different organizations or those different departments, but some of the things that are the same are what really made up the

bar. Some of the behaviors that we could

bar. Some of the behaviors that we could actually observe in the hiring process is what we focused on to keep the standard consistent and then we adapted that evidence to the role. So, um I'm

it's been really amazing to see how it's played out because we also built it in uh partnership with our executives and with the functions themselves. Yeah,

it's it's I mean it's really cool. I

remember my mind was kind of getting blown when y'all took it and built this out and you know there's obviously nothing fancy about a a really nice rubric and and document, but when you

read through it, it's really really thoughtful. Um it's really pragmatic and

thoughtful. Um it's really pragmatic and you can see how as a manager or as a hiring as a recruiter, you could read through this and you could evaluate someone's fluency or just skills, even

interests, right? you can start to gauge

interests, right? you can start to gauge those things. Um Cameron here in the

those things. Um Cameron here in the chat is uh you you're dead on Cameron.

We have a builder culture here at Zapier by the nature of our product. So there's

kind of this like internal builder culture at at Zapier. Um I you know my answer to like where are you finding highly affluent people um is kind of

like every everywhere you know like it's um you know it they kind of sprout out like like flowers. But I don't know if you have an answer to that, Colin or Tracy. Where where are we seeing people

Tracy. Where where are we seeing people sprout up?

Well, I agree with you. I think that this can be found everywhere at every level of experience and every function.

There are people who are leaning in, experimenting, and tinkering with AI either as a part of their work or just because they're interested in it and they're doing it for for personal reasons too. So, I agree with that. I

reasons too. So, I agree with that. I

will also say that as we raised the level of rigor for our bar for AI fluency at Zapier, we also raised the level of support for our candidates as

well. So we gave a lot of resources as

well. So we gave a lot of resources as soon as somebody applies at Zapier. We

explain AI fluency is really important to us. Here's why we're, you know,

to us. Here's why we're, you know, striving to be AI native company, etc. This is why we feel like it's important.

And if you don't feel like you have, you know, really tinkered with AI or you have built your level of AI fluency that would meet our bar, here are resources you can use. Some were Zapier resources.

We pointed people to courses with Open AI and Enthropic. So we did that very early on in the process so that candidates would have the chance to actually jump in and start to self-e

even throughout our hiring process. So

we felt really strongly raise the bar but raise the level of support to meet it. And so we hope that even if people

it. And so we hope that even if people don't ultimately end up joining Zapier, they also feel like our hiring process is a place where they learn something along the way and they've even improved their own skills, whether it's for a job

with us or a job elsewhere, too.

Yeah, that's a that's a great point. I

think a reason why I love this this accomplishment for the team that you've you've reached 100% evaluation of all candidates um being measured against the

AI fluency rubric. Tracy, do you want to talk a little bit about getting there?

you kind of touched on it already a little bit.

Yeah, we I mean the the data is clear.

Every single person who goes through our hiring process is is evaluated on this at multiple points in the funnel, too.

Um and so we we talk about it super transparently. Again, like I said, we

transparently. Again, like I said, we share supports, but we also do this because we want to share the clear signal and expectation about the environment that they're going to jump into. because it's not just hiring as

into. because it's not just hiring as the sort of the gatekeeper for joining Zapier. It's hiring as a way to set the

Zapier. It's hiring as a way to set the expectation for what we're going to expect from people once they get here specifically around AI. And so we want to set up people for success and know

that when you're going to be here, you're going to be building on day one and this is how we operate. So I think the important thing here just to note is like yes, it was important to get this right for hiring, but we didn't use this

as a hiring rubric for hiring in isolation. It was really the first part

isolation. It was really the first part of the employee journey to set the stage. And so we we did this in concert

stage. And so we we did this in concert with onboarding, with performance expectations. You know, hiring is one

expectations. You know, hiring is one layer of the system, but it's not a substitute for development. We don't

just hire people at a bar and then expect them to be great. And that's sort of where Colin's team came in when we when we did that handoff.

Totally. Um, and we'll we'll get to the handoff. One thing I do want to

handoff. One thing I do want to highlight before we we get there and this obviously flows through is the thoughtfulness I was talking about is like there are these components to fluency. So you don't have to come in

fluency. So you don't have to come in and be a total wizard that's reaching the transformative bar of our fluency framework. It's kind of like let's look

framework. It's kind of like let's look at these components um yeah through and I can walk through them because what I I hope people take away from this is it's really holistic and we hope more

more durable than just any given snapshot in time. Right. A lot of people think of AI fluency and think, oh, can they use these tools? And we're very, very particular to say it's not just

about being able to build with AI tools.

So that is, as you can see, a part of it. So I'll start from the beginning. So

it. So I'll start from the beginning. So

we really evaluate mindset early on.

That's specifically a curiosity, a willingness to iterate, to experiment, to fail fast, to revisit how work gets done. And then strategy or is really the

done. And then strategy or is really the strategic acumen of um somebody how they think about their function. So how are they thinking about the future of work in recruiting or accounting or

engineering? Are they starting to piece

engineering? Are they starting to piece together how they could imagine an AI native world in their function and start to move towards that understanding you know what the outcome should look like

and also deciding where AI should belong. building. Again, fairly obvious.

belong. building. Again, fairly obvious.

We lean on some of the components of Anthropics AI fluency index. So, we're

really looking for things like people using AI as a thought partner, iterating, um, producing something useful, and that goes into accountability and discernment, which is ensuring people understand what quality

is, making sure people are responsible for the result. Um, they know what good looks like. They are not just popping

looks like. They are not just popping something in and accepting it for for what it is, right? So being very discerning, having taste related to AI um is something that's really important.

And so um a couple other things to call out is these are the components, but there's also some important ways about how we think about this holistically. So

we want to really ensure that we're looking not just where a candidate is on the day that they they apply, but how they've also really grown in their AI

fluency over time. So how has their practice changed over time? maybe the

six months before they get to us and then also how are they learning and growing with AI even throughout the four to five week hiring process. So we call this looking at the slope and not the snapshot. That gives us a much better

snapshot. That gives us a much better signal of how they're going to continue to grow at Zapier versus just like I know how to use this tool today and I'm I'm really really good at it. Um so

that's one thing to call out. And the

other piece that I think is really important especially in in V2 is that because this is such a big part of how we work, manager roles carry this additional requirement because we want

you to demonstrate how you have created the conditions for um your team to thrive in this new AI world, right? It's

like how have you supported psychological safety and space to experiment and change management and clear expectations and things like that.

So, so cool. And I think this kind of gets to the next thing, Colin, as we kind of transition into onboarding here.

Y'all were thinking about these two areas alongside each other. They weren't

necessarily totally isolated different talent initiatives. Could you talk about

talent initiatives. Could you talk about that as we kind of get into onboarding here?

For sure. Um, yeah, I mean, we're always, it's really interesting for us how this developed is, um, you know, at the time my team was in charge of new hireer onboarding. Tracy's team was

hireer onboarding. Tracy's team was doing hiring and we were actually kind of thinking along the same lines about um on the onboarding side like oh we need to there's new tools available there's new ways of thinking there's

whole new ways of working like we need to we people are coming into the company and we need to make sure that we're enabling them right away. Um Tracy did a great job of um you know communicating

with us about like hey we're also thinking about that on the hiring side and the benefit of what her team is doing with hiring is they have a really clear sense of who they're bringing in and kind of like what their strengths

are and what we're selecting for so that in the onboarding program then we kind of we had a better sense of who we are getting and in this case in a lot of ways like the bar was raised and so we knew these folks were coming in with

these like baseline abilities of those those different aspect of AI that Tracy was just walking through. And so we could kind of like start at a higher level. Um it's always easier anytime

level. Um it's always easier anytime you're building, you know, training or enablement or any type of experience for people um to kind of like have a baseline understanding of where they're

coming from. And that's what this gave

coming from. And that's what this gave us. And so we're coming in, we already

us. And so we're coming in, we already knew that they had some of these these skills, some of these mindsets. we were

able to just kind of basically build experiences that helped um you know just be consistent like continue the momentum and then helped kind of introduce them to how that works here at Zapier, the

tools that we have available. Um and

then I'll get into this more later but just for some context too our onboarding program is cohortbased.

Um so people are coming in as a group and it just they could be from different functions and different teams. Um, but they're working together and and the way we did a lot of our new hire enablement

for AI is in a live group setting building together. Um, and that's

building together. Um, and that's something I'm I'm just going to keep coming back to as a recommendation for we just found so much value in people learning from each other. Uh but also

just kind of the social um I talked a little bit about like the identity aspect before where people feel safe trying things um together trying things

that don't work you know together seeing seeing uh what what you can experiment with and what you learn from that. Um

that was a really important part of that new hire experience as well so that we could just like introduce people into the company into not just like hey here's the tools go use AI but it's like

hey here's how we build together here's how we can learn from each other uh to kind of keep that journey going over their their full employee experience.

Yeah, I I I really like one of the things, you know, as as an employee that you were thinking about, you know, and the experience that I had and kind of

when I zoom out and think about what y'all created over time was obviously like the framework, um, the rubric created an an

actual like named expectation for everybody of kind of like, all right, here's like the standards. And those

standards have evolved over time, but it's been always been very clear when we're kind of like pushing the ball forward. Then Colin, I know you were at

forward. Then Colin, I know you were at the center of a lot of gruesome work which was like access [laughter] and like just figuring out like th those

sorts of things where there was tons of friction for people and and we'll get into this a little bit but like you have a marketer setting up an agent harness for the first time which is sometimes

like setting up a developer environment like that is so much friction it is going to completely snuff learning and growth in the process. And so it kind of like rooted those things out and you

started to build paths and golden paths and then like you said the practice and so you know there's there's a really nice method to the mess uh to the madness here that I think you you

brought in. That was my my experience.

brought in. That was my my experience.

I'm sure it felt um less clean to you as you were orchestrating a lot of these things. Um, but it was actually like

things. Um, but it was actually like very very helpful for employees and I think in the end you're just trying to like raise the floor in so many ways because like who knows what the ceiling

is for a lot of people as well.

Yeah, absolutely. I I think that's a great way to put it. I also think about it as just removing friction that gets in the way of people. So like you know

again if we're hiring people that we know are already um interested in AI, we know that they're getting here and they're already motivated. we don't need to motivate people, but it became apparent that there's stuff that can get

in the way like access to tools and, you know, security guidelines and things like that. So we can talk about some of

like that. So we can talk about some of that but there's definitely just like baseline there where I see as part of our responsibility is is like training on the ceiling but absolutely some of it

is like oh boy we just need to get that that baseline infrastructure set so that um you know just get out of their way and remove stuff that uh it might be a barrier to them jumping in.

Right. This might be a totally duh thing to the folks on the call here, but the thing that was all like illuminating to me is y'all picked hiring and onboarding because those are really structured um

parts of of a employee or future employees process. You know, once you're

employees process. You know, once you're off in the org, you know, you're off in the org. There's almost like a

the org. There's almost like a wilderness aspect to it in that, right?

People have their functions, they're busy, they have all their thing, but these are actually processes that you can really influence with this. And so I think that's like a highle

recommendation to everyone is like these are great places to start because you have the best ability to influence um people and their AI journey at the

company ultimately.

Um, well, we talked about the paths. Um,

Colin, let's dig into something that I don't know if this came from somewhere else or uh, you know, you introduced it, but you have written and helped the organization with a lot of golden paths.

Um, can you just talk about what these are, how they work here at Zapier?

Yeah, absolutely. I mean, it's it's really it's a version of what I was just saying, which is removing friction. Um and and for us a

removing friction. Um and and for us a golden path basically just describes um a recommended way to get set up in this

case on a particular AI tool. Um and so you know it doesn't mean you can't diverge from that. It doesn't mean like you can't build on that or it might be customized in some way for your team.

But like you were just saying Ryan, at least during like initial onboarding and setup, like we want to get some momentum like to get set set up on the right foot and um not have people like reinvent the

wheel every time about getting set up on things. So um I can show you should I

things. So um I can show you should I show our um some of the actual guidance here?

Yeah, let's do it. And while you pull up the screen, I'll give a little personal anecdote. So um we started using cursor

anecdote. So um we started using cursor a lot internally and I like adopted it early after seeing a colleague show me something really cool and I was like oh I want to do that too. So I attempted to

set up cursor by myself you know without any guidance and I was kind of you know having a good time. I was learning some new things and then Colin came up with

this our our golden path here and I reset up cursor and then I was like having the best time like I had like a lot of the friction I was feeling my cursor just wasn't set up in the right

way and Colin here really actually helped me clean that out and I was off and running afterwards. So I'll let you you take over here Colin.

Yeah. So this is the context is we had many phases to our roll out but at one point we had a lot of folks getting set up on cursor. Um and it was more complicated than people were used to

especially if you're not an engineer.

We're asking people um you know if you're not familiar with cursor they you might you're setting up um a local file that you're pulling from with a bunch of context in it. It was the first time a

lot of people got into skills and shared skills and how that worked. We're having

people connect to our MCP at Zapier so they could they could communicate with other tools effectively. Um we use Glean internally so we're connecting to Glean

we connecting to data bricks. There's

like so much that once you got set up to your point Ryan like people were were like great this just works. Um but there was a lot there's a good example of a lot of friction for setup friction and

uh I know many of you are probably in this now where there's just like these like pretty tangible things that are getting in the way. So um this you know this was like a project of love for for

many days that I had to update all the time because also these tools are getting updated and my you know my screenshots would get out of date. But

there's a bunch of different pieces to this that um both are existing employees and then all the new hires were going through to set up cursor for the first

time and it's just like super stepbystep screenshots. Click this, go here, set up

screenshots. Click this, go here, set up this, connect here. Um I, you know, it's not ideal that you would need this much instruction, right? Like ideally these

instruction, right? Like ideally these tools just get better and better, which they have been, but at the time we are like, we want to jump into this. And

again, what do we need to do to set people up for success and reduce any friction that they might be experiencing? And so people go in and

experiencing? And so people go in and they would they would set this up. new

hires were doing this on their first day at Zapier um because we wanted them to get access to it um and make sure that they could start building on their first day using this tool. And so this was a

lot of the So this is all just getting set up. There's also other components of

set up. There's also other components of this which I can build on later, but I'd also created all these getting started guides which were really set up so um

people in teams could build some activities that are more customized for their teams. Um but I basically gave them a template to start with um of

activities of exercises actually using the tool to walk through that would kind of force them to use different features. So we wanted to make

different features. So we wanted to make sure they understand that just how to use the chat and different models. We

wanted them to understand how to use skills. We wanted them to use MCP. So we

skills. We wanted them to use MCP. So we

had a bunch of exercises here too that we scoped out of like walking through all of those. Um, and like I said, a lot of teams would would take this and build

on that or run their own sessions with with these as uh inspiration.

Yeah. And I think there's a question here from from Jerry. This is also a good um transition into kind of like the employee experience side of this. Um,

obviously there are a certain number of new hires every month that made this worth it, but this is also for all all employees um too. And I think the cool part is the

um too. And I think the cool part is the alternative to this is you're going through generic cursor docs, you know, which not everyone has the appetite for

and it's not always as helpful, right?

So, this is actually really dialed for Zapier's environment, which I think makes it, you know, 5x more useful to you as an employer, a new hire. And it

just required ultimately Colin taking the time and thoughtfulness to really break it down and provide all these resources for folks.

Yeah. I don't I don't think I would have done it if it was just for new hires. So

it was for every we are all I mean this is one of the other things just in general that we always try to do with our new hires is we don't treat that as a we don't want to create resources that are just for new hires that are not for

everybody else. It should all just be

everybody else. It should all just be the same. So when you join Zap year like

the same. So when you join Zap year like you're using the same resources and going through the same you know experiences that everybody is. So I

think that works really well in this case with AI too where you just we give you a little bit more of a curated experience when you're new. We walk you through it and you know we kind of help you manage your time a little bit more

because you're still figuring out how to do that. But the resources are the same.

do that. But the resources are the same.

And so in almost every case, things like this, any any artifacts we created, any supports we created, it's it's not it's for everybody, including new hires. It's

not just for new hires or not just for existing employees. Um, and you'll see

existing employees. Um, and you'll see that with as we've talked about the rubric and different models that we have. It's the same for that, too. That

have. It's the same for that, too. That

that consistency has worked really well.

And I think, you know, like so much is new here and changing all the time. I

think one of the unsaid things that makes this fun, hard, exciting, challenging, all those things is um what you're kind of talking about Colin is

like there is this convergence and divergence like loop that you keep going through. So you know you might be asking

through. So you know you might be asking oh why was Zapier adopting agent harnesses? Well, the answer was as

harnesses? Well, the answer was as people were experimenting, folks that were getting like bigger gains were getting into these tools and doing new things. So, the thing that pulled me

things. So, the thing that pulled me into cursor is I watched my colleague Joe basically take our messaging and positioning and personas, pop it into cursor, grab all the HTML from the

website, and then write five new versions of our homepage for like five different roles that we would target here at Zapier. And I was just like,

whoa, that's like really different. Like

the old way we were doing things like we would, you know, do that all by hand or like we could, you know, we could never do that kind of thing. And so I think we're seeing some early adopters do that. And it's like, all right, well now

that. And it's like, all right, well now the challenge is can we bring everyone else along to where the the early adopters are going. So now we're going to have this convergence moment, right?

where you need to build these resources for everyone and kind of figure those things out on the fly and make them standardized for everyone.

Yeah, I I just want to plus one on that and it's important point where you know if you know we are as a people team we did not just kind of in a vacuum

like come up with a rubric or come up with this guide or come up because we actually didn't know right we were not the experts right and so there was absolutely the other thing that kind of happened for us with

that leadership urgency is there was a gap between the lead leadership urgency like do stuff and us rolling out a lot of the the resources that you're seeing

today. And that was like it could be

today. And that was like it could be stressful and you know there's there's downsides to that. But it was also the way that like we needed to learn from people like Joe you know and our

colleagues who were pushing into this about best practices about what was working about the best ways to do things about what was leading to success and what wasn't. And then we were kind of

what wasn't. And then we were kind of just trying to be like, okay, at what point do we start to coalesce this and say like, yeah, we think this is the best way to do it right now. Um, you

know, when I created this cursor guide, I had never used cursor before, right?

So, I was doing it on the fly and but I was leaning on people who had already been doing the work. There were some folks in marketing who had been setting up cursor who are not engineers and had already gone through the process

just because they're interested in this.

Yeah. like what it took and how they understood it and that was vital. Like I

don't know how I would have gotten that information. Um except for the fact that

information. Um except for the fact that they were experimenting with it. They

had been through it. They kind of became the subject matter experts on how this all works. And then teams like Tracy and

all works. And then teams like Tracy and I are we're just trying to be like at the companywide level, how do we start to build some centralized resourcing around this that starts to feel like

yeah, we're all kind of landing on these patterns as being helpful based on all the the wonderful work our employees have already been doing.

Yeah, I I love that point. Maybe I'll

lead you to your next tab [snorts] here too, which is like there are different ways of doing this. one is kind of like documentation like this, but I think I think Colin, you also have an HQ article

there of kind of how you take wins or failures from builders across the org and kind of turn them into like teachable moments and stuff like that for the organ. And that's another, you

know, divergent thing. Let's converge

it. Let's help the organization learn from people like on the front lines having success and failure.

Absolutely. I you know my stance with learning and development is we don't we haven't done a lot of like centrally organized training like we will now do a 10p part

course on how to use AI. I'm not saying that's worthless but it was it just seems really hard to figure out like how would you put that together in a way that doesn't become outdated the next week or in a way that's really relevant

for everyone. So, our approach has

for everyone. So, our approach has really been to try, like I said, to kind of like poke at where we're already seeing momentum and then to try to build some resources

uh that we could just share across the company and help people learn from each other. And this this is our this is a

other. And this this is our this is a company our intranet and this is like our homepage where I've tried to start to compile a lot of these resources. So,

we have like here's what our vision is.

Here's um some of those infrastructure pieces I was talking about before like what tools are available. How do you access those tools? At Zapper, we have access to basically everything. You can

use claude, you can use codeex, you can use um Gemini, you can use cursor, you know, everything. Um how do you choose?

know, everything. Um how do you choose?

How do you you know all those things? We

have documentation about that here. Um

we haven't talked too much yet about like security guidelines, privacy guidelines. That's a big deal when these

guidelines. That's a big deal when these tools are working with data sets and people had all sorts of questions about what can you do, what can you not do. Um

that was another area where for a while we were like we're not sure, you know, we had to have some learnings about like something going wrong and be like oh that's that's a privacy concern. We need

to put a guideline in place that we didn't used to have. And then we started to standardize those and post them in a place like this so that it becomes a bit more official. Uh we have some of our

more official. Uh we have some of our frameworks that we've already talked about here.

um example workflows and cataloges for people to look at. Um we did curate a lot of resources um that we didn't create but on that were publicly available and from other

groups if people wanted to learn more directly uh about AI that way. And then

we had things like um I talked before about how we like to emphasize peer-to-peer learning. And so we're just

peer-to-peer learning. And so we're just trying to point out places where this is already happening. Slack. We have a lot

already happening. Slack. We have a lot of Slack channels where people are sharing examples. Um we have some

sharing examples. Um we have some sessions that people have kind [snorts] of self-organized where they're sharing thoughts and insights with each other like uh engineering was doing this for

themselves. Um and so this is kind of

themselves. Um and so this is kind of our this is our hub for a lot of the information we try to keep updated. It's

another example of this is for current employees uh to look at as a reference.

It's also what new hires look at in their first week um to make sure that they're familiar with what's available to them.

So, um so cool to see all this. Well,

Colin, we are um I'm mindful of the time. What I'm going to do is actually

time. What I'm going to do is actually move us forward just slightly because I think like you know as we dig into employee experience um and if folks want to download the the PDF the slides here

if you hover over the slides you should see a little download button. I'm going

to just zoom a little forward because I think there is this um piece that's really important here. It's like what is this all in service of? Like you said, Tracy, you know, it is in service of of

growing the skills of our whole workforce from candidate all the way to tenur employee, but I think it's also in service of, you know, it's pretty easy

to change the way individuals work or they will like find ways to change the way they work. The stuff that actually this is like trying to set a foundation

for is like at the team level and at the business impact level like how do you like kind of ladder um all this skill building all this permission all this

energy all these things of of that and so you know I think one way you found value Tracy is like 100% of our candidates get evaluated column maybe

you want to kind of talk through just some ROIs some examples where we've had bigger hits and this is where we are. I

think a lot of companies are like how do you actually like prove the value, not just we're changing the way people are work, they're changing the way they work kind of thing.

Yeah, it's it's such a tough one, right?

Um, I think we're all trying to figure this out, but I will I will say like it was it was obvious early on that getting people individually to have increased

productivity is a great and crucial starting point, but is not the end game.

um that is in service ultimately of we think there's bigger impact that you can have across the company to like uncover new new value for our customers

essentially. Like there's things that

essentially. Like there's things that are probably possible now that were not possible before at a at a pretty high level that would make a a meaningful difference to our business's customers.

Like what what are those things? Um and

that's that is much harder than the personal productivity the personal productivity piece, right? Um, and I'll just I'll maybe I'll start by saying some of the barriers to that that we

found. Um, one is that um the incentives

found. Um, one is that um the incentives aren't always there. You know, it's like people are um it's not that people are inherently selfish, but like you're incentivized to

for your own productivity, right? And

there's not always a person who has kind of a broad view of the entire team or or different teams or there's fewer people in those roles who kind of like understand that broad picture and also

have the kind of technical knowhow to know maybe what's possible and wasn't possible before. Um so there's some like

possible before. Um so there's some like workflow redesign things that a lot of companies don't always have a muscle for that needs to kind of interact with with

the business impact piece. Um but um in order so so the main way that we tried to address that is we had folks we worked with what we called AI

transformation managers where which is we advocated to have representatives from kind of each of the major teams within our organization which were these AI transformation managers kind of

representatives in that team and that was their job. Their job was to um we we asked them to create a case study each quarter and the case study was essentially like

okay what is kind of like a higher team level solution that you can build that is not just about an individual but is like changing the way that your team is

is doing something in order to produce more value. And we were all kind of

more value. And we were all kind of doing that together because we're all learning from each other about like how do you do that? What are the barriers to that? like I don't even know what to do.

that? like I don't even know what to do.

What are you doing? Like what ideas do you have? I don't know what's possible.

you have? I don't know what's possible.

Um and so it was a messy process, but again, peer-to-peer learning. I would

say like we're really transparent about it. Let's learn together, make it a safe

it. Let's learn together, make it a safe place. Um and then we did have some like

place. Um and then we did have some like I do think like the authority was important to like have someone in charge of that who's like driving that work because left to their own devices, we found like people wouldn't necessarily

organize in that way. They would tend to be p, you know, individually focused.

Um, and so formal roles that were not just side up desk that were people thinking teamwide and then some transparency and like rituals to have them report on what are you doing? How's

it going? Like how can we learn from each other? Um, let's share insights.

each other? Um, let's share insights.

That's a great uh you're you're seeing the divergence convergence model just all in in that too. Tracy, anything you want to add on business impact before I think we can dive into just a couple

examples and then we'll we'll start to get on to any Q&A we have.

I'll just say one thing really briefly that was really amazing to see because obviously, you know, I'm tracking our new hires and I'm tracking their cohorts as they go through the onboarding

process and, you know, on to being like full full-on full-on Zapians.

That's so nice. It was it was [laughter] really it was really nice to see that people who were coming in at a higher level of AI fluency were able to jump in as Colin said on day one week one

building and already starting to solve pain points in their own roles. Like we

believe at Zapier that the people closest to the work are the ones that should solve their own pain points. And

so equipping people and empowering people with that from the very beginning with the golden with the the knowledge that they already had coming in and then coupled that with all the golden paths and all of the resources that Collins

team built, we saw really empower people very early on and much faster than than was typical before to start really reshaping the way that they work um in this new AI world. And so we gave them

the tools, we gave them the space to experiment. they came in with a level

experiment. they came in with a level of, you know, curiosity and strategy and so they were able just to jump into this and I think I've seen that really transform not only the way that they

acclimate to the organization and contribute but then also how they're raising the bar for all of their peers as well. And so it's been just an

as well. And so it's been just an awesome piece to see the peer learning that Colin talked about in so many different ways because it really does feel collaborative. It feels like we're

feel collaborative. It feels like we're we're moving together and then we're we're unlocking creativity together. So

some of cases may may start with this more adaptive way of working where we are thinking individually but it's unlocking the next step of creativity to do more and build more and it's just it's awesome to see honestly

it is yeah well you're talking about the slope not the snapshot and I think it's this very nice image to think about how these systems from hiring to onboarding

to being you know a tenur employee um what it's like to actually like put somebody on the trajectory so then they have a really good 90 days, you know, and they come in obviously with their

own level of AI fluency, but then they come into an environment they were expecting. It's like they fit right in

expecting. It's like they fit right in and all of a sudden they're raising the game within their first 90 days. I can

think of a few people on the marketing team that just came in and it was like they had been here for years kind of thing, you know, and they were like making really big system improvements.

Totally. Well, let's hit a let's hit a couple of these. I don't know if you want to do the the finance um example. I

think this is like an example of a transformative win, a big a big win that we had. And then we'll run through a

we had. And then we'll run through a couple more and we'll make sure we answer everyone's questions as we go, too.

I know examples are always the best, right? I'm always like, what but what

right? I'm always like, what but what what are you doing? So, here here are some these are examples of some of the a lot of these are pulled from the case studies that I mentioned where we're asking people to document. Um there's a bunch of these and and Tracy can speak

to some next but here's one an example where the accounting team was doing these BVAs business versus actual reports. Um we already had it

reports. Um we already had it semi-automated because we're we are Zap year and we had that but the AI allowed us to at a team level. This was not something just an individual was doing,

but they had this a shared team workflow that they basically created a um just you know you can imagine how AI would help them kind of organize all this

disparate data that's coming in and do some analysis and this was a report they were ultimately sending to leaders or to executives to take a look at um for their departments and this was an

example of the the solution that the person was able to come up with um not only increased the speed, so they were able to do it quite a bit faster. Took

about half the time um but it actually improved the quality which we knew because we caught some errors. So

somebody had just you know in this example I think somebody had just made an error in a spreadsheet two years ago and uh for with a formula and they we just kept using that same spreadsheet

over and over and over again and nobody caught the error and that the AI solution found it like right away the first time. um and like sol you know

first time. um and like sol you know answer some questions where like why why do these numbers never match up? Well,

there you go. Um so that's one really really quick example that like changed uh just like a really common team workflow happens every month.

Yeah. Uh Crit is asking for one uh here.

I'm just going to throw some other ones on the screen, but I'll I'll hit the product marketing team one. This is one that was built um collaboratively by uh Joe and Charles on our product marketing

team. Um, but we we did have a

team. Um, but we we did have a competitive intelligence tool and you know I think the usage was helpful but it was pretty light overall and this is

a good example of they were talking about it and they're like I think we can kind of build enough of what we need here um with with Zapier in this case

but they basically built a a little intelligence pod that pipes insights into Slack. it checks certain things and

into Slack. it checks certain things and we were able to just kind of remove that tool from the tool stack and it's a really nice thing for them to be able to say hey we were not just accepting the

status quo of this tool here we went and built something and we're essentially saving you know 25k in annual value kind of replacing that I think that's a good example where there's the permission

there's the learning and there's the let's let's try this out let's see how it goes it's going pretty well let's take this tool out of the tool stack and actually like simplify things for for

our stack overall. So hopefully that's a kind of a good example as well.

I like the business partner calibration example on on here. We just did our performance reviews which we call impact reviews. As part of that, managers get

reviews. As part of that, managers get in a room and they they look at uh the ratings and what people have um documented about their folks. And um

that's really hard. It's a complicated process. Our business partners this time

process. Our business partners this time used an AI solution to help um pre-populate that calibration activity with uh basically like commentary on things to pay attention to. It's

basically comparing people's ratings with what our actual criteria was and like flag stuff. And

one of the tangible things that allowed us to do is a we got a lot of positive feedback that those sessions were much more helpful than they have been in the past. And then second, that was one of

past. And then second, that was one of the factors that allowed us to reduce that whole cycle from like four or five weeks to two weeks.

Um because calibration was one of the things that took us the longest and this just like fast forwarded that.

Yeah. So good. I loved that this time around. It was like, you know, there's a

around. It was like, you know, there's a lot of groaning at that time of year because people are trying to do their jobs and then there's a lot of important work for managers to do and just to

condense it and make it a a sprint and it was super helpful this year. Tracy,

any others come to mind? We also have some um HR related examples. I was just going to just give one because someone also asked in in the questions in the Q&A section about what was maybe a use

case in hiring and so this one in the middle we also just like uh marketing placed an inter uh replaced an external tool that we use so very easy to justify

to our CFO saving $30,000 but um we take reference checks really seriously at Zapier and if anybody has been on a recruiting team reference checks are usually like a pretty generic tool that

feels like a check the box activity not helpful. Um people who fill them out

helpful. Um people who fill them out don't really like doing it. Um and so we really wanted to change that and make it a lot more intelligent, so much more personalized and customized to the

candidates at scale. And so we basically created a whole workflow where we're using automations and agents to before we ask a reference for any information

about a candidate, we are pulling all of the information from scorecards, all of the information from transcripts um from uh when we when we do the video interviews, we have transcripts so we have a sense of everything that the

candidate talked about. We compare that to our internal rubrics, the job description, the job criteria, and then we're able to ask really focused and specific candidates per role and per

candidate about the information and the gaps that we need to know in order to make an offer. Um, and so it's just become really personalized, really tailored. The people we ask to fill out

tailored. The people we ask to fill out the reference form or to talk to us live it because they're like, "Wow, you really know this person and like I'm happy to dig in." And we get so much better information. So, it's a great

better information. So, it's a great example of not only did it save us a lot of time, but the quality of the insights that we're then able to get and tee up from the references themselves are so

much better because the AI is helping us pull up the insights from our entire process and we don't have to rely on on something generic. And then because we

something generic. And then because we built it internally um on my talent team, we were able to eliminate this tool and actually save dollars too um with a product that is so much better.

So, I like that example because it's it's very applicable to how we run. Oh,

it's so awesome. Um, I'm gonna fire a link off to folks and then, um, I'll I'll shoot a couple more. Somebody was

asking if they could get some help. Um,

so first one, a couple use cases here if you just want to see some of the ones from our HR team. A lot of these are are still really in in action today. Um, and

then the other thing I wanted to highlight is if you're looking for some help, we've got some awesome strategists, some awesome experts um, can come and do a consultation call. Um,

so who was asking about that in the chat? Um, gosh, uh, I'm I'm missing

chat? Um, gosh, uh, I'm I'm missing them. Aka, it's it's there for you. Um,

them. Aka, it's it's there for you. Um,

and then the last thing I wanted to highlight too, um, is, uh, next month we have Zap Connect. Um, and so we'll have more sessions like this. We'll also have sessions where there will be building

being done. So I I couldn't recommend a

being done. So I I couldn't recommend a better like four to five hours. Um,

block it off, come to Zap Connect. There

is a lot um to learn from leaders like Tracy and Colin um and share and and a lot from builders who are building some of these solutions as well. Um gosh, we

are almost out of time. So Tracy, Colin, I did a terrible job of time management to leave a few questions at the end, but we were hitting them as we go. Um

anything from the Q&A here, if you switch to that tab, that you want to take on to close this out?

Um I we chatted a little bit about use cases. I can say um very quickly in

cases. I can say um very quickly in terms of tracking progress against new hires when we did something like roll out AI fluency it wasn't just a hypothesis. We wanted to ensure that

hypothesis. We wanted to ensure that like actually raising the bar for AI fluency was meaningful in terms of employee productivity, engagement,

ability to build. And so um we have a way that we track through onboarding surveys, hiring manager onboarding surveys, qualitative studies as well as a pretty extensive like quality of hire

algorithm we use. So all to say is there's a number of ways that we track this, but it felt really important to be data driven in if we actually believed and if we could prove that raising the

bar for AI fluency actually then led to a greater level of productivity, engagement, etc. with our workforce and not just like a hey, we think this is a good idea because it's really popular to

talk about AI fluency right now.

Nice. Colin, maybe I'll ask you uh with a minute left for our audience if they could kind of take away one thing from today's presentation or just your experiences. What's kind of the one

experiences. What's kind of the one thing they could take back to the team, their teams today?

Yeah, good question. Um,

you know, I just go back to uh thinking of that soccer field. Um, you need the cheering section, but you also need the support section. Um, think about what

support section. Um, think about what that support looks like tangibly but in a way that is that is like crowdsourced from your wonderful employees and then

just look for those bright spots that you can help pull together what's working well that you can influence the rest of the organization with. I think

that's been our model that that's worked really well and helps you kind of like ride the waves as this thing keeps changing um and helps you stay really flexible.

I love it. Well, that's a perfect place to end. We're at the top of the hour.

to end. We're at the top of the hour.

Just wanted to thank uh both you, Tracy and Colin, for joining us and everyone in the audience. Thanks so much for joining us and we'll do more of these in the next couple months. And come to Zap Connect. There'll be a lot more there,

Connect. There'll be a lot more there, too. So, thanks everybody. Appreciate

too. So, thanks everybody. Appreciate

you spending an hour with us.

Thanks everyone. Thanks Ryan.

Yeah. See you everybody.