Talent Trends: A Data-Driven Discussion on the State of Hiring

Key Takeaways

1. Scott Bonneau connected candidate frustration and employer frustration as part of the same feedback loop. When candidates apply and never hear back, they apply to more and more jobs over time, which lowers signal quality for employers and increases noise in the funnel. Faster candidate communication is no longer just a candidate experience initiative. It directly impacts application quality upstream.
2. Traditional top-of-funnel signals are rapidly losing value in an AI-enabled hiring market. Scott emphasized that resumes and LinkedIn profiles are becoming easier to optimize at scale, pushing recruiters to look for stronger proof-of-work signals like GitHub activity, published research, portfolios, or public demonstrations of craft.
3. Felicia Menon shared that referrals continue to outperform because relationships still create stronger signal than automation alone. Her team treats referrals as a high-rigor process, not a shortcut, distinguishing between true referrals from trusted collaborators and simple leads.
4. Scott pushed back on the assumption that inbound candidates are inherently lower quality. Shopify found no statistical correlation between source of hire and on-the-job performance. The challenge is not whether strong inbound candidates exist. It’s helping recruiters identify them efficiently inside very large applicant pools.
5. Both Scott and Felicia described recruiter productivity as relatively stable despite dramatically higher application volume and increased AI adoption. The bottleneck is no longer administrative throughput alone. Hiring remains fundamentally human, with recruiters spending more time on evaluation, alignment, and relationship-building. 
6. Both leaders described workforce planning becoming significantly more dynamic. Rather than relying on static annual plans, recruiting teams are increasingly operating with rolling forecasts, live operational data, and flexible capacity models that allow them to respond more quickly to changing business demand.
7. Scott and Felicia both described AI as changing what hiring teams evaluate during interviews. The focus is shifting away from whether candidates can complete isolated tasks on their own and toward whether they can solve real business problems while effectively leveraging AI tools, judgment, and domain expertise.

Transcript

Note: transcript is auto-generated and may contain minor inaccuracies.

Kelsey Peterson: Very excited to have you both here today. Uh, Felicia Menon is the head of talent acquisition at Juniper Square, where she has spent the last four years scaling the team and evolving the talent function during a period of rapid growth. She began her journey as a senior recruiter and has since grown into her leadership role today. Felicia holds a degree from the University of Michigan and is based in Milwaukee. She is an avid traveler and reader who brings curiosity and energy into both her personal and professional life. Welcome, Felicia. 

Felicia Menon: Thanks, Kelsey.

Kelsey Peterson: We also have Scott Bonneau here today. He is head of recruiting at Shopify. After nearly 20 years, two decades, as a software engineer and, and as well as VP of engineering and CTO across companies big and small, Scott very wisely chose to pivot into TA HR in 2018 while working at Indeed, where he was tasked with rethinking recruiting from a place of first principles. Outside of work, Scott is a professional musician. He owns and operates a recording studio outside of Austin, Texas. Welcome, Scott. 

Scott Bonneau: Thank you.

Kelsey Peterson: Okay, what can you expect from us today? We're going to talk about three disparate topics, though they're all interrelated, and we'll do a very quick wrap-up. It's going to be a very punchy conversation. So without further ado, we're gonna start with the topic of inbound volume and referral stagnation. So what we are seeing is that applications per hire tripled between 2021 to 2024, and it remains elevated still. That's probably not news to most folks in this room. There is a ton of activity. Needless to say, our recruiters are busy. And inbound applicants-- applications make up fifty-two percent of all hires across Ashby customers Yet inbound candidates are least likely to secure an interview. So high application volume, least likely to secure an interview, but most likely to be hired. So there's something interesting there we wanna spend some time talking about. And what else are we seeing? Internal and referred candidates are the most likely to move from interview stage to offer. However, the relative share of applications from referrals is decreasing over time. Okay, I just went very quickly here. Basically, we are seeing more than ever candidates entering into the funnel. We know the teams are busy, but it is requiring more work per hire. Volume is up, signal is harder to find in a high-noise environment, and that pressure is showing up directly in recruiter capacity. So to the team, let's talk a little bit about this. Scott, uh, when we were catching up earlier, you shared something interesting. You said that you believe job seeker frustration and company frustration are causally linked. Can you spend a little bit of time talking about that? 

Scott Bonneau: Yeah, for sure. Uh, one of the things that we looked at a lot when I was at Indeed, which is obviously focused on helping folks get jobs, was when we looked at our, uh, job seeker audience, the number one complaint that they had was what we call the black hole problem, which is you apply to a job, you never hear back. Um, and so if you apply to one job and you never hear back, you might apply to two jobs, three jobs, four jobs. You never hear back, you apply to a lot of jobs. Um, on the flip side, what employers talked about was a concern about application quality. So what they were seeing is a flood of applications, but not a lot of quality in that application pool, so kind of very low signal-to-noise ratio at the top of the funnel. And it's clear to see how these two problems feed into one another. If you're a job seeker and you reach out to your ideal employer that you think you're a good match for, you apply, but you never hear back, you know, you're gonna start getting into a loop of applying to more and more places. Um, and statistically speaking, uh, what we found at Indeed was the, uh, first job that a job seeker applied for is generally the one that they are most well-suited for. So by the time you're applying to five, 10, 20, 100 jobs, your sort of quality of apply tends to drop off as well. Um, and so one of the key things that I think comes out of this is it's imperative for employers who are looking for strong applicants in that application pool to, uh, to be able to address candidates as many, uh, as, as directly and as quickly as you can, um, and make sure that you're getting back to candidates so that you aren't sort of perpetuating that, uh, that black hole problem that candidates are experiencing.

Kelsey Peterson: That makes complete sense. And timely recommend- timely replies, certainly. What else would you recommend to sort of prevent the consternation of additional applications over time? 

Scott Bonneau: Yeah, I mean, I think the biggest challenge that a lot of us are facing right now, and you talked about here with just the volume of applies that everyone is seeing right now, is the traditional signals that we're used to using when looking at candidates at the top of the funnel, their resume, their work history, and the like, are sort of rapidly trending to zero. Uh, and particularly in a world now where, you know, we've seen, uh, the ways that, that as, uh, TA professionals we're able to leverage AI, well, candidates have those same tools as well. So on some level you can... it's very, you know, low cost or free for candidates to produce, you know, the ideal resume for any particular, uh, you know, inter- or for any particular job that they might be applying to. And so one of the things that I think is gonna become increasingly important and something that we're certainly digging into on Shopify, is looking for alternate signals, not just what you see on the resume or a LinkedIn profile, but looking for other sort of, you know, sort of sources of proof of competency or proof of work, you know, out on, on the internet. Whether that's, you know, if you're engineers looking at public GitHub profiles or if you're a researcher, your presence on Google Scholar and the like there, trying to find other things that might be an indicator that candidates are particularly well-suited for the role, which allows recruiters to spend, spend their time on the highest probability candidates.

Kelsey Peterson: Thanks for sharing that. I'm gonna come to you, Felicia. Uh, when you talk about inbound versus referrals, you're actually seeing something very different than the broader market, and I would love for you to share what you're seeing and how you think about signal today. 

Felicia Menon: Yeah, absolutely. So we are actually seeing a trending upward of our referral percentage. So over twenty percent of candidates typically are at Juniper Square are hired via referrals. And I think we've talked a lot about it today. We are seeing huge advancements in AI and technology, and it makes it easier for candidates to apply. It also makes it easier for recruiters to go through applications quickly. But what tooling and automation doesn't do is build relationships for you, and the signals that we get from that, those relationships that people build, um, through referrals is really positive. So we wanna foster, as a talent team, as much of that as we can. So we have done a couple things to invest in that, uh, at Juniper Square. It is a very, um, you know, kind of referral-forward culture. We talk about it at all-- every all-hands as a company, uh, and we go through what is the distinction between a referral, someone you've worked with before and can attest to their work, versus a lead, so someone that, um, you know, you could say you like them, you think they could be a good fit. Um, so we go through those, those distinctions. Our talent team also is joining every onboarding sync, and we actually have every new employee pull up their LinkedIn with the recruiting team. Uh, so the recruiting team serves as a partner in driving referrals versus the receiver of, of those referrals. And then we can, you know, really help guide in the quality and what that looks like as well. So, uh, we've done a couple things to invest there, but I do think that there can be a misconception sometimes that, and I'm sure, you know, many people can relate to this, but that referrals are easier or quicker to move through the process, don't take as much as involvement from a talent team. And I think that's actually, you know, the opposite of, of the way that we think about it. And one example that comes to mind for this is- There was a leader that I've worked with for years, um, who viewed, you know, the, the referral pool and the, the recruiting team as, you know, kind of the person to process the offer once we, we get introduced. Um, and over years and, and building trust, we really changed that perspective, um, to be actually, "Hey, referral, our recruiting team has an extremely high bar. Um, you need to connect with them and go through their process and their rigor to have a spot at Juniper Square." And I think that's actually something that then has led to, you know, us being able to say to our clients, we have an extremely talent-dense team. We've built those relationships through our network, um, and hired those people at Juniper Square. So it has a compounding impact as, as you continue. 

Kelsey Peterson: Yeah. Thanks for sharing that. It seems like early, often, consistent, maintaining that bar- Yeah ... um, not easy to do, but worthwhile. 

Felicia Menon: Very worthwhile, but yes, not easy to do.

Kelsey Peterson: Um, Scott, I'm gonna come back to you. There's a narrative that, uh, some folks say that inbound is lower quality. Um, you've pushed back on that pretty publicly. Share with me, like, how do you think about quality across source? I think folks would be curious. 

Scott Bonneau: Yeah, there's a couple things here. I mean, one, I think the...it is definitely true that the success rate for direct applies is, is gonna be lower than your sourced or referred candidates just by nature of sort of large numbers, right? Um, uh, but one of the things that, that we've seen at Shopify, one is, uh, direct applies, uh, make up a plurality of our hires. So, um, it's the most significant source of hire for us right now, even though the success rate broadly is quite low. Um, but one of the things that we've actually dug into is taking a look at, uh, on-the-job success after folks have hired and trying to do a, an analysis or correlation study to see whether or not, for example, folks that were referred or sourced are- tend to perform better on the job on average than folks that come from direct applies. And what we have found is no statistical correlation there. So what we are seeing is there absolutely are very good candidates that are coming in as direct applies. I think the challenge becomes how do we make sure, um, as TA leaders, we are having our recruiters or putting our recruiters in position where they can spend their time on the most likely, uh, high-probability candidates. And in a world where the application pools are so large, that becomes a real significant challenge. 

Kelsey Peterson: Uh, it's interesting hearing you talk about focusing recruiter time. Maybe a rapid-fire question. How are you thinking about where to invest time as it stands today with the recruiting team? 

Felicia Menon: Absolutely. I think to your point, uh, at the top of funnel, there is so much that can, can exist, um, of just size and candidate pool to go through. And a lot of what our tooling and Ashby and even some of the stuff that we saw today is allowing recruiters to spend less of their time there. So where can they use that time? Um, I think it's spending time with their stakeholders and hiring managers and really being embedded. I think those conversations that we have with hiring teams then shift from, "Hey, here's an update on my pipeline and, and, and funnel metrics," to, "Hey, this is what I'm hearing in the business. This is what this could mean in the next few months for us. This is what we need to do." And I think that type of forward conversation is really exciting. 

Scott Bonneau: Yeah. I mean, similarly, I think one of the things that's, that's gonna be very important is understanding where traditionally have we found from a source of hire standpoint, whether it's, you know, a competitive set, a sector, a geo, those types of things where folks have in particular been successful to allow recruiters to be able to narrow down their focus. But again, also I think finding ways both to gather sort of passive signal based on, you know, information that's on the public internet, demonstrating proof of ability. Uh, but also looking for, for creative ways to allow candidates to demonstrate their ability at the top of the funnel, uh, that don't necessarily require a half an hour or an hour of a recruiter's time to be able to do that assessment. Those are things that are gonna be really, uh, I think areas of lots of investment in the near future. 

Kelsey Peterson: Makes perfect sense. Um, folks in the back, feel free to filter your way in. We're gonna pivot a little bit and talk a bit more about, uh, recruiter capacity. This is a great time to get yourselves comfortable. Um, okay. So what are we seeing in recruiter capacity? We just talked a little bit about the sort of relentless need to focus in certain areas. Um, hires per recruiter for business roles have reached a five-year high, and it's notable here. You can see the tech hires, I believe are pink. Um, they have largely stabilized. This likely won't come as a surprise to anybody in this room, but, uh, time to first fill for these tech roles has remained consistently longer than the business roles. And so we have teams interviewing significantly more candidates per hire, and yet fewer supporting roles like coordinators. Therefore, there's a forced reliance on, um, more tooling. So we've been talking about signal breakdown at the top of the funnel, and what we're seeing in the data is that even as the volume is increasing and tooling continues to improve, recruiter output is staying relatively stable. If the output is relatively stable, begs a pretty earnest question of like, what's changing and what's not? And so we're gonna unpack that a little bit and dig into recruiter capacity. Scott, I'll start with you here. Uh, we're seeing this, that recruiter output reasonably stable, um, even with a big increase in volume and in AI tooling. What's your best guess or how do you see that showing up at Shopify? 

Scott Bonneau: I think there's a couple things. I mean, I think one is, uh, even with the advances that we're seeing in technology, the ways that we're able to apply AI to what it is that we're doing right now. At the end of the day, hiring, at least for the time being, I think remains a fundamentally very human process. Uh, you've got individuals who are making life decisions. Those things take time. You've got companies that are trying to make sure that you are bringing the right people in, particularly in a world right now where hiring volumes might be a little bit lower than they were, say maybe five years ago. But at the same time, the market for candidates has a significantly larger, uh, number of folks than might have been on the market in, say, you know, a 2019 to, say, 2021 timeframe. Um, but there are some, I think, what I would call like a speed of light problem here, which is how fast can you move any individual candidate through the process? I don't think that that's materially changed over the course of the last little bit. So as we look at, um, capacity and productivity metrics, and the real questions again, I think stem toward closer to the top of the funnel. How do we make sure that recruiters are spending their time on those folks that are likely to be both good fits for the role and likely to accept an offer at the end of the process? So that's where we are, we are spending a lot of our, you know, time and energy these days. 

Kelsey Peterson: Um, now Felicia, your team hired, I think it was over 500 people with a very lean team last year. Uh, what is your perspective? How do you think about recruiter capacity? 

Felicia Menon: Well, one, I think about recruiter capacity all the time. So it is something I think as talent leaders, we're spending a lot of our time thinking about. Um, and I think one, you know, having a great team, that helps and, and they're here today, so a huge shout-out to the Juniper Square recruiting team. But, um, leading into the year, I think with 300, um, planned roles to fill. So that's obviously a pretty significant difference. And a, a few things that helped us do that, I think one, um, we have a team that is very embedded in our business, but is also cross-trained, so we are able to, you know, uh, adjust that capacity across recruiters. Um, but we actually look at our numbers, um, and our targets per month. So we target, um, four hires per recruiter per month, um, for business, five in tech. Um, and that has helped us, um, not necessarily set, you know, the metric of what we're aiming for, um, but more to, um, set the target and, and communicate to the business this is what we can commit to in, in a year. So we were able to, you know, you know, work across our team to align on some of those numbers. But as I highlighted, there's still a gap there. There's still a gap in, in, in the, the hires that we ended up making. And so we have something that I refer to as the faucet model, and I will say it feels a lot more like turning the faucet on than turning the faucet off. But, um, what are our levers as a recruiting team to be able to respond to those spikes in business need? Because they will come and, and for good reason, we want the, our, our business to be growing. Um, so some of those things and what that's looked like for us has been tooling. Ashley has been a big partner for us in that. Um, but also in, you know, elastic layers across RPOs. We've worked closely with Talentful, um, agencies and across geos, there's differentiators there, and that's really helped us respond to that, that need. Um, and I think it's important to call out it's not just the numbers. The capacity isn't just we filled 500 people, it's also holding that talent bar. So we hired 500 good people, and that takes a lot of time to do. So, um, yeah, that's how we've thought about it. 

Kelsey Peterson: The faucet model, I bet it resonates with a lot of folks. It's, the faucet is on. 

Felicia Menon: It is just so on. Yeah. 

Kelsey Peterson: Um, I'm curious, how do you decide you've got this elasticity you're talking about? When are you investing in full-time head count? When are you thinking about these other levers? Like- Yeah ... what does that look like for you all? 

Felicia Menon: Yeah. I think it depends o- obviously on, on the need. Are we talking about, you know, kind of a long-term prediction or quarterly, you know, we see a spike. Um, so it depends on, on the timeframe that we're talking about. Um, but I think it also depends, and what I try to communicate to leaders as I'm going through these activities, 'cause it is a, a, a group and a business decision, is it does come with a cost trade-off. So I think a lot of, um, you know, executive leaders can understand that language of like, "Hey, we can respond to this need, but it does come with this cost, and the recruiting and our, the people team, um, needs to be inc- um, uh, captured in that." And I think we've done a lot in the past year where we have that seat at the table. Um, but it also comes with a hiring manager continuity trade-off, and I don't know if we, we always talk about that enough. But as much as I love elastic layers and they've been deeply helpful in helping us succeed, um, it does come with hiring manager, real hiring manager fatigue. So, um, you know, it, it takes time to build trust with a recruiter. It takes time to define, you know, your best, um, profiles and what you're targeting for a role. So when, when we are responding to those short-term hikes, it's my job to make sure that we're still maintaining that con- continuity and ultimately, you know, holding that talent bar.

Kelsey Peterson: Yeah. Um, Scott, I'm curious, you had mentioned earlier, uh, traditional capacity modeling, very different now, uh, from your time at Indeed. Uh, would be curious what you've seen change over time here. 

Scott Bonneau: Yeah. Shopify takes, uh, a, a very interesting approach to how we think about headcount, which I, which I think is, it's certainly been unique in my experience. There's a couple things here. One is we, we do a, uh, a six-month look ahead at, at what the sort of target shape of the organization, uh, that we wanna achieve is, and we do that every quarter. So basically every 90 days we are running a forward-looking model that expands over the course of six months, and we do that in what we call this sort of a, a, a principled approach. So the idea is we've defined a number of things about what we think the ideal characteristics of the shape of the organization are that we're gonna wanna have. So that includes things like the, uh, wingspan for managers, the ratio of more senior talent to more junior talent, uh, the ratio of, for example, product managers to engineers, or designers to product managers. And when we take a look then at the macro level of what we are trying to do with the shape of the organization globally, we are looking at it through the lens of these principles that describe what that target shape should be. And then we work back from that on a quarterly basis to recruiting to look at where and how we should allocate the capacity that we have within TA. And that also provides us some signals about are we right sized in TA. So similar to, uh, Felicia at Juniper Square, you know, we have a hybrid model of like largely doing our work with full-time recruiters, but we also have that ability to kind of flex up and down, uh, using, uh, contractors and, and the like to support the hiring that we're doing there. But what I think what we have found is, uh, the notion of trying to plan annually at this point, it seems like a bit of an archaic, uh, uh, exercise, at least given the, the, the domain that, that we exist in. And so rather, uh, than trying to conform to that, we've just kind of acknowledged the fact that we know that we're gonna be on a little bit of a rollercoaster ride, but let's try to approach that in a principled way, uh, to where we know that we are never in any given quarter going to like achieve that target shape, and we also know that this target shape is gonna change as business demand change and as the landscape changes. Uh, but we know that we are gonna be constantly making adjustments to get ourselves closer to what that shape looks to be. 

Kelsey Peterson: I think for this audience, that probably resonates deeply. I'd be interested outside of this space, like, what does it look like getting buy-in across the organization? I could see an organization maybe saying, "No, we need to have clarity at the beginning of the year." How are you, uh, either of you getting sort of, like, consent or support to, to be agile in this way? Because it isn't static. It isn't static today. 

Scott Bonneau: Yeah, you know, it's one of the things I've seen about, uh, about Shopify that I think is really amazing, and one of the things that, that drew me there in the first place, is even 20-plus years in and at the scale that we operate, it is more like a startup than the actual startup that I had been at, you know, for the two years prior. Um, so it moves really quickly. The leadership team, everybody is deeply invested in the direction that we are taking and has sort of bought into the systems there. So in terms of, uh, getting buy-in with the leadership team, the executive team and, and the like there, that really hasn't been the issue. The, the challenge really is, um, being able to keep up with processes. We all know, you know, in, in the recruiting world, um, if it takes 30 days or 45 days or 50 days to fill a role, uh, if you're constantly changing your priorities every four, five, six weeks, we run into an actual problem there. And so that's part of the challenge there is just trying to get the practical reality of the minimum amount of time it takes to execute searches to work with the speed at which the business wants to change and, uh, and find that, that correct balance, which certainly, uh, keeps us on our toes. But, uh, I've been very happy with the way that our team has executed on that in the time that I've been here. 

Kelsey Peterson: What about that recalibration or that pivoting at Juniper Square? 

Felicia Menon: Yes. There's a lot of recalibrating, and it feels like daily. Um, and I will say, I remember sitting in this Ashview crowd a year ago, um, and thinking, "Wow, the people on stage, like, they have it really figured out because they, they, they sound, um, like they have it all together." And I, and I think from a recalibration perspective, what I've learned at least is it is just constantly daily making as best of an educated decision as you can with the data that's in front of you. That data might change in a few weeks, days even, um, but that's what, that's what you need to do. And for us, uh, Ashby reporting has been an extremely valuable partner in doing that. Um, in part- partnership with Kendra, our people analytics manager, we've pulled together something across, um, Ashby data, uh, Team Ohana, which is our headcount planning tool, uh, and Claude and Cowork. And to be able to get a live time view, not only into retroactively what we've seen, but looking forward, like you said, Scott, you know, that kind of six-month view. What is the demand from the business, and what are the priorities, and then what's the capacity of the recruiting team, and then what are my levers? Like, what could I... If, if we invest in X, you know, what changes? And so I think having the data to make those decisions is extremely critical because when things change that fast, again, you're just making the best decision with what's in front of you today.

Kelsey Peterson: Speaking of things changing fast, um, you're welcome to answer either parts or both of this question. Um, if you look back, let's say, two, three years and you look now, where, where are folks, where are recruiters spending more time and/or where are they spending less time? Like, where are you- what are you taking off their plates? I'd be interested, maybe like a, in the spirit of time, quick hit on this question. 

Scott Bonneau: Yeah, I mean, I think certainly one thing we're seeing right now is more time spent at the top of the funnel vetting candidates, um, you know, doing sourcing and the like there, just because the landscape is quite different, you know, sort of post second half of 2022 till maybe the last, uh, six or nine months or so. Um, and I think, uh, thankfully, less time on a number of the administrative tasks, thanks to a lot of the tooling. Many of the features that exist in Ashby have made our lives quite a bit easier on that front. And, and ideally, that's really what I'd love to see that team spending more time on, is sp- those things that are candidate-facing, the things that are pertaining to making sure we are aligning what the expectations of the business are with the candidate pools that we're developing and, and less time just focused on the mechanics of making the process run.

Kelsey Peterson: Yeah, makes sense. And also a good call back to one of your early, earlier notes of, like, the human part of hiring is the focus. I'd be interested, a bit self-interested, but any particular part of, of Ashby or your tooling that has reduced that administrative burden? Like, anything in particular where the team has been like, "Oh."

Scott Bonneau: I mean, I think, like, the scheduling capabilities is really amazing. And managing interviewer pools is a- is another big one. We do quite a bit of interviewing, um, and making sure that we are balancing interviews, interviewers appropriately, making sure that we have the right set of people lined up for an interview as we go to schedule a candidate. Uh, all those things have been a massive help. And I know we've seen, for example, like, the drop in the number of coordinators, for example, per, you know, per, per, uh, in each organization. We have no full-time coordinators at Shopify today, and we are still able to make sure that all of those things happen, while also allowing our recruiters to maintain sort of historical throughput, which has been really good and, and, uh, sign of great tooling on Ashby's part.

Felicia Menon: Yes, a lot of different features in Ashby. I will say to your previous question on where we're spending more time, I think just the short answer is in the business. And I think there's a feedback loop here of, like, the more excellent execution that a, a recruiting team does, the more that they're embedded in their teams. And we've seen positive reward for the talent team because of that. So our, our recruiting team is going to the engineering off-sites, um, to be with the team and sitting with them in those meetings. Our sales recruiter was recognized to go to President's Club. So it's just-- it, it moves it from, "Hey, here's the talent function ex- executing on these roles," and it moves more of the recruiter's time into being actually business partners on what we're accomplishing.

Kelsey Peterson: I'm sure Inquiring Minds, at least I wanted to know, how did you kind of get into being in the business? Like, what did it look like to sort of shift and be that, like, strategic partner? 

Felicia Menon: Yeah. I think, one, just having, like, a passion and a care for what your, your teams are building, uh, goes a long way, and showing up and being really curious on what it is, not just the person in the seat and what they're doing, but the impact of that to the broader organization. I think that has helped us, and then it is just forming those trusted relationships with your leaders, and I think those trusted relationships, um, it might sound like they could come from like, "Oh, yeah, we filled ten hires for you. You must trust me." But I think it's more in the, the situations where they didn't get filled on time, or there's hard conversations, or something didn't go right. Um, you learn and, and show up, um, through those situations, and I think that trust builds and, you know, it's gotten to where we've seen it today, so. 

Kelsey Peterson: Awesome. Scott, anything to add to that around the sort of business dynamic and the partnership internally? 

Scott Bonneau: Yeah. One of the things that, that we do at Shopify, which is unique in my experience, is we have centralized a lot of the hi- a lot, a lot of our hiring. Um, and about 80% of our hiring is done very intentionally through pipelines. So only about 20% of our hiring is sort of rec-based or individual specific searches. And, uh, one of the things that that centralization has allowed us to do is, uh, my recruiting leaders that are, are, are working, for example, say, on, on commercial hiring or engineering hiring, are very regularly squaring off directly against what we call the discipline leaders of, you know, VP level person who is responsible for hiring within that discipline, and meeting at least bi-weekly, going over progress on pipeline, where we are seeing, you know, green flags, red flags, anything that's showing up. For example, if we see a drop in apply rate or a, an increase in offer decline rate, something like that, we're spending time directly with discipline leaders on that, uh, on, on those issues. So there's quite a bit of visibility, and there's a very strong expectation that TA is gonna show up with the data and a clear opinion and perspective on what it is that's happening. But the other thing that that's allowed us to do is because we've centralized that hiring and a lot of that decision-making in those discipline leaders, recruiters aren't spending a whole bunch of time with, you know, 30 different hiring managers in the business. That allows us to really focus on what is the big picture that that discipline is trying to accomplish? How is that gonna show up in terms of the roles that we're trying to hire? It allows us to get aligned on those things and sort of eliminate some of the noise or friction that might otherwise, uh, exist in sort of more bespoke processes. 

Kelsey Peterson: That's really interesting about the eighty percent to twenty. Thanks for sharing that. All right, our final topic for today. Uh, we've talked about the importance of the human element of hiring. We've talked a little bit about AI. We'll talk a bit more. It's embedded across hiring workflows, uh, note-taking, screening, coordination. Um, teams are beginning to use AI not just for efficiency, but even around decision quality. Both these two- teams here today are using AI Notetakers, and so we wanna talk, that's what we'll focus. There are innumerable things we could talk about on this topic, but we'll talk about, uh, Notetaker very specifically. Uh, quick data set up from us. Um, scorecards with AI Notetaker, we know that they include quite a bit more detail, and we know that detail can help inform hiring decisions, so a primary benefit here. Um, and also, AI Notetaker very much speeds up scorecard submission. There's nothing quite like waiting for that feedback form to come in. And when it comes in a little bit faster, it makes a material difference. So, uh, while note-taking and transcription are just a couple examples, we know that AI is no longer experimental. It's a part of the infrastructure that teams are relying on to navigate the volume that we're talking about, um, and really recover that signal. So where interviews used to be more of a black box, and now we have insight into behavior and experience more so than we've ever had before. Uh, both Shopify and Juniper Square, as I mentioned, they're, they're leveraging AI, and they're using these insights from tools to help better their teams. So wanna take the next five or so minutes to wrap us here today and talk a little bit about this. Um, first question, I'm gonna go to Felicia. Uh, you had mentioned when we were chatting earlier that you were using AI-generated interviewing data to advise and fact-check on hiring decisions. Um, we're seeing that happening more and more. Uh, what does that actually, like, tactically look like in practice at Juniper Square? 

Felicia Menon: Absolutely. I think any time where you can get more data into something, and specifically something as important as a hiring decision, uh, many benefit from it. Um, but I think one thing it really allows the recruiter to do is to show up much more as a coach and advisor, um, to a positive or successful hire. And so we're utilizing that data, and we actually, um, did a training session for recruiters earlier this year. Um, the role of recruiter as a coach is not new. That has always been something that we've done with our hiring leaders. Um, but now you have so much more data to be able to do it. How do you actually tactically utilize that data? So we went through a couple examples, and, um, there was one even last week where a recruiter of mine was working with a leader. Um, there was a debrief and a lot of mixed feedback, and I know that's a very common situation. Um, but now with, with the, the data that we have from an AI notetaker, we can say, um, we can go back to that with the hiring manager, so I do think it should be a, a, a partnership there. Um, and that hiring manager is actually able to connect with their interviewers and realize that there is actually some misunderstandings and some gaps on, you know, what we were assessing. There was coaching for the interview team, um, but there was also data that we could point to of like, "Hey, actually, this candidate does demonstrate this skill set," and it led to, to an offer. Um, and I think the flip side has been true of times where it doesn't lead-- we don't move forward with a candidate, and I think that's equally positive. Um, so, so we've, we've learned a lot there. So I do think it also comes with some trade-offs, though, of just making sure our, our hiring teams are, um, the ones who are inputting what their feedback is on the candidate. So while transcripts are great, there is still a requirement to assess what is a good answer versus a great answer, and we need those teams to do it. So it needs to, to kinda work in partnership there. 

Kelsey Peterson: I'd be curious, uh, w- how would you say your confidence has changed with the introduction of AI in, in this part of your process?

Felicia Menon: I think, uh, like I said, like, any time you can backtrack something, my confidence goes up , so, so I'll, I'll take it. But, um, I, I think we can, um, still there's still room to coach, like, hiring leaders on how to use it in a way that we're not just reinforcing bias, uh, through, through, you know, kind of, you know, if you've liked one thing, AI is great at recognizing patterns, so making sure that, you know, we're assessing for a diverse candidate pool. So it's definitely made me more confident, and then I think we have a lot of room as leaders to continue to coach on how we use it. 

Kelsey Peterson: Yeah, that makes sense. Uh, Scott, earlier you were talking in the very first topic around, uh, getting additional information, like finding a developer's GitHub or looking at some documentation. Uh, you had said earlier, I think it was AI makes everybody a seven out of 10 on anything or everything. Um, what does that mean for, for you with hiring? 

Scott Bonneau: Yeah. This is one of the things that Toby Lütke, the founder and CEO of ChatFly talks about, is that AI kind of makes everybody a seven out of 10 at, at, at everything. Um, and so the real question is, what, what, what can we do with that now? And, um, you know, I think one of the implications of that is that, uh, the distance between a good idea and a, an initial, like, V0 implementation of it is very close to zero now, irrespective of what your particular background, uh, would have been, which I think creates a really big opportunity for, uh, for all sorts of new, you know, building and, and, uh, ideation and creation in a world that may have not necessarily been accessible to folks before. You know, uh, for example, my, my background is in software engineering. I have zero, like, design or aesthetic sense. But with AI, I can put together something that not just, like, functions, but also looks a little bit pretty in a V0, which might be enough to get a, say, a real designer to be interested in what it is that I'm working on. But as we think about that in the world of recruiting, I think one of the things that's really interesting that's shifting is the, you know, core problem that we, or core question we've been trying to answer in interviews since the beginning of time is really can this candidate do- this function or this job. Um, that question I think is shifting in a subtle but important way moving forward, which is less about can this candidate do this job, but more can this candidate get this job done potentially by leveraging AI or a fleet of agents to make, to make that happen. Um, and, uh, more so than ever, the, the interesting signal out of interviews, particularly when you're trying to measure someone's ability to execute within their craft, is less about whether or not they can solve some toy problem that you can fit into a 45-minute interview that gives you an approximation for whether or not they can do the job. But with the, you know, with the advances that we've seen in AI even just over the last couple of years, it's now getting to the point where we can actually provide candidates with a real business problem to solve and give them the tools and watch them and see whether or not they're able to accomplish it. How do they leverage AI? How do they, uh, bring their craft excellence to the table as well? How do they know when to press on the AI 'cause it still is gonna get things wrong, or they might not prompt it appropriately. And so it's really opening up an avenue for us to do, I think, much more in-depth, realistic assessments of candidates in, uh, you know, in the interview process, uh, than, you know, was available even, you know, maybe 18 or 24 months ago.

Kelsey Peterson: I'd be curious, how do you counsel candidates around AI as part of the, like, practical exercises that you're sharing with them? 

Scott Bonneau: Yeah, it's funny. I mean, if you dial the clock back even just two years ago, I think one of the major questions everybody in the room was probably dealing with was like, what if people use AI to cheat on the interview process? It's like, well, that's gone. Um, now the real question is how can you design problems that a candidate could only reasonably solve if they appropriately leveraged AI in the time window that, uh, that, that you, you make available to them. But when you look at things like the, the note taker, one of the things I think is really important is the observability of all this. What we actually wanna see is not just the end product. We wanna actually see how the candidate decomposed the problem, how they, uh, prompted, how they, you know, injected their own knowledge and experience into the equation, and how they achieved the outcome at the end. Um, as well as being able to spend time with them, um, you know, maybe even in a debrief asking them questions about, "Okay, I saw that you did this.

Why did you prompt it this way? Why did you question the AI on this result but not on this other result?" And these are the kinds of things like the, the, the interview itself used to be this black box that we didn't really know what was going on. We just had to kind of trust that, oh yeah, we think Scott's a good interviewer or not a good interviewer. We know whether to trust or not trust his results. Now we have the ability to actually inspect all of that, um, not only for the, the candidate, but also the interviewers themselves. It provides, uh, I think a great tool for, for, uh, interviewing itself, but also to as a, as a coaching opportunity for interviewers to get better at the craft of interviewing as well.

Kelsey Peterson: Thanks for sharing that and so much more to come. I would love another hour with you all, but we have a jam-packed agenda. So wrapping up, here is a QR code if you'd like to subscribe to Talent Trends, you can get more information. Enjoy the rest of Ashby One, and please join me in thanking Scott and Felicia for joining here today.