Smaller Teams, Not Faster Recruiters: Andy Mountney on Talent From Atlassian to Elliptic

Show notes

In this episode of the Leaders in Talent podcast, host Adriaan Kolff interviews Andy Mountney, VP of Talent Acquisition & People Operations at Elliptic, the blockchain analytics and crypto compliance company that announced a $120 million Series D this year. Andy has spent 25 years in talent: he founded a recruiting agency, set up the first in-house recruiting functions for European tech companies, led global talent acquisition at Atlassian as it grew from about 4,500 to over 12,000 people and hiring went from 300 to 1,000 people a quarter, and led talent at Chainlink before joining Elliptic this year.

The conversation starts with the decoupling of revenue and headcount and what it means for talent teams. Andy runs a 10-person recruiting team with no coordinators, yet he hasn't met a talent leader who has seen AI push productivity per recruiter through the roof: functions are getting smaller while the individual recruiter does more things, and recruiter capacity is still capped by business and interviewer capacity. He explains why your values have to survive the interview when anyone can produce beautiful employer brand content, why discernment and diligence are what stand out in Anthropic's four Ds of AI fluency, where AI interviewers belong (clarification and speed, not replacing the human conversation for high-value talent), how he thinks about buy versus build in the people tech stack, the four org design shifts he sees everywhere, and why he would rather throttle the inbound and speed up the ramp than build a hiring engine that is hard to switch off.

Timecodes

01:18 Welcome: Andy's path from agency founder to Atlassian and Elliptic

05:37 Scaling Atlassian from 300 to 1,000 hires a quarter

07:16 Decoupling revenue from headcount

10:22 A recruiting team with no coordinators

13:20 Employer branding at AI speed, and why your values must be real

15:40 AI-made work and Anthropic's four Ds

18:10 AI-written CVs, inbound volume and AI interviewers

25:37 Why AI hasn't made recruiters more productive yet

27:45 Buy versus build in the people tech stack

32:23 Four org design shifts in the age of agents

36:13 Throttle the inbound, speed up the ramp

38:55 Andy's first 90 days at Elliptic

___________________________

Connect with us on LinkedIn: https://www.linkedin.com/company/matchr/

Get in touch with us: https://matchr.io/contact

___________________________

Connect with Andy Mountney: https://www.linkedin.com/in/andymountney/

Connect with Adriaan Kolff: https://www.linkedin.com/in/adriaankolff/

___________________________

Resources mentioned in this episode:

Anthropic's AI Fluency course (the 4D framework): https://academy.claude.com/courses/ai-fluency-framework-foundations

Dave Hazlehurst, co-founder of Agentic People: https://www.linkedin.com/in/googledave/

___________________________

RSS feed: https://media.rss.com/leaders-in-talent/feed.xml

Transcription

[00:01:18] Adriaan: All right, ladies and gentlemen, and welcome to another episode of the Leaders in Talent podcast. I'm excited that today I'm joined by Andy Mountney, who currently leads TA at Elliptic. Andy, welcome to the podcast.

[00:01:32] Andy: Great to be here. Appreciate you making the time, and yeah, excited to have a discussion today. So thanks, Adriaan.

[00:01:37] Adriaan: Well, I must say, it's 4:00 in the afternoon for me, but it's almost 9:00 for you in the UK, so I should be the one that's grateful. So thank you.

[00:01:44] Andy: All good, all good.

[00:01:46] Adriaan: Andy, before we dive in, give our listeners a little bit of background: who you are, what exciting companies you've worked at, and also what your role at Elliptic really entails. And give us a little bit more background info on what Elliptic actually does.

[00:02:01] Andy: Yeah, thank you. Appreciate the opportunity to share. So I've been around the talent space for the best part of 25, 26 years now. And my journey, like most people's, starts with falling into the work that we do, in the industry that we're in. And I was really fortunate to get the chance to found a recruiting agency really early on in my career, and that was predominantly, initially, around placing people into headhunting firms and recruiting organizations. But the thread of my career really developed when we set up a business that was part of the end of the 2000s, early 2010s, where businesses were really taking ownership of their recruiting brand, and how they sourced candidates in the market, for the first time. And so we had a business in Europe that really went in and set up the first in-house recruiting functions for a lot of tech teams, a lot of tech startups. We supported the really big FAANG companies, and then we were also there alongside the high-growth companies that were going through those early-stage funding rounds as the European tech scene kicked off. And that then led into another phase of consulting, where I'd go in-house and build the first phase for a seed or a Series A team, put the technology in there, do the first recruiting, and then leave someone behind in place. And that whetted the appetite for really being part of something, as opposed to having my own supplier business. So I was very fortunate to have a client back in 2018 offer me the chance to join them, and it was really a foundational moment for me in my career. I joined Atlassian, went to work for an amazing client, and learned about working with some incredible leaders there. And the thread again from there, over the last eight years or so of being in growing leadership roles in talent acquisition, was really about working for values-driven businesses that deliver incredible platform experiences for their customers and users. And so I was fortunate enough to move to Australia and lead a local recruiting team, and eventually left Atlassian as VP of talent acquisition, with a big global team that was really driving them through the US, India, and APAC. When we returned to the UK, that was my first taste of Web3. So blockchain, Web3, crypto, however you want to frame it, really that world that we now talk about in terms of digital assets. And that's a world today where we're seeing traditional finance meeting that DeFi world. And DeFi, for people, is the decentralized way that blockchains operate, whether it's a smart contract transaction or cryptocurrencies that you'd be familiar with, like Bitcoin. I started that journey in a global business, 500 people, growing them, a remote organization, so very specific challenges about finding the very best people, but with that deep understanding about building an infrastructure platform. And more recently, I'm super proud to have joined Elliptic this year and expanded my role. I've had the opportunity in my career to interim lead some of the teams in the people team outside of talent acquisition. This is my first go-around leading people operations full-time and expanding that role, and it's really part of how I see that whole talent stack, from hire all the way through to onboarding, through to performance and success, and then offboarding folks at the other end. And I joined at a time when we announced our Series D earlier this year, $120 million.

[00:05:06] Adriaan: Yeah, I saw that.

[00:05:07] Andy: Thank you. Yeah, so a really exciting time. And that's really driving big change in an industry where I'm now on the other side of this. We are helping with the regulation, the compliance, the anti-money laundering, helping the good guys beat the bad guys around crypto and blockchain. Our software is the leading compliance software globally, and we've got some incredible opportunities in the agentic space. And that's driving both hiring for skills, as most of us are right now, particularly around AI and agentic, and then an opportunity to do some more work in the US again, which is something I've always enjoyed.

[00:05:37] Adriaan: Just for context, right? So when you joined Atlassian and when you left, how big was Atlassian when you joined, and how big was it when you left?

[00:05:44] Andy: Yeah, I mean, huge change, a different kind of scale. So I guesstimate I joined when we were about 4,500 people, and we peaked at over 12,000 people. And so when I was moving into a global role there for the first time, the big ask from the leadership team was: how do we go from hiring 300 people a quarter, and fighting to get there, to having an efficient way of delivering 1,000 people a quarter seamlessly? And before we drew back on the volumes, we got there, and that was a really proud moment for the team.

[00:06:14] Adriaan: Yeah, crazy, crazy volume. So now going into Elliptic, how big is the company? Around how many employees now?

[00:06:21] Andy: Different size. So we were going through the 200 mark, we're well over 250 people today, and on the way to being larger still and breaking through that 300 mark at the end of the year. So a much smaller business, which I think talks to two different things, right? One is a great opportunity to be at a different phase of a growth journey. And also, and I'm sure we'll get onto this a little bit later, is what all organizations are really going through in tech now, which is that decoupling of revenue and headcount, in the way that hockey stick used to rise. And so if you go back to my Atlassian days, and maybe even early in my time at Chainlink, there was always this real association: the bigger your headcount, the more productive you will be as a company, the more revenue you'll get. And I'm sure we're going to dive into this, but it's such a fascinating time now to think about, in all of our functions, whether we're in recruiting, revenue, engineering: how do we break some of those traditional rules while not losing the human side of who we're trying to be and the values that drive our organizations?

[00:07:16] Adriaan: So let's dive right into that, because that's such a fascinating conversation. I actually just listened to a podcast with a very successful CEO of a company called Bird, and at his peak he had 1,200 employees. He has now scaled back to 120, and he also shared that he was following the traditional growth playbook of SaaS companies, Silicon Valley, however you want to call it: more people, more revenue, a bigger engineering team to uphold the tech stack, until he realized, "Hey, this is not the right format." And he's now back to 120, and he does five times the revenue that he did when he was a 1,200-person company, which is pretty insane. Talk to me a little bit about what you see happening now, front and center: how you, as a business leader, are thinking about that playbook that has changed.

[00:08:04] Andy: Yeah, I think it's fascinating, because that's someone who's placed a bet, and I think there are a lot of leaders who've placed those bets at a business level and frankly don't know whether the choice they've made is going to work out or not, right? We've seen some of the early movers in the space who shredded their customer support teams, or maybe their more junior engineering functions, and then are turning around two or three years later and saying, "We moved before the technology allowed us to, and we lost some of the heart of our business." So for me, in our current situation, it's a really fascinating time of being ahead of that growth and saying, "We can make choices now about how we grow and how we shape our organizations," as opposed to a lot of businesses really having to say, "We're already saddled with this cost base," and I don't just mean headcount, I mean cost base of operating models, cost base of old technology, cost base of amortized commitments to spend on things year over year, before we can go and make those trade-offs and decisions in our systems, in our ways of operating. And so for me, that means you've got this twin track of a market. And if you think about the world I operate in today, the digital assets world, we are a technology supplier, and so we build and operate like a technology company, but we're operating in parallel to financial services companies. I was talking to a peer earlier this week who was saying it's great that a lot of teams now, in talent acquisition or any business team, are saying, "We're going to use Claude, we're going to use AI, we're going to do the things that are best for us. We're going to make trade-offs. Are we going to buy or are we going to build?" And we're in a really great position, growing, to make those decisions now, because we don't have legacy that we're building on. We can go and choose the best HRIS and ATS. I think where it's so hard is if you're a regulated financial services business where you have 50,000 engineers globally in your team, and even if you have this goal of saying, "Maybe we can do this with 35,000 people around the world," how do you even start that journey when you haven't even rolled out AI to your engineers today? And so that, for me, for businesses, is such a fascinating stage, where you have a first-mover opportunity, if you're smaller and you're nimbler and you're not carrying legacy, to move much faster through to being market dominant than you would've done in probably any other era in technology in the last 25 years.

[00:10:22] Adriaan: Let's take that to your current playing field, right? You're head of TA, but also talent ops, right? Especially on talent ops, there's a lot happening. Tell me a little bit more about what you're seeing, what is front and center, what are some of the conversations that you're having, and what are you driving internally?

[00:10:38] Andy: Yeah, I mean, it's a space that's changed so much. And again, I was talking to a peer this week about the last five, 10 years, and there are some really obvious places where there's change, right? If you look back, how did you schedule your interviews? You had coordinators, big teams of coordinators, maybe one coordinator to every three or five recruiting team members, depending on the efficiency goal that you were driving to, and it was simply because the coordination pain was so significant. I think there are two things that have changed significantly there today. One is the technology, as in the ATS, or potentially the products that were even online then but perhaps weren't as intuitive and AI-driven as today, are significantly improved. And then also, the tooling that we're all using being more modern allows the recruiter to do a lot of what ops would've done before. And so the interesting thing for me is there are two parts to this. There's the efficiency side, and then there's: what does that mean for people who operate in talent acquisition, specifically in talent ops, over time? So if you think about the training ground, historically speaking, you had roles which could still be specialist, like a sourcer. It was possibly an entry-level role, but it also had this career trajectory upwards to being a senior sourcer or a principal sourcer, and you could go down that career route, but you had the space to build that within the craft, because you could enter at a more junior level. The reality is there are some tools now where, with the agentic capability on the identification side, you can put that in the hands of a recruiter just as fast, or you can automate that through your hiring manager, or even put that in the hands of your business leaders. That significantly impacts the entry points into our functions, and it means that not only are they smaller, but our ability to develop craft skill is going to change. Again, if you think about that recruiting coordinator side, today I run a team that has no coordinators. I don't think I could have had a 10-person recruiting team even two or three years ago that could be efficient and not have coordinators. That's a tooling-based decision. Now, there'll potentially be a tipping point at some point, but it means there are roles that simply don't exist in our function today that would've done historically. Another space that I think is really interesting from a specialism perspective is employer branding. You can work so much faster now if you've got a marketing team and you have recruitment. Between the two of you, you can move incredibly fast. I don't think it replaces the value of a true employer branding specialist. That's always going to be great. But again, if I go back to my time at Atlassian, when I would've had a large branding team, I would've localized that branding team to be able to meet the market where it was. That feels like an incredible luxury today, right?

[00:13:19] Adriaan: Yeah. And I'm smiling because I spoke to Dave Hazlehurst today. Do you know Dave?

[00:13:25] Andy: Yeah.

[00:13:25] Adriaan: Yeah, so for listeners, Dave is the former partner of Ph.Creative, one of the more well-known employer branding agencies in the world. He left that, and he's now building an agentic-first employer branding company, because he says that for one of their larger global companies, it took them eight to 12 months of just doing the research alone before they were able to do anything. He says, "Now we can do this in weeks, in terms of the speed and what we're able to achieve."

[00:13:56] Andy: Yeah, I mean, it's such a great point. We launched a new website at Elliptic last week. Two things really stood out for me when we did that. We did that in a week when we had what our company calls our big day of rest.

[00:14:08] Adriaan: Oh, you've got to tell me a little bit more about that. Yeah, we'll come back to that point. Yes, go on.

[00:14:12] Andy: So everyone has a day off on the Monday, and even though there was a deadline for a new website on a Thursday, people were ready because they prepped it and they were good to go. Now, we had a new careers site as part of that as well, and we've seen an incredible spike in our application rate in the 48 hours that followed. But the other thing that's so different is we're reading, in real time, what people are consuming, and within the first two to four weeks, we'll be able to iterate and update that website. Whereas if you think back to where Dave would have been working eight, 10 years ago, you're doing static research to create something that you then update at a point in time. And you haven't guided me there, but let's take that step into the human side of what we do. The advantage that everyone has now is that you can produce content that looks beautiful and tells a story. The difference will still be the humanity of it, which will be: when I scrape away at what you told me were your values and behaviors, or your key operating principles, or the things that you stand for as a company, when I then come and interview with you, are those real or not? That, I think, will still be the fundamental difference: do you tell something truthful about your business and your organization? That remains the hook that gets people excited, and you can make that stand out from everything that's auto-generated and just, "That's what we're supposed to say." A bit like when award season comes around and people get excited about being best employer of the year.

[00:15:40] Adriaan: Yeah, yeah, yeah. So good, so good. And it's funny that you say that, because I don't know if you've experienced it in your company, but I see two movements happening with AI. On the one hand, especially some of my fellow founders, including myself, spend hours on all the stuff that we can now build, and we're almost getting into this loop of never-ending prompting and building things, and at a certain point it's like, "Wait, is this still efficient? Is this still what we need?" Right? That's one. And then the second part is that we've now had to implement a rule within our management team: you are only allowed to send anything AI-created, in terms of text or documentation, if you specifically state what part has been yours and what part has been AI-generated, and also give context there. Because every time I'm on the receiving end, when someone has sent me a beautifully designed doc where I immediately understand this is AI-generated, I almost lose interest and/or am very skeptical of reading it.

[00:16:39] Andy: Yeah, completely. I think it makes sense. If you look at Anthropic's four Ds when it comes to AI, everyone's managed to move fast, right, on delegation and description, but the thing that makes someone stand out is discernment and diligence. And I think fundamentally what you're describing is true of work, of interviewing, of presentations, but also of the things that we're going to choose to engage with, right? We're not going to choose to engage with the four things that look exactly the same as the last four things we saw. We want to have that context of how somebody came up with that idea, why they're sharing it, why they think that idea is valuable, which I'm sure, in your leadership meetings, is really what you want to understand from someone, right?

[00:17:22] Adriaan: Exactly, exactly. And it's now almost working against you when you use it, because the message is differently received, versus when you created it, it always feels like, "Oh, wow, this is a great piece of content. I'm so excited to share this with the world." And to your point, right, we launched our website. It took us a month and a half to build, which normally would've cost us six months, but the big differentiator was that our head of marketing, who had never built a website in his life before, built the entire website from scratch, which saved us 15,000 to 20,000 dollars on the website building costs alone. And now, to your point, we can update everything in real time. It's super easy to update everything because we're in full control.

[00:18:08] Andy: Such a strength.

[00:18:08] Adriaan: What do you see? One of the things that I see a lot of our clients actually dealing with is that because of AI, the amount of inbound applicants has tremendously increased, because people are using AI to create amazing resumes, and also to apply, to easily apply. Is that something that you are experiencing as well? Is that something that you're looking into, of where AI can help or not help? And do you maybe see a role for AI interviewers in the future as well?

[00:18:40] Andy: Yeah, I think it's a fascinating area. I think I'm probably slightly cynical around this. I don't think the problem is that different to what we saw two, four, six years ago. I think volumes are high, but in many cases for the same natural reasons that they often are, in terms of the strength of the market and the geography of where you're trying to hire. And certainly my current experience is that where I'm seeing application surges, they're very typical of roles and locations where you would expect it. For instance, a software engineering role in the United States will drive a heck of a lot of inbound applications, often from people who don't meet the working rights requirements, or the skills and experiences that are required. I don't think that's changed hugely. The interesting gap that you're talking about around the application piece is that fit piece: has this person tailored what they're saying they can do, or what they believe their skills are, specifically to the job description? And certainly from what I've seen in recent times, not in huge volumes, but when you're screening, and yes, I do still screen CVs, there are tools there that are helping give me some indications around the match between the skills and the requirements. But you still have to make that human decision yourself. Am I seeing mistakes where someone literally leaves the text from the job description in their own CV? Absolutely. Am I seeing prompts in CVs that say, "Make this fit job description X, Y, Z"? Yes, we are. Are they super prevalent? No. They stand out, and you react like that because they're not there all the time. And I think that capability to really match the two, the role to the candidate, and assess them is the hardest part of everything that we're doing. So when you then talk about the AI interviewer piece, I think it's a really fascinating area, and there are two different parts. I think people can get stuck in each of them a little bit. Part one is interviewing for AI fluency: are we hiring people who can use AI, and discern, and be diligent, and assess risk against our own expectations, and explain that effectively in an interview? And again, not everyone's doing that, or not everyone's doing that all the time, but we've certainly built an expectation into our own job descriptions and adverts around our expectations for a role that we feel are reasonable, and then we're assessing for that. And I think that will become more targeted over time, and clearer. The AI use in the interview is the piece I find fascinating. Personally, my view is I'm ready to look at these tools, but I'm ready to look at them from the point of view of optionality. I do not like the idea when we're hiring for people, when we're competing for high-value talent. When you are in a market saying that you're a bar-raising organization with high talent density, you need to meet those people at the same place, and there's a human element to that, in my personal opinion. What I do like is AI for clarification. When you think about all of those inbound applications, what do you do with the 10% where you're not sure about fit, simply because they weren't well written or they didn't explain themselves? Is that the opportunity to go back with an AI interview, just literally asking a few questions? Clarify A, B. "Tell us about your fit. Did you tell us about your location?" Simplifying something without needing both people to be immediately available. And I think it works as much for the candidate. You know what? I want to answer things like that on my own time. I don't want to have a call booked in the middle of the day with someone who happens to be available, and go through maybe the pleasantries of a Zoom introduction. It's a three-minute job, and I want it to be a three-minute job, but I want that to help the fairness of how I'm assessed. Then similarly, where I think this is going to be interesting, again, it's an extension of that silver medalist piece, will be: do you want to move faster? If I can offer you the chance to have an accelerated process, because you can do this AI interview that has the same content as the interview, I'll tell you what, you can do it this Saturday morning, on your own time, 24/7. But should we be telling people they have to do that? I'm not at that stage yet.

[00:22:57] Adriaan: So would you then give candidates the opportunity: "You can either do the first interview through an AI interviewer or with one of our recruiters. With the recruiters, it will most likely take a little bit longer before someone is available." And give candidates those options?

[00:23:16] Andy: I think I can definitely see it on the roadmap. It's something that I'm thinking about right now. I'm talking to vendors right now. And again, I don't know: is it the recruiter screen? Is it possibly the places where you actually tend to see the blockers in the process, which can be a technical interview sometimes, a sales interview, could even be a hiring manager screen? I don't think it's going to take away from the amount of human connection within the overall assessment structure, but I think it can plug some of the gaps and put clock speed into the system at the places where it typically doesn't exist today. I also think it's probably going to create stronger signals, because with these tools, I would assume you're typically training them against your own interviewer data as well. I think it's going to be fascinating to test: where does fairness exist? Because fairness, to a lot of people, needs to be in the eye of the human. But actually, the human, so long as they're making a decision, are they going to make slightly different assessments where they haven't been doing the prompting and building a relationship sometimes, and they actually have the rawness of the transaction data and the transcript, which is fairer on both parties? I don't say that with a particularly strong view yet, but I'm fascinated to see what, frankly, true research and academic research is going to tell us about that in the next three to five years.

[00:24:34] Adriaan: I mean, it's the question that I constantly ask myself, right? I remember vividly, there's this famous clip of people being interviewed on the streets when the mobile phone came out, and they ask, "Hey, would you like to have a mobile phone?" And everyone says, "No. Why would I want people to call me? Leave me alone, right? I have a phone at home. There's no need for it," right? And am I looking with my current knowledge of, "No, I want to talk to a human," versus three, four, five years from now, the technology will be so good, and especially the adoption of AI will be so far advanced, that I will find it very normal that we'll be having a conversation through that lens? And what does that then mean for my business, right? Because that's the big question for me: is there then going to be less need for recruiters? Is the role of the recruiter just significantly going to change? Is it going to be further down the process? Or is the process so far automated that the first steps are all done through AI, and only the hiring manager gets to speak as a real person?

[00:25:37] Andy: I think that part's really fascinating, because, well, there are two things that I like to throw in around this conversation, in broad terms. So the one is: who's doing what? So let's go and look at Anthropic, let's look at OpenAI. Their business model, to all intents and purposes, their headcount model externally, when I look at it, the way they're building recruiting teams, appears to be the SaaS model, the FAANG model. So if we are all heading in a very different direction, but we should assume that they probably know things that we don't know today, what is that, and why? So I think that's something really interesting to research and understand. And it goes to my second point, which is that, to the question you've just asked about more or fewer recruiters, I haven't met a talent leader yet who has found this answer where AI has increased productivity through the roof per recruiter, to the point where that changes that dynamic in that model. Where I'm seeing it is about the time the recruiter spends and where they're spending it, and the value of those activities. So whole functions, as we talked about a little bit earlier, are definitely getting smaller, but the actual individual recruiter who carries the rec load, I don't see their productivity necessarily increasing. I think they're doing more things. Some of that maybe was done by other people before. And there's also a question about whether they're then spending more time being a true advisor to the business, which I think is the goal we all want in how we're using our time, as much as we're just increasing the volume of hires the recruiter can make. And the asterisk that goes next to that, as well: recruiter capacity is informed by business capacity, informed by interviewer capacity. The recruiter on its own, supported by AI, doesn't suddenly solve for those other two things that are bottlenecks within the hiring system.

[00:27:36] Adriaan: So true. Which is often overlooked by the business planning, et cetera, right? And that's constantly what we need to push back on, or at least give clarity on. How do you look at working with vendors, getting outside tooling in, for now, and also having the opportunity to build a lot of stuff on your own? You are an engineering company at heart. You are in this interesting field where you could make a decision: "Hey, maybe we should develop this in-house." What's your take there?

[00:28:04] Andy: Yeah, I think it's interesting. I've worked for companies in my career that would maybe describe themselves as sales-led, product-led, engineering-led. And we're a product-led organization, and I think I always look to the business first, in terms of: how are we helping the business achieve the goals of the business? And there are two responsibilities I think you have there. One is: is the work that we're doing the high-value impact things that help us achieve those goals? And then the second thing is: are we applying our resources, which is dollars, tooling, people, to help us do that? I think the really fascinating thing for me at the moment is, one, if you're buying, are you buying a tool that's going to sustain? I've always had a view that you don't want a tech stack, a tooling stack, in the people team, not just in TA, where you have dozens and dozens of vendors, competing solutions. You're duplicating, you're removing efficiency, you're losing speed, probably through integration success or failure. You should really be picking the things that are high value and allow you to operate effectively, but around a hub. Now, going to the buy versus build piece, a few years ago you really only had two hubs, I'd say. You had your HRIS and you had your ATS. I think the build piece then was, "Well, what really integrates with those? Or where can engineering give me some capability?" So I've been in organizations where there wasn't a strong enough tool for interviewer pools, so we built it. But eventually it wasn't maintained, so it didn't scale with the business. That, I think, is a really fascinating example of where that will change, because we can build that now, and it will scale more effectively, because the maintenance element is a lower bar. But I still always come back to: what's my hub, and is that effectively insulated from market risk, for want of a better phrase? If I'm going to go and buy an AI interviewer, I'm going to buy one that is run by and associated with an established product, that has a foundation in some form of that tooling and operating model, that gives me some form of certainty that in two years' time I'm not having to buy another vendor again. Now, that's really hard for a market where you've got lots of new AI products coming that are all tremendous solutions, but the reality is their funding and their life cycle don't tell us that they're going to be here in a year's time, or that their support is going to be available at the time of day I need them to be if something falls over. So I think it's a balance of those things. And then the final piece with the build is: if you're building, that is still time. So is that the best use of time and of those dollars? And that's before we even get into the cost of tokens and that. We're fortunate: we all have Claude, we all value it. I use it daily. I would say most of my team do. I already hear of peers saying, "Oh, we've pulled back some of this tooling," or certainly more advanced frontier models, from teams outside of engineering. So you then get into: well, should we have gone down that line, or should we have been more simple and just looked at automations through Zapier or something, and worked on that lower-hanging fruit and built an integrated system? In simple terms, the integrated system that serves your business need, by both cost and quality and output, and gives you the right data, doesn't matter whether that's today or three years' time or four years ago, should be the philosophical principle that makes your choice, for me.

[00:31:46] Adriaan: Yeah, interesting. We are now, for the first time, hiring a full-stack engineer, and we're a services business, right? Let me give you an example. On our HR system, we spend 15,000 to 20,000 euros, and in the end, what is it? It's a database where we keep our contracts and we keep people's data. And we're looking, and we can build our own, with compliance and safety, and then connect it with all the other systems, arguably for a fraction of the price, right? We're careful about what to do and what not to touch, but all of us are having these conversations, right? Which is fascinating, absolutely fascinating. For you, Andy, in terms of the organization, and this is something I'm constantly asking my guests, and I also don't know if you have the answer for it: in terms of org chart design and how that's being thought of, has a change happened in how your organization is looking at that? You've raised a tremendous amount of money, right? We discussed it a little bit in the beginning: the SaaS playbook of, "Well, this amount of money, this sales team should be adding this much revenue. With this much revenue, we need X amount of engineers," right? In terms of how that scales. With AI, with agents, with tokens, what are the discussions that you see happening, and what is the same, what is maybe different?

[00:33:05] Andy: Yeah, I think so. I probably haven't been here long enough to give you a specific answer, but there are probably four trends that I see consistently that I think are playing out everywhere, and I'd include us in this. So first of all is spans of control. I think broadly, over the last three years, and increasingly with AI and agents, there is now an assumption that you will want to have fewer layers between your CEO and your frontline delivery teams, in whatever function they are, and therefore managers are going to have broader spans of control. Which, again, if you roll back to five, six years ago, you had these playbooks that said, "Someone's got five direct reports. We assume we keep growing. We now need to hire another manager, split it, to grow again." I haven't heard someone say that to me in a very long time, and quite the opposite. People are saying, "A manager will be responsible for a team and work, and that isn't just the team of humans. There will be an agentic element to that." Are people there yet today? Maybe not, but I hear people talking about that from an org design perspective. The other parts to this, in terms of principle, are geography and support functions, which are really interesting. There was always a discussion about: are you near, or are you offshore? Some of those assumptions were about being near to the customer, or language skills, that came into play. I think there are organizations where they can see the ability to break that, and where an agent alongside a human can create 24/7 coverage. So customer support's a classic example where you would see this: you're not going to have to create the same follow-the-sun models that you would have done a couple of years ago. That, as well, leads to a really interesting footprint discussion about where you need to be and why. What systems do you engage with, in terms of employment law, countries, different regions, to service them effectively? And that's certainly something I'm hearing people say: you create operating cost and inefficiency the more places that you're trying to operate and the more places you have headcount. How can I simplify that and keep that more consistent? And I think that's a space where people are looking at that agent augmentation piece a little bit as well. And then the final thing I'd say is the change, the big change, is the speed. The planning process for headcount planning and for org design was typically a multi-year horizon. I was part of processes where we were planning headcount two, three years in advance, and building engines to do that, and thinking about promotion cycles and what that meant for the grad scheme in two years' time, and what that meant for how many managers you need to buy versus build. I don't hear anyone talking about headcount with a two-year horizon today. People are looking at this quarter by quarter, and are probably looking at 12-month horizons maximum. And that's, I think, in part because of uncertainty around what AI's going to do. I think in part it's just the speed at which business is changing, and frontline revenue is changing, that you want to be much more nimble. You don't want to risk sticking to a plan that could be outdated.

[00:36:13] Adriaan: But that's interesting, right? So how do you deal with that level of volatility and uncertainty? Because if you can plan two, three years ahead, I mean, that's the dream, almost. But every quarter, headcount up, headcount down: how do you deal with that capacity planning internally?

[00:36:28] Andy: I think, for me, I've tended to veer towards the side of saying, "Let's throttle a little bit of the growth. Let's keep productivity as the aim." And actually, for me, I think of people as a system, right? And so coming into a business is important. We want to hire great people. We want to build our talent density. But the reality is there are three phases that come after this: there's onboarding, then there's ramp, and then there's a period of high productivity and high impact. Now, if you're spending all of your time thinking about your hiring engine on the front end, you're really failing your business in terms of its ability to do the things it says it's going to do in a six to 18-month horizon, particularly if you're in a business where you're adding 50% or 100% headcount over a period of time. So if your business goal and business strategy is about market share or about revenue, I'm putting my pin in on: well, how are we our most productive per headcount in a year's time? And in the old model, as we talked about a little bit earlier, it came back to: how many people do we have? Today it comes back to: how do we not just find the great people, but how do we unlock them fast so they ramp and are productive faster? And then how do we get them to a higher stage of productivity, and then maintain or raise that bar of productivity? And I think that gets super lost sometimes. But actually, if you build a system that focuses on that, and then you throttle the inbound and make it contingent on those outputs, I think it's always slightly easier to build up your output from a recruiting sense than it is to ever switch it off, and we've all learned the pain of that over the years.

[00:38:14] Adriaan: Yeah, that's interesting. We actually had tremendous growth in Q1, then we slowed down a little bit in Q2, and now we see tremendous growth again. And I was having a one-on-one with our head of recruitment internally, and her challenge now is to get the pace back up and get everyone really running like we were in Q1. It will happen, but it's interesting that that's very front and center with us right now.

[00:38:37] Andy: Completely, completely. And it just saves that shock for all founders like you and all chief execs: you don't want the conversation of, "We built the company, we have the foundations. Oh, but revenue's off. Now we've got to focus for a quarter just on those core fundamentals." And so I think that's where it just becomes such a key part, for me.

[00:38:55] Adriaan: Yeah. Andy, tell us, what excites you? You've been how long with the company now?

[00:39:01] Andy: So I'm just over 90 days. I'm about three, four months in.

[00:39:04] Adriaan: Yeah. So fresh off the boat. I read this, I can't remember what the name of the management book is, but it basically states that your first 90 days, that's when you usually make almost the most impact, because you're fresh, you come with new ideas, you're not strangled yet by any sort of rules or preconceptions that exist. So tell me a little bit more about what you're excited about. What's happening in your world?

[00:39:28] Andy: Yeah, I think it's so interesting: one of the things I've had to do here is pull back a bit from that early kind of "make decisions and move." We're a super thoughtful organization, so I've really appreciated having the space to really do the research, and to look, and to understand the org. That's really hard, though, when you're growing, because I want to be in there. I think when I was coming here, and I talk a lot about this, there are three reasons why, for me, you join an organization. You've got your values and your behaviors, or your principles, and you want to see that those things that someone puts on a website or a career site are true. Then you've got your people, your leaders in particular, the people you interact with every day. And then you've got your product-market fit. And if you get those three things right, and I mean right for you, because everyone's different, then you should always be excited by what's in front of you. And so for me, the great thing has been coming in and seeing these things are real, right? Our values, when you go on our website, are real: people's curiosity, people's support, a group of people who say "team before I." The leadership team, I knew, was going to be amazing. The fascinating thing then, and I come back to this because I was talking a little bit about this earlier, about the business connection to the people experience, is that we operate in a market where we have a moated opportunity, a regulated market that's growing, with a buyer base that is being compelled to buy us or a very small group of competitors. And so one thing that excites me on joining is getting under the hood and seeing, "Hang on, this product's incredible." We have a first-mover position in a market which has tremendous value to be successful, and that, to me, is incredible. So then, to be able to go to candidates and actually say: this isn't something in a world where you're being told about this AI product, or "if this comes off," or "maybe you get super rich if this is successful," but you're placing a bet in a sea of different AI things that may or may not work, that all have 20 competitors, and one's going to win. Being somewhere where you can say, "Look, you're going to win." So now it's about: what's your role going to be in winning, and what's your point of impact going to be? And I think when you're on a people team, that's so compelling, to go: right, well, what is our role here? How do we help people perform to the best of their abilities? How do we help hold people accountable, alongside the rest of the business, to really accentuating those values? How do we close that ramp gap at the very beginning, so that the people of tomorrow are as effective as the people who have been here for a longer period of time, because we're willing to learn from that? And naturally, because we're growing, we get to do that through both hiring, bringing great people in who you see have impact, as well as that ability to influence that downstream bit. And so that's the excitement. I've always found the conversation we were just having, about the headcount planning and business planning, slightly hard, because I've always had this view that we shouldn't just live in the moment in recruiting. We should always be thinking about: where are we going to be in a year? Are the things we're doing today building us towards where that's going to be? That's a really fragile space, because over the last three years we've gone through a downturn in the market. We've seen redundancies, great friends lose their jobs. That feels tough to do. And then on the flip side, we're coming through this period now where it's like, well, everything's happening so fast. But I still think of that lens of: what are you doing today that tells you that you had impact in six months' or 12 months' time? And that's the thing that's exciting me right now: the people we're bringing in, but having the foundation of new tools, new team. We're all pretty new in our function. And this market opportunity opening up in some new geographies: we've hired in Miami and Hong Kong this year. Those are exciting to be a part of, and I just love doing that.

[00:43:15] Adriaan: So good. Andy, this was such a pleasure. It was so great talking to you. I completely forgot the time: 45 minutes. This is so good. Andy, I appreciate you. What's the best way for people to connect with you? LinkedIn, I assume?

[00:43:28] Andy: It's always best to try and connect with me on LinkedIn, and then don't be angry when it takes me too long to accept the connection and it goes into that sea of LinkedIn horror. But yeah, go there. I look forward to connecting with folks. And Adriaan, as always, it's been incredible talking to you. I love speaking with you, and love listening to you and your journey as well.

[00:43:44] Adriaan: Awesome. Thanks, Andy.

[00:43:45] Andy: No worries.