The New Blueprint for Building Talent Teams

Reece Batchelor — Talent Acquisition Manager, Wayve

I joined my previous company, Dust, as a first talent person, a blank canvas with no playbook. I came into that role with a lot of assumptions about how AI and the product we were building at Dust would change the way that we build talent acquisition. And over time, I realized that many of those assumptions were incorrect.

And I've recently just started a new role at Wayve, and whilst the scale of the company is much bigger, what I've realized is that we're all trying to solve the same challenges. So the lessons I learned at Dust ring true in this next chapter. So I want to share with you the assumptions that I had when joining Dust, what challenged them, and how I'm approaching this challenge differently with Wayve. Now the truth is that you don't know what you don't know, and in a world where AI is changing how we work so rapidly, what we've done in the past will be so different to what we do in the future.

But looking back, there are three assumptions that had the biggest impact on how I approached building the function, and those assumptions were: AI means we could build a lean team; if we build with AI, the ROI will naturally follow; and lastly, we can hire for rec ops later. Now, as a team of one, I thought that I could handle significantly more hiring whilst outsourcing a lot of the work to agents through sourcing tools and other agents that we were building.

I was partly right. The ground-level repetitive work was really good when it came to AI, but I misunderstood what it meant in practice. The capacity didn't just disappear as a whole, it just shifted elsewhere. For example, interview hours per hire are climbing, and this is due to the competitive landscape that we're in right now.

Also, roles are changing and new roles are emerging, and this requires more human hours, not less. The other reality check is that human connections cannot be automated. When I joined Dust, we had an offer acceptance rate of around about sixty percent. And when I dived into the data, the candidate feedback, and looking at the systems that we had, it was really clear what was missing, which was humans in the loop.

So we made a change. We went back to the basics. Talent took ownership of the full candidate journey. We spent more time with candidates, preparing them for interviews, giving them a space to ask questions outside of an interview setting that they might not feel comfortable asking otherwise. And what probably seems like such a basic change for most of us in here, it had a significant impact, and our offer acceptance rate increased to above ninety percent.

The lesson I learned here was that AI can absolutely increase the output and the quality of our work, but it does not remove the need for human capacity to hire well. Now growing up, I got constantly unwell. I was in and out of hospital, one antibiotic after another. And it wasn't until one doctor stopped and asked, "Why do you keep getting ill in the first place?" that I found out I was celiac. And after reluctantly changing my diet to gluten-free bread, the symptoms disappeared. Now I know you're probably all thinking, what the hell has this got to do with AI? But here's the thing. A lot of us feel we need AI to fix our problems when actually we need a behavior change.

Without the behavior change, AI is useless. And I saw this when we tried to fix a problem with our referral process. Referrals wanted more visibility into what was happening in the process. So we built an agent to do that. And what happened was the agent didn't work as we were expecting. And when we looked into the reason as to why, it's because the pipelines were not being updated correctly and the quality of the feedback within the ATS was not good enough.

Until we fixed that, the AI was never going to work. And that leads me to another lesson, which is there is no adoption without enablement. Everyone's at different stages on this AI journey, and just because you can spin up an agent and understand how it works doesn't mean that everyone else is going to follow you.

You need to take a step back, meet people where they are, and take them on the journey with you and help them understand how it fits into their process as well. And the final lesson for me here is that not everything you can build is actually worth building. And I saw this when we created an agent to do pipeline management and forecasting.

The V1 worked really, really well for us because what it did was it organized the data and helped us connect talent with the business. But then when we pushed it further to help explain why a role was off-tracked and predict the future for us, we spent so much time going back and forth with the agent constantly to correct it.

What I learned was that AI can organize the signals, but it cannot replace strong human judgment. So being more intentional about what you're building with AI is how you move from lots of individual experiments to an infrastructure that actually makes a team more efficient and effective. And this leads me to my third assumption that we can hire for rec ops later.

The more I thought about it, the more I realized that we needed someone to own the build work much sooner. And the benchmarks I had previously used were based on Ashby's customer guidance, which is around about three hundred employees. But the truth is that the number of employees doesn't tell you how mature your talent function actually is.

The systems that you need to connect, the infrastructure that you need to build, and even consolidating your tools in a platform like Ashby, there's still so much happening that someone needs to own this. We were tripling the team at Dust, but we're also trying to build a scalable function for the future.

I thought everyone could just split their time versus build and run, but it just ended up spreading Dust way too thin. And that's why we decided to hire for ops earlier than the conventional benchmarks. And the thing is, the business case wasn't just talent needs more support. It's that with better systems, we're going to create operational leverage for the entire business.

And that led me to think about the talent function in three different areas. Recruiting owns the search, candidate relationships, and the hiring manager partnerships. Rec ops owns the systems, data, process, and governance. And an AI operator owns the AI. And it doesn't necessarily mean three separate hires from day one, but it does mean three areas of accountability.

And that to me is how you build a leaner team. Not by asking everyone to do everything, but by giving the right people ownership of the right work. Now, as we know, talent is changing so quickly, and the behaviors I believe matter the most right now are AI fluency and relationships over volume. Now, AI fluency is such a taboo, and what the hell does this even mean?

Well, for me, it's not just about whether someone knows how to build on Claude. It's about whether you can actually help the business change with the work that needs to happen next. We have this unique position in talent where we have this view of the company, we can see all of the roles that are being hired for and what's emerging and what's coming next.

What I want to know is how curious are you about how AI is changing the work and the roles that you're hiring for? Are you staying up to date with what's actually happening? Are you influencing teams and roles that are going to be designed around that? And that's what AI fluency means to me. Talent shouldn't just be adapting to the future of work, we should be helping shape this.

At Dust, beyond that, we use the same framework to measure success for every single candidate we hired and every single employee, which was aptitude, attitude, and agency. Now aptitude is simply can this person do the work at the level that we need them to? Attitude was does this person make the people and the team around them better by being in it?

How they approach change, ambiguity, collaboration and feedback. And then lastly is agency, which is about does this person make progress without having to be told? Do they take ownership and do they find problems without people having to surface it to them? Now all three matter, but attitude and agency were the biggest things for us.

That prioritization matters so much in this world right now because the work that you did in the past is changing, and the people that are going to be successful in this environment will not just be able to do the work as it exists today, they're going to be able to do it as it exists tomorrow. And then lastly, relationships over volume.

Now, I've always been a big believer of sourcing, and I still am. For a long time, the maths was simple for me. If we reached out to more people, more people would respond and the hires would follow. But the market has changed massively. Everyone's fishing from the same small pond. So I wanted to understand what we could learn from the best agencies right now and how they're attacking the market.

So I reached out to the guys at Sequoia, and they gave me an analogy that really stuck with me, which was when you are buying a very expensive house, you don't go searching on Zoopla and Rightmove. You find one estate agent that has access to the right houses. You give them your requirements, and you trust them to filter the market for you.

And this is exactly what's happening in the hiring market right now. The best candidates are receiving so much noise that they need people to filter the opportunities for them. So the question for talent teams is: how do we create this same level of access and trust with candidates? And for me, the best talent teams will start to operate like executive search firms.

They'll be building relationships before you need them, understanding the networks around your roles that you're hiring for, and making it easier for the right people to introduce the right candidates to you, and that's not just internally, that's externally as well. So to come full circle, I would encourage you all to do this exercise.

Map your talent function across three areas: recruiting, rec ops, and AI operator. Ask yourself who owns each of these areas, where should humans stay in the loop, and what should be outsourced to AI? AI is exciting. It's going to change the way that we operate. It's 100% going to make us more efficient and increase the output of our work.

But it won't remove the need for judgment, ownership, and human relationships, and in many ways, it's going to make these things even more important. And for me, the teams that win will not be the teams that automate the most. They will be the teams that are deliberate about where AI creates leverage and where humans create trust.

Thank you.