In 2021, Trent Cotton was leading talent acquisition, and he jokes that he hired anyone with “a pulse and a prayer.” The hard part was finding enough qualified candidates.
Five years later, the candidates show up, and the hiring process loses them after the interview. “The pain is in the middle part of the funnel,” he says. And that middle doesn’t belong to recruiting.
Trent is now Head of Talent Insights and Analyst Relations at ICIMS, where he interprets hiring data for the practitioners who have to act on it. Before that, he spent about two decades on the operating side of talent, including a long stretch in finance. He wrote Sprint Recruiting and hosts The Human Capitalist podcast.
He explains how candidates go missing in two places that sit with leaders: how fast they give feedback, and whether the role was settled before the search began. The first one is entirely on engineering leaders.
Slow Feedback Costs the Hiring Process Its Best Candidates
ICIMS’ September Workforce Report shows job openings up 13% over last year, while hires are up just 2%. ICIMS’ Frontline Hiring Report asked candidates and hiring managers directly where the pain sits, and both pointed to interviewing and screening. “It’s almost like the hiring managers and candidates are looking to recruiters and the HR tech and going, for the love of God, can you solve this problem here?” Trent says.
According to Trent, 32% of the drop-off happens after the interview stage, partly because systems move too slowly. Candidates interview at several companies at once, and in an unsettled market, “whoever gets the first offer, chances are that’s the one that they’re going to take,” he says. Every day a hiring manager holds onto interview feedback, a competitor has time to make that offer first.
Some Roles Get Pulled at the Offer Stage
The second decision starts with the posting itself. Trent hears from companies that post roles while they’re still deciding whether those roles should exist. During his years in finance, the executive team would approve a posting and then attach a 16-step approval process at the end.
“And about half the time, you would have the candidate be ready to extend the offer, and then someone in this position control group would come in and go, ‘Yeah, we’re not going to fill that role,’” he says. He sees the same hesitation today as companies weigh AI’s impact on the workforce.
If a role is uncertain, companies may not be ready to make a permanent hire. Settling the headcount question before the search begins protects candidates and recruiters alike. When the answer is still unclear, a flexible staffing model can carry the work while the talent acquisition strategy catches up.
Sprint Recruiting Starts With the Roles That Can’t Wait
Trent’s fix came from six weeks of Agile and Kanban training that his CHRO sent him to. He says that he went in “kicking and screaming”, but left convinced that recruiting could borrow the same systems.
Leaders start by naming the roles the company can’t go without over the next year. Recruiters work those roles in two-week sprints, ranked by points. Once they submit five screened candidates for a role, they stop and wait for feedback.
That pause breaks a cycle Trent knows well. “What I found is that a lot of recruiters, they submit candidates, and then they go, and they submit more candidates, and then we get angry with the hiring manager because they have FOMO,” he says. “Well, we’re creating that.” For an engineering leader, the sprint moves only as fast as their feedback.
Getting candidates in the door used to be the hard part. Now they reach the interview, and the middle of the funnel loses them.
Trent says recruiting is limited by two things: the recruiter’s focus and the hours in a day. Fast feedback and a role that’s settled before the search starts give recruiters both back. Fix those two, and the candidates who make it to the interview stay through the offer.
Transcript
Trent Cotton:
What I found is that a lot of recruiters, they submit candidates, and then they go and they submit more candidates, and then we get angry with the hiring manager because they have FOMO. Well, we’re creating that. So once we would get that feedback, then we would move it to the next part of the process, and then we would continue to source and screen. Recruiting has two limits: the focus of the recruiter and the amount of time in a day. And you’ve got to work within those constraints.
Gemma Versace:
Hey everyone, and welcome to Keep Moving Forward, the podcast from X-Team for tech professionals who are passionate about growth, leadership, and innovation.
I’m your host, Gemma Versace, Chief Client Officer at X-Team.
Five years ago, the hiring problem sat at the top of the funnel. Companies couldn’t find enough qualified candidates. Today the problem has moved. Candidates are applying and interviews are happening, but people slip away before the offer.
AI skills follow a similar pattern. Workers are teaching themselves faster than employers can train them, and leaders struggle to tell real capability from tool familiarity.
Today I’m joined by Trent Cotton, Head of Talent Insights and Analyst Relations at ICIMS. Trent spent about twenty years on the operating side of talent, including a long stretch in finance and years as a talent acquisition leader. He’s the author of Sprint Recruiting and the host of The Human Capitalist podcast. He translates the data for the practitioners who have to act on it.
We talk about why candidates drop off after the interview, and how two-week sprints help a stretched recruiting team focus. Trent shares the screen-share exercise he uses to see how candidates think with AI, and why AI training can win over candidates better than a bigger paycheck. We also cover what a CTO should measure to know AI is paying off.
Let’s get started.
Welcome, Trent, to Keep Moving Forward. Thanks so much for joining us today.
Trent Cotton:
Thank you. I’m really excited about the conversation.
Gemma Versace:
Excellent. Let’s get straight into it. You ran recruiting and talent teams for two decades, and now you interpret the data for everyone else. Walk us through how you got there.
Trent Cotton:
I have always had a passion for numbers. I jokingly tell people that I trust numbers, I don’t trust people. And that’s after working two decades in HR. But I actually started as a banker. I put myself through school thinking that that’s what I wanted to do.
And so that’s where my love and my affinity for data kind of began, because it was something that I could trust. And I always believed that if you interrogate the data correctly, it tells you the right story. And I brought that over with me into HR. So whenever I made the career pivot, I was the ugly duckling of HR, one, because I didn’t have HR experience. So then two, I was always asking for data. That sounds like a great idea. Where’s the data to support it?
And that did not land well early in my career, but I’m glad to see that HR has finally kind of come around and people analytics has become a function within HR. As a talent leader, and even, you know, working as an HR leader, I was always looking at benchmarks externally, because when you’re just comparing your data to your own data, you’re only getting a fourth of the story.
And so one of the reasons that I was so excited about this job is that ICIMS does put out benchmark data that TA leaders can use to evaluate themselves, but then also create very successful talent strategies. And it sounds kind of cheesy, but I really believe that it’s our data plus our insights enables our clients and our prospects to be able to make talent strategies that are sound and based on data, which in my experience usually lends itself to be a little bit more successful.
Gemma Versace:
Amazing, excellent advice for those people listening today. We’ll be digging into ICIMS’ September Workforce Report today. So let’s start with the hiring market. Openings are up 13% over last year, but hires are up just 2%. Where are the companies losing people?
Trent Cotton:
You know, it’s interesting, ’cause if you look back in 2021, I was a TA leader at that point, and I always joke and say that I hired anyone with a pulse and a prayer, because the problem was at the top of the funnel. You couldn’t get enough qualified candidates. Now, fast forward five years, and the pain is in the middle part of the funnel.
In fall of 2025, we did a Frontline Hiring Report. And we looked at, from the point of view of the candidate and the manager, where do they see the pain? Because we hear from the recruiters, we wanted to have those two different vantage points. And it was really interesting, because both of them settled in on interviewing and screening. It’s almost like the hiring managers and candidates are looking to recruiters and the HR tech and going, for the love of God, can you solve this problem here?
And whenever I look at the funnel and look at some of the other data that we have, 32% of the drop-off falls after the interview stage. So it’s not getting candidates through screening. It’s something happens after that interview. And there’s a couple of things that I think, based off of the data, are happening.
One, systems are moving too slow. So you interview the candidate and they’re actively interviewing with other places. And in this market, whoever gets the first offer, chances are that’s the one that they’re gonna take, because the market is so unsettling right now. Again, compare that to five years ago, candidates would take their time, see who had the best offer. Those times are not now. So it is changing where the bottlenecks are happening.
The other interesting part, and I hear this a lot whenever I’m in market doing our AI Power Labs, is that companies are putting positions out there. They’re trying to, you know, get candidates, but at the same time, they’re also evaluating, is this position something that we actually need to fill? And in my experience, I spent a lot of my time in finance, and, you know, during times like this, it was almost like the executive team said, go ahead and post the position, but we’re gonna put in this 16-step approval process at the end that will kind of reevaluate whether or not we really want the position to be filled.
And about half the time, you would have the candidate be ready to extend the offer, and then someone in this position control group would come in and go, yeah, we’re not gonna fill that role. And so I think that there’s some of that going on just because of the great anomaly of AI and the impact on the workforce.
Gemma Versace:
Yes. And I mean, you wrote a book on running recruiting like an Agile team. What’s the first thing an engineering leader should change about how they hire?
Trent Cotton:
Well, it’s interesting, because the book, Sprint Recruiting, is for all of recruiting. My CHRO at the time put me through Agile, Design Thinking, and Kanban training, which was about six weeks. And I don’t sit still well in class, so I went kicking and screaming, but I came out of it and kind of had this realization. All right, if Apple and Facebook and Google are using these systems to produce client-centric products quicker, more efficiently, and higher quality, there’s gotta be some kind of application to recruiting.
So if I’m looking at, you know, I’ve got these critical roles in my organization, and I’m gonna speak from like the C-suite, these are the critical roles that are do or die for me over the next year in terms of making sure that my company is still around to compete. I need to prioritize that, because if we look at a lot of recruiting teams, they’ve been cut in half, if not more. The work has not moved, and now they’ve got this huge influx of candidates, and it’s causing a lot of consternation.
So the best way to do that is say, okay, one, what is our prioritization? And then two, limiting the amount of time. So rather than looking at, I’ve got 600 openings right now, what are the top roles that I need to get filled over the next two weeks? So you’re shortening that amount of time, and within that short amount of time, I want you to prioritize. And I used points. You can use high, medium, low. I’m a numbers guy, so I wanted points.
And what we would do is we would work those in order, and our goal was to get to a certain threshold and stop working on that position while we waited on feedback. So as an example, if I’ve got this top priority role, that’s a hundred points, my job was to source candidates, interview, screen them, and then submit at least five to the hiring manager. And then I would stop working on that position and then work on the next one in progress, or in the process. And what that did is it gave the recruiter some type of a mechanism and a process to focus on efficiency.
Because what I found is that a lot of recruiters, they submit candidates, and then they go and they submit more candidates, and then we get angry with the hiring manager because they have FOMO. Well, we’re creating that. So once we would get that feedback, then we would move it to the next part of the process, and then we would continue to source and screen.
And so if I’m looking at this market, that’s exactly the process that I would continue to use, because it’s forcing prioritization and you can report on it. Because at the end of two weeks, I could say, you gave us 10 critical roles and we got seven offers out in two weeks. That reinforces to that senior executive, I need to continue prioritizing to the team, because recruiting has two limits: the focus of the recruiter and the amount of time in a day. And you’ve got to work within those constraints.
Gemma Versace:
And shout out to your book, Sprint Recruiting, that you mentioned there. We’ll make sure that we give some details to the listeners who can seek that book out. Most job seekers describe their AI skills as ChatGPT, Copilot, or Gemini. How does a hiring manager tell familiarity from real ability?
Trent Cotton:
You know, I love case studies, and this is a great opportunity to do a share screen. So whenever ChatGPT first became popular, I was actually looking for a recruiter, and I had my initial interview screen with ’em. Everything went well. I had a hiring manager interview them just to kind of look at the relationship. But then I had a follow-up, and I said, I want you to share your screen. I’m gonna give you some data and I’m gonna give you some context. Just walk me through, like, show me how you use AI.
And I wasn’t grading them on, did they arrive at the right answer? I wanted to see, one, do they leverage AI to help them understand data? Because I think that’s kryptonite for a lot of people, and I think that’s a strength for AI. Two, I wanted to just kind of understand their thought process. Like, how do they prompt? What are some questions that they asked?
I let them chase rabbits. So there was one interview, I think he forgot that he was being interviewed, ’cause he was asking all these questions and chasing it. And I liked that, because it showed me that he’s intellectually curious. So I think this is a fantastic way that you can evaluate AI skills, but then also go in and evaluate some soft skills that are going to be absolutely critical for success.
One being critical thinking. The next is, how do you talk to machines? Do you understand the logic and the language that’s needed? And then three, having that strategic mindset. You know, that’s thrown around so much, but in that example, I understood something about his mindset. He wasn’t going to just take that first answer. He was doing the five whys, and I’m a huge proponent of the five whys.
So if you’re interviewing and you want to see how capable is this person on ChatGPT or Gemini, or, I’m a Claudine, I love my Claude, I would get them to share the screen. And again, you’re not evaluating them on, did they get to the right answer? You’re more evaluating, what’s their thought process behind it?
The other added benefit is, let’s say they’re 75% of the way there. That difference of 25%, as a hiring manager, you need to evaluate, is that something that I can teach them? Or is that part of their mindset, or how they go about things, or their soft skills? That’s not something that I have the time or the ability to be able to coach and develop. And I think that’s gonna help you make a stronger hiring decision in that kind of a context.
Gemma Versace:
Yeah, absolutely. And, you know, workers are teaching themselves AI faster than employers are training them. Why do you think employer training has stayed flat?
Trent Cotton:
Two reasons. One, there are two areas in HR that always get cut whenever there is an economic downturn: recruiting and L&D. I have never understood it and I will never understand it, but it is what it is. The second part of it is, I don’t know that a lot of organizations really understand what are the AI skills that we need to train for. So if they don’t have a target, you can’t develop curriculum to be able to offer to, you know, your employees. So I think it’s kind of a juxtaposition of both of those kind of fighting that’s creating this gap in the employer training side of it.
What I really like, though, is that the market is not waiting. You know, if I look at, unfortunately I’ve been around a lot of transformational technology, or technology transformations. You know, when everything first went to mobile, people were a little bit hesitant about it. This is the first technology I’m really seeing the average consumer and the average candidate really lean into it and really try to maximize.
And I think that companies are going to come to a point where they’re going to need to look at reevaluating their educational and experience qualifications. All right. Again, I’m gonna get back to, or kind of harp back on skills. You need to start hiring for outcomes, not necessarily experience. And I think this is a perfect example.
So if I’m looking at hiring someone that has got, you know, decent skills at doing consulting, and I want to evaluate their AI skills, I’m gonna do it in the interview like I just did in the last question. But the other part of it is, I’m going to look at how did they learn this? Did they teach themselves? Did they watch YouTube? I call YouTube the best free university that’s out there. That’s how I learned AI. I spent probably six hours this weekend. I listen to it while I’m at the gym. It’s a fantastic educational opportunity.
So, you know, we’ve been looking at, do you have a four-year degree? Do you have four years of experience? And assume that that equates to the skills that are gonna drive the outcome that we’re hiring for. We haven’t learned over the last 20 years that those two don’t necessarily predict success.
I think now AI is really transforming, and having candidates that take it upon themselves to learn this new skill, not waiting for a university, not waiting for employers to do it, they’re doing it on their own. I think that in and of itself puts that candidate ahead if I’m looking at an entire candidate pipeline, because that shows intellectual curiosity. That shows initiative, entrepreneurship, ownership. Those are the kinds of people that I want on my team, because I know that they drive enterprise value, hands down, usually 10 times higher than your average performer.
Gemma Versace:
Yeah. And this next question, I’m very interested in what your answer is, because we are seeing that some job seekers told ICIMS that they’d take lower pay in exchange for AI training. Is AI training becoming a new form of compensation?
Trent Cotton:
I think it’s becoming a new form of benefit, because candidates can go and assume what employers are looking for. All right? And that’s what they’re doing. They’re assuming that these are the skills that they need to train for. I do think that companies can layer on top of that and say, you’ve gotten thus far, here’s where we’re going to take you from beyond, and actually do like more application type of training.
So I do think that companies should leverage that as a benefit, offering certificates, offering apprenticeships. I really hope, I don’t know if I hope or I think, I’m kind of debating between the two, that apprenticeship is gonna come back, especially whenever it comes to AI, because there needs to be some knowledge transfer, especially with the generational changes that we’re seeing in the workforce happening over the next five years.
So we’ve got a lot of external variables that are affecting the workforce that I think is going to force companies to really reassess and go, okay, maybe I can’t compete with company A that is paying 20% higher, but I have made an investment. I have created these internal work groups that someone can join and solve problems leveraging AI and being taught, and I’ll certify them internally. I think those companies will actually be able to outperform the ones that are not offering that kind of training but offering a higher compensation, because the candidate is gonna see that as being something that’s a little bit more stable.
Think about a candidate now. Yes, I may want that 20% extra compensation, but that’s going to get me only so far. If I have a company that’s saying, I’m going to continuously reinvest in you and make sure that you have got the most crisp, most current AI skills that are available in the market, that provides me some longevity. And that’s a safe bet. And I think that’s where candidates are probably going to double down and make that investment back into the company by accepting their roles.
On the apprentice piece, I mean, think about, I’m assuming I’m probably a generation ahead of you. I’m getting old, but I learned so much that got me from where I am today by those apprenticeships. They weren’t formal apprenticeships. It was me riding in a car with a senior lender and listening to how he engaged with a client. Whenever I first got into recruiting, it was me going on those interviews with hiring executives and listening how they positioned the brand and how they positioned the opportunity. It helped me refine my tools.
And so those were, again, they weren’t formal apprenticeships, but I learned things that I didn’t learn in school. It wasn’t on some web-based training that the company offered. It was those things that you just get from humans. And I think the intensity at which technology is taking over certain aspects of work, I think that yearning, that natural human yearning to have human connection, I think putting in that apprenticeship is, one, going to kind of satisfy that yearning. But then two, it’s really going to be able to help with that knowledge transfer, that knowledge advancement, and evolving as well. Super passionate about apprenticeships, so we should probably change the subject.
Gemma Versace:
So ICIMS found that demand for AI skills varies widely across the US, the UK, France, and the Middle East. How does AI upskilling change when your team spans countries and employers?
Trent Cotton:
So I think this is where, if you’re running a global enterprise, using this kind of cookie-cutter approach really hurts you, because I always like to take a glocal approach. So I like global standards, but localization at the company or the country level. And I think that, you know, when you look at EMEA and even, you know, some of the other little areas that we looked at, it was fascinating to me. I mean, in all of them, finance was leading in terms of the AI skills saturation.
And so what we looked at is, a year ago we looked at a group of job descriptions and kind of notated how many times AI skills were listed in the job description compared to, you know, like maybe July of this year. And we kind of measured that change in terms of saturation. So did we see a higher concentration of AI skills in that same job over the course of one year? Where we saw the biggest hit was in the finance industry, in all of the areas that we looked at.
But then when you start looking at the different levels, or the different areas, it changed a little bit. And to me, that’s what’s so fascinating about our ability, is that we can look at kind of globally that it’s not the same data story. So you can’t have the same strategy. And that’s one of the things that I would encourage anyone that downloads the report, like, really dig into it and make sure that your approach to the AI skills race is looking at the specific data in a specific region, in your specific industry, ’cause that’s what’s gonna determine your success, not a cookie-cutter approach.
Gemma Versace:
Hmm. Yeah, some really good insights there. We have a lot of CTOs and engineering leaders listening to the podcast. And what should they watch to know the organization is actually becoming AI ready and not just rolling out tools? What are some things that you can provide some guidance for them as to what they should be watching regularly to be able to really see that their organization is AI ready and not just rolling out tools?
Trent Cotton:
The biggest thing, and this is with any technology, if you think of a technology transformation, and I’ve led several global projects, you put all of this time in the planning, you know, the evaluation of the product, the procurement, the legal, the risk, and then you implement it. And then you’re surprised that, wow, this didn’t really go the way that I wanted to, because no one’s looking at the people aspect of it. How do we get these people to adopt? What are their fears about this new technology that we need to overcome by giving them some exposure?
Again, that’s where these internal AI universities, you know, doing internal hackathons, it increases the exposure and it mitigates the trust factor that a lot of multi-generational people have, according to our surveys, with AI. Exposing and kind of seeing, all right, how do we do all of that? Because the true measure of success is AI adoption and then the ROI that you can absolutely track back from before leveraging AI to currently using AI.
And there’s a ton of data out there. It’s a little murky. We’re not seeing a lot of transfer, right? So I think McKinsey just came out with their State of AI report in September, and what was really interesting to me is that 80% of organizations reported a per-employee increase of productivity, but only 37% of organizations could trace that to the P&L. So there’s a huge leak going on. Like, all of this investment going into AI, but we’re not seeing at least a one-to-one transfer into the P&L.
So I think if 2026 was the year of operationalizing AI, 2027 is going to be evaluating and making sure that we’re able to track that ROI. So if I were a CTO, I would be looking at what percentage of my teams in the organization are leveraging AI, and track their productivity. You know, if it’s a risk function, have we been able to mitigate more risk? If it’s a fraud function, have we been able to close more fraud cases in a shorter amount of time? And I would look at that almost down to the FTE level, and then be able to really understand, okay, yes, AI tools are impacting our ROI and efficiency per employee.
Because if not, then that’s where you need to go back and go, okay, there’s some adoption issues. Maybe there’s some trust issues with the technology. Because I think one of the leading obstacles and transformations that needs to happen is going to have to happen in leadership, because leaders are now going to be expected to lead a hybrid workforce. And I think that’s going to be the do or die whenever it comes to AI transformation, AI adoption, and AI ROI.
Gemma Versace:
Yes. Yeah. Well, thank you so much, Trent. I have thoroughly enjoyed this conversation. We do have one final question for our guests at all times, and that is, what keeps you moving forward? What gets you out of bed in the morning, Trent?
Trent Cotton:
I, and I’m not saying this ’cause it’s recorded, I do have a dream job. I get to stay in data all day. I get to hang out and meet with practitioners and translate that data, and use the 20 years of good and bad decisions to help coach them. And I get to hang out a lot with analysts. So what gets me going in the morning is being able to have this cross-sectional view of what’s going on in the market, but being able to zero in on the particular areas that I think has the most noise.
So just an example, we did our 2026 AI Adoption Report in partnership with Aptitude Research. We released it in April of this year. And what’s interesting is, whenever I’m doing these in-market events, I always ask, all right, what percentage of recruiting organizations are leveraging agentic AI? Now, agentic is defined as do this like me, do this for me, do this with me. No one can guess that it’s only 16%. Because if you listen in the market, you’re going to think that everyone but your recruiting team is leveraging agentic AI. It’s not the case.
So that’s where my passion is: how can I make sure that our clients, our prospects, and the industry at large is the most informed, so that way they can make the best decisions? Because in the end, that kind of trickles down to that candidate that’s looking for that next role, and that’s where I get my passion. So it’s just at a different altitude.
Gemma Versace:
Oh, amazing. Well, thank you so much for sharing a lot of your passion with us today in this podcast, and thank you so much. We really appreciate it.
Trent Cotton:
Awesome. Thank you.
Gemma Versace:
Hiring for outcomes sounds simple until you try it. Most processes still treat a degree and years of experience as proof of capability, and neither reliably predicts success. AI is making that gap impossible to ignore.
The screen-share exercise shows what can replace them. Hand a candidate some data and ask them to work through it with AI. The answer matters less than the path. Watch whether they accept the first result or keep asking why, or get so curious they forget they’re in an interview. That tells you more than any list of tools on a resume. And if they get seventy-five percent of the way there, the call is simple. Either the missing twenty-five percent is teachable, or it’s a mindset no one has time to coach.
The same thinking applies after the hire. AI training and apprenticeships offer something a bigger paycheck can’t, and that’s longevity. People who taught themselves AI notice which employers will keep investing in them.
The organizations that pull ahead will track AI adoption and results all the way to the P&L. That work starts with leaders who can earn the trust of a hybrid workforce.
Join us next time for more conversations with technology leaders who inspire us to grow, lead, and innovate. You can find us on Apple Podcasts, Spotify, or YouTube. If you enjoyed this episode, please share it with your network.
We’ll see you next time.