The acceleration of AI in talent acquisition has sparked both excitement and concern across the industry. While many fear automation might make recruiting more mechanical and impersonal, what if the opposite were true? Traditional recruiting processes already force candidates and recruiters into transactional relationships – with endless applications, screening calls, and administrative tasks leaving little room for meaningful human connection. At the same time, increasing expectations for personalized experiences are putting pressure on talent teams who don’t have enough hours in the day.
So, can AI actually make hiring more human rather than less?
My guest this week is Diana Tsai, Co-founder and CEO of Upwage, a company building AI interviewing agents that have been proven to reduce turnover by as much as 48%. Diana shares valuable insights from working with employers to deploy over 4,000 AI agents. We also discuss her recently published book “AI For Good”, which maps out a positive vision for the future of recruiting where AI doesn’t replace human connection but creates space for it to flourish.
In the interview, we discuss:
• Why Diana wrote a book about the positive future of AI in recruiting
• How the role of recruiters is evolving with AI empowerment
• Where will the data come from to power advanced AI talent systems?
• How to ensure AI reduces bias rather than amplifying it
• The human-AI balance and what AI will never replace
• Key signals that indicate the transformation of recruiting is accelerating
• What will recruiting look like in 2035?
Follow this podcast on Apple Podcast.
Follow this podcast on Spotify.
00:00
Matt Alder
What if finding the right job wasn’t a frustrating maze of applications and rejections, but a single conversation with an AI agent that truly understands your potential? While many fear automation, what if AI actually transformed recruiting into something more human rather than less? Keep listening to find out how support for this podcast comes from Upwage. Upwage builds custom AI interviewing and assessment agents that act as intelligent extensions of your recruiting team, dramatically improving quality of hire and scaling recruiter productivity. Trained in the STAR methodology and behavioral science, Upwage’s agents conduct structured competency based interviews that surface high retention, high performance talent with human level insight. The impact for their clients has been a 48% reduction in turnover, 2 times recruiter productivity, 50% faster interview to offer ratios and millions saved in preventable churn?
01:11
Matt Alder
Unlike glorified chatbots, Upwage’s agents function as autonomous AI teammates, interviewing thousands of candidates in parallel, probing for real behaviors, flagging top talent and delivering decision ready insights for recruiters and hiring managers. They operate 247 without bias, fatigue or bottlenecks. Candidates love them too, giving the experience an average NPS score of 9.01. With over 4000 AI agents deployed, companies using Upwage aren’t just hiring faster, they’re hiring smarter, more equitably and with unprecedented precision. Hi there, welcome to episode 696 of Recruiting Future with me, Matt Alder. The acceleration of AI in talent acquisition has sparked both excitement and concern across the While many fear automation might make recruiting more mechanical and impersonal, what if the opposite was true? Traditional recruiting processes already force candidates and recruiters into transactional relationships with endless applications, screening calls and administrative tasks, leaving little room for meaningful human connection.
02:49
Matt Alder
At the same time, increasing expectations for personalised experiences are putting pressure on talent acquisition teams who just don’t have an enough hours in the day. So can AI actually make hiring more human rather than less? My guest this week is Diana Tsai, co founder and CEO of Upwage, a company building AI interviewing agents. Diana shares valuable insights from working with employers to deploy AI interviewing agents and we also discuss her recently published book AI for Good, which maps out a positive vision of the future of recruiting where AI doesn’t replace human connection but creates a space for it to flourish.
03:33
Matt Alder
Hi Diana and welcome to the podcast.
03:36
Diana Tsai
So excited to be here. Matt.
03:37
Matt Alder
It’s a pleasure to be talking to you. Could you just introduce yourself and tell everyone what you do?
03:43
Diana Tsai
Co-founder, CEO at Upwage, we build AI agents that decrease turnover, increase performance and help extend recruiting teams. So very fun stuff.
03:53
Matt Alder
Fantastic. Now you’ve just put together a really interesting book about the future of recruiting and AI and all those kind of things, which is just about to come out.
04:03
Matt Alder
You very kindly sent me a copy.
04:04
Matt Alder
Of it, which I’ve read, and it’s a really interesting vision of the future. Tell us a little bit about the book and why you decided to produce it.
04:14
Diana Tsai
Yeah, so my co founder, Greg and Yvonne Johnson, our head of AI and I essentially decided, you know, we wanted to write this book and the book title is AI for Good. How AI Can Transform Hiring for Good. It’s all about agentic AI. The other title we had for it was Agents of Change because that’s really what it’s about. We wrote the book because last year, so last year we deployed about 4,017 different AI agents and had really quite a lot of conversations with TA leaders about what the future of work would look like. And what we noticed is a lot of actually uncertainty and fear about the future.
04:47
Diana Tsai
And, and our thought is always, you know, if we have a very optimistic future, mostly because I think out of necessity you either have to choose to have an optimistic view or you can sort of just descend into the depths of despair around the. Directly just be like, well, that’s it.
05:01
Matt Alder
Right, yeah.
05:03
Diana Tsai
And, and I think my co founder and team and I, we’re not going to sort of roll over and just be like, well, this is the world that we’re entering. All AI robots are taking over all our jobs. Our thought was, is there a way that we can leverage this technology for good? Is there a way that we can take the power of AI and essentially combat the disruptive impact of AI as well in one fell swoop. And so that was the premise of the book. Our, the premise of the book was really, let’s paint a vision and a roadmap to 2035 of what the world can look like with AI leveraged in the best possible way. Now, of course, I can’t guarantee that future because it relies on all of us working together on this front. But we do want to.
05:41
Diana Tsai
I mean, there’s no possibility of building a better future if we don’t even have the way to articulate what that might look like. So that was why we decided to write this book.
05:48
Matt Alder
And I couldn’t agree with that, I couldn’t agree with your thinking more there because I Think it’s just so important to do that because these tools and these technologies have such tremendous pet potential. I really think that they can take us to this sort of recruiting nirvana where, you know, recruiting works, as you say, it combats the threat of the, of AI to jobs and really, you know, everyone can get the job they need at the right time and all those sort of things. And I think, you know, without a vision, but as you say, we’re never gonna get there. So it was really kind of refreshing to, to read that. Now let’s dive into that vision a little bit deeper.
06:29
Diana Tsai
Yeah.
06:29
Matt Alder
In the book you talk about sort of three waves of AI transformation in recruiting.
06:36
Matt Alder
Two questions really.
06:37
Matt Alder
What, what’s it talk about each wave and what it entails? And why did you pick the timescales that you did for this transformation?
06:46
Diana Tsai
Yeah, so the three waves, where the three waves came from. So Greg, call my co founder and I, we always sit down and we do some future thinking at the end of the year. So December last year, we sat down and said, and the question we ask each other is how far into the future can you see and with what conviction level? And based on that, all of our company strategy emerges from this. And so we spent two full days discussing this, trying to figure out like, okay, what do we see? What are we aligned on? What do we, what basis is on all the inputs that we’ve had building these thousands of AI agents, talking to all the TA leaders, deploying them in real hiring environments. And we’re like, okay, we see three waves emerging.
07:21
Diana Tsai
The first wave, where most of us are familiar with now, it’s already happening. We had put 2026, 2027 is essentially when we rise to really see like this wave go mainstream. Well, we call that wave the transformation of recruiting tasks. And so by 2026 we’re already seeing this already. But the world of AI driven interviews, AI top of funnel is really going to proliferate. And the reason for that, it’s simple, it’s just economics. Behind the scenes we know the ROI on these tools and it’s too good to not adopt just in any system that is capitalistic, like it’s impossible not to do it. But additionally, the transformation of recruiting tasks means the transformation of the role of a recruiter that is these meatiest conversations we’re having.
08:03
Diana Tsai
I think it’s critically important for us to have a vision for what the role of the recruiter is in this new world. And it’s Actually a really freaking an exciting role. And so we have the transformation of the role of the recruiter. We also have the emergence of different types of businesses. So the companies that adopt AI, right, we see this with companies that we’re working with that are using our AI agencies like faster hiring, better hires, lower turnover, so massive long term business growth. There’s going to be a distinct difference between those businesses that do adopt AI into their processes, including recruiting, versus those that don’t. It’s just the pace of growth. You can’t compete against a company. Like when a company adopts AI agents across its different business units.
08:42
Diana Tsai
That’s like you literally just taking your workforce and making it and turn it into 10x your current workforce. And now you’re competing against companies that are still at 1x workforce, which is like okay, you can see where the math works out. So that’s really wave one, right? Wave two emerged. Our vision for wave two emerged out of real customer conversations. And actually we have an amazing advisory board of talent leaders. And a lot what we talked about there is. So when you deploy hundreds of AI agents across your organization for let’s start with basic tasks, interviewing assessments, right. Onboarding is another more basic. So as these AI agents are deployed in partnership with TA leaders and TA teams, you rapidly will see the emergence and the need for a new kind of AI manager that can manage these teams of agents.
09:25
Diana Tsai
So a coordinator and the best example I can give of this is one of our partners has 962 AI agents deployed across their different roles. And so the question now becomes how do those agents pass information to each other? So if a candidate’s going through one interview for a job and actually they’re really qualified for three other interviews that our agents are interviewing for, how can we effectively know in the middle of that conversation, redirect them, reroute them. And then the other part is we’re already starting to build agents downstream. So not just in the top of funnel, but retention check AIs that actually engage with your employees after.
10:00
Diana Tsai
So what then happens is how do you feed that data back into the interviewing agents at top funnel to make sure that those learnings of who is actually a great fit, great performer in that job end up trickling up to the top. So to be able to do that you need an AI manager. And so that’s sort of your multi agent system type construct where you have this AI manager that’s reporting to TA leader and HR leaders to unlock deeper insights and strategic decision Making. So we think that’s 2028, 2029. It could be earlier with how fast things are moving. We’re already doing a lot of R and D into how to build that AI manager, but the coordination piece is still rather complex. Like the coordination of agents.
10:36
Matt Alder
Oh, absolutely, yeah.
10:37
Diana Tsai
In the final wave three, we put 20, 30 plus. But really we want it to happen tomorrow. This is the whole reason we even wanted to build this company is the transformation of the labor market. How AI transforms the labor market. So today, you know, if you think about today, right, we’re powering thousands of AI agents. They, they’re powering interviewing assessments for a variety of jobs across hourly, all the way up to executive level roles. But on top of these job interviewing agents, you can actually build a universal interviewer that can replace the traditional application. And why that’s so exciting is today, you know, the kinds of wins that we’re delivering to TA teams are all around time to fill, time to hire. And then you’ve got quality of hire, you’ve turnover metrics, your performance metrics.
11:19
Diana Tsai
But that translates into a far superior candidate experience. Because what it means when we decrease time to fill is a candidate gets a job faster. That’s like, that’s awesome. So on the other side of this universal interviewer concept, you can almost imagine a GPT like interface where a candidate can just go to a single interface and just start talking about like, hey, I just got laid off. You know, X company just laid off 10,000 people. All these folks can just come to the universal AI interviewer, say, this is my background. I’m going to drop in my LinkedIn profile, here’s my resume. This is actually what I’m passionate about. Because to be honest, I didn’t really love my job anyway. I’d rather do this. But I do have these skills. Can you tell me what’s out there?
11:55
Diana Tsai
And the AI interviewer can connect to millions of interviewing agents that are live in the market. So that’s our goal, is to build as many job agents as possible so that we can essentially layer them foundationally underneath the universal interviewer. And the universal interviewer can help job seekers get to the right job much faster and consolidate, you know, 45, 50 interviews for a single type of role into one interview, find opportunities for better roles in the middle of the conversation, all that great stuff. And the intent there is to create a world where layoffs no longer derail careers. Right Today there’s just, especially in the United States, there’s no system of protection There is no safety net for layoffs. There’s really nothing there. It’s just good luck, go apply to a bunch of jobs.
12:37
Diana Tsai
So, and this is going to become even bigger of a problem with AI and automation. We just are already seeing the job market evolving. So the question is, how do you leverage the power of AI, build this safety net of AI agents to essentially mitigate what is coming?
12:52
Matt Alder
But yeah, and it’s a fantastic vision because that whole, even just that whole job discovery piece, we’ve not moved on since you looked for jobs in newspapers, you know, job boards and all of that kind of stuff. And the search engines, that search engines that we’ve had, they still make it very difficult for people to find jobs that are suitable for them, to find jobs that are still open. And I think that the potential of AI to fix that is just kind of. It’s just sort of fantastic. And it’s interesting about timescales as well, because from an AI perspective, you know, 10 years seems like an absolute age.
13:31
Matt Alder
But having said that, I’ve done quite a lot of work examining how people have talked about the future in the last 10 years of this podcast, and change takes a very long time in this industry. So it’s that kind of balance between the two is obviously quite difficult to, quite difficult to predict. Whichever way that we look at it, though, the role of recruiters is, is evolving. What kind of skills and competencies do you think that TA leaders should be focusing on in their teams right now? So, you know, the transformation has started. It’s probably debate about which stage we’re kind of at and all that sort of stuff and how long it’s going to take. But, you know, fundamentally, the role of recruiter is already changing.
14:13
Matt Alder
So what should people kind of focus on in terms of making sure they have the right skills for the near future?
14:20
Diana Tsai
I love that question. And so I think rather than answering from my point of view, I’ll just share what I’ve learned from some of the chief people officers, the TA leaders we’re working with, and how they’ve retrained. They’re retraining their teams, they’re empowering their teams, because I think that’s just going to be more valuable, their perspective. For sure. We get to learn a lot from amazing TA leaders. How they’re applying technology, I would say, typically starts with the philosophy. So it starts with a point of view on how we’re raising our game collectively. So what I’VE seen from exceptional TA leaders is that they’re recognizing that there’s a fierce new reality that we’re competing in, which is just that you have accelerating tech stacks, agent stacks that are being built across all different. And this isn’t even unique to TA leaders.
15:04
Diana Tsai
It’s also every single business unit. So leaders as a whole of business units are realizing we’re the new market we’re entering. All of our competitors are in a like AI adoption war. And so essentially, how do we actually equip our teams with the right technology to outperform or to continue to perform at the same levels that we are today in the recruiting space specifically, what this means is also candidate expectations are increasing significantly. So high touch, rapid engagement, personalized interviews, no more chatbots, no more application forms is a baseline. It’s no, it’s no longer a nice to have. It’s like, oh, whoa. Our expectations, the expectations that we are held to as TA have significantly increased since AI. The other part of it is the opportunity that is opening up for recruiting teams, right?
15:51
Diana Tsai
So TA leaders realize that, okay, my recruiter team is now price is truly priceless. Like, the time is really important. So how do I free my recruiters from all of these manual screening tasks, the endless phone tags, scheduling as admin, like, how do I get them to what they’re actually meant to be doing? And so this is the foundations. That’s the beginning of the change is when a leader starts thinking, okay, how do we raise our game? So the second part of this is AI empowered recruiters. So I think it’s like distinctly understanding that there’s a difference between recruiters and AI empowered recruiters. And this is at the point where a TAT leader will say, hey, I want to empower my recruiters with the best technology possible so that they can function as 5x10x versions of themselves.
16:31
Diana Tsai
So as the company grows, we can extend our reach, our ability. I think the number that even we’re talking about internally outside of the recruiting space is revenue per headcount. Like, it really comes down to how do we increase revenue per headcount? Because when technology accelerates, it’s like, okay, everyone becomes more productive. And so in this world, what we’re seeing is TA leaders that are recognizing that the old way of doing things for recruiters is, you know, scheduling time suck, backlog of candidates, no call, no shows, lots of drop offs, and there’s almost zero time for relationship building with top candidates. Highly transactional conversations, 20 phone screens, 15 minutes. Let’s go, let’s go. Right? There’s no, like, I really want to dig into this part of the conversation, really want to get to know you.
17:09
Diana Tsai
And a lot of that manifests in the bottom of the funnel in terms of forfeit candidates or low performance or not a culture fit or turnover. You know, all the time we’re hearing from candidates like, this was not the job that I thought I was signing up for. I don’t want to be here, like. And so. And those are easy to mitigate. It’s. If you have enough time to spend with a candidate, you can. This is mitigable. So that’s one side in terms of how we’re seeing TA leaders evolve. The relation, the role of a recruiter, to say, hey, okay, when we have AI empowered recruiters, we’re gonna have our recruiters focus a lot more on deepening their relationships with candidates, especially top candidates, you know, because you’re not just vetting that candidate’s vetting you. And so what is the process for how.
17:47
Diana Tsai
How do you invest in that time and make it less transactional? Super important. The other side we’re seeing TA leaders lean into is actually hap, like training their recruiters to spend more time with hiring managers. So rather than saying, we only have time for an intake form, now we got to go back to our phone screens. It’s like, let’s sit in on the business. Let’s get to know those teams deeper. Let’s build much deeper relationships with the hiring managers. More, More deeply. It’s just. This is what it opens up, right? This is what AI opens up, is the opportunity for recruiters to lean into the. This small sliver of time that was previously allocated for real relationship building and expand that dramatically, which has massive ramifications for the business from a quality perspective.
18:26
Diana Tsai
So in this new world, then, the AI empowered recruiter has 247 AI interviewers that are automatically taking care of their interviews overnight. They’re screening, and they can jump into those screening and interviewing results in the morning when they come to work, dig into those conversations, find the great candidates, still get them on a phone call. But actually, the phone call sounds totally different now, like, totally different conversation happening. More bandwidth for those deep conversations, more bandwidth to bring those candidates back to hiring managers, go back and forth, really, to and understand caliber. It’s like you get to be more human. And so I think of it as, like, shifting from defense to offense mode. That’s what we’ve noticed the most in the teams that we’re working with is they’re like not running around hair on fire anymore.
19:04
Diana Tsai
It’s more like, okay, we can be on offense, we now have the tools to be on offense. And I think the best leaders I’ve seen in this space are really thinking about this technology as investing in their recruiting team. So each recruiter gets a team of AI agents to support them on first round interviewing and assessment for any role. This is like super critical and to also really train recruiters to understand like you are the decision maker, you’re still doing what you’re doing just at a higher quality and a higher bar. What’s really critical is you now have a team supporting you. That’s, that’s what it comes down to. So I think, and I think that’s universal, like even outside of recruiting it’s each individual employee team member now has an AI agent team that can support them.
19:41
Diana Tsai
And that’s amazing from a productivity perspective and also just from a, where we can focus our human talents on perspective.
19:51
Matt Alder
Oh absolutely.
19:52
Matt Alder
I’ll come back to that in a second with some kind of sort of follow ups around that sort of human machine split. Before I do though, just going back to the kind of the overall sort of vision. You know you mentioned AI talent managers kind of helping people with their career and all this kind of stuff. Now difficult question to answer because we’re talking about the future and there were unknowns and all that sort of stuff. But one of the issues with recruiting at the moment is around almost a lack of data. So you know, resumes are not, don’t have very much on them. LinkedIn profiles really don’t either. You’ve got job descriptions, you’ve got disparate HR systems and all that sort stuff in order to sort of support these kind of super AI agents of the future. Where’s the data coming from?
20:41
Matt Alder
What kind of needs to shift to sort of build the foundation for that to happen?
20:46
Diana Tsai
I love this question. So this is funny that you asked this because we haven’t even talked about this, but we are. So we’re building out an AI builder agent that releases at the end of the month. And what it does is it collects all of the context to build AI interviewers. And so it’s based off of all of our learnings from building the last 4017 AI agents. But essentially what we realized is it all comes down to resumes. LinkedIn profiles, incomplete data produces like bad data in bad data out. So it’s like, it’s not surprising then. So, so right now we’re matching at the top of the funnel with relatively bad data, a resume and then relatively more bad data, which is like just an application. So it’s like just together. This is a, this is sort of like the blind leading the blind.
21:27
Diana Tsai
So the world that we’re entering now with our AI builder agent, what we’re actually doing is enabling recruiters and TA leaders to be able to build their own AI agents by inputting a couple key data sets. Base minimum, job description, a well written job description. We take that, inhale it, convert it over into competencies, skills, the definition of those competencies and skills for a business. And then the questions, the interview questions, the behavioral interview questions and assessment questions that match up to those competencies skills. So the first data point like the base minimum is job description. This is literally if you have like nothing else. But what we find is if you want to really create an incredibly robust AI interviewer, we also include typically company mission, company values, sometimes because you want culture fit questions that are actually screening for company values.
22:11
Diana Tsai
Previous interview guides, we can learn from that tone of voice. Examples are super important as well. And then also what I love is I think about one of the most ingenious applications, like I think about one of the best AI interviewers that one of our TA partners created it decre turnover in a patient service representative role by 48%. And it was because he took performance reviews at the bottom of. So he actually literally took these performance reviews and said, I want my AI interviewer to understand what we’re looking for in an employee, a high performing employee. And I was like, this is freaking brilliant. Like perfect, right? So that’s how you leverage this base data set of inputs that are far more robust than anything else to then create your AI interviewers.
22:53
Diana Tsai
And so then the AI interviewers can understand, okay, I have job description, I’ve got the interview guys, I have performance. Okay, I have the hiring intake form. I’m going to aggregate all this and now recommend a set of competencies to the recruiter or to the TA leader. And for this particular role, and it’s highly customized, so that’s how we generate then competencies, usually three to five or skills related to that particular role. And then underneath that you nestle the questions so then you can generate interview questions that approach those particular skills and competencies. And when you do it this way, you’re essentially engineering the interviewing process from the bottom up. You’re really going deep in funnel.
23:27
Diana Tsai
And so what’s super interesting about that is everything that I just described with our AI builder agent you can do in less than five minutes and the moment all those inputs go in, the agent’s created. So it’s instantaneous agent creation. So we’re super stoked about it because it’s just going to enable a lot of really quick ability to capture a lot of the things that have been almost impossible to capture before.
23:47
Matt Alder
In the book you also talk about AI’s potential to reduce bias in hiring and it’s something that I fundamentally kind of agree with. I think that there’s a huge opportunity for AI to do that and sort of fix recruiting from that perspective. On the flip side, there’s also the opportunity or the fact that it might do completely the opposite if not handled properly.
24:10
Matt Alder
Or.
24:13
Matt Alder
There are scenarios where it increases bias in hiring potentially. What can employers do right now as they’re sort of learning about AI deploying AI purchasing, AI building, AI, whatever it is to make sure that you know, we’re going down that path of reducing bias rather than amplifying existing bias.
24:34
Diana Tsai
I love that question. So first I think start with the regulations. So New York State Law 144. We also have a list of all our regulations on our trust center as just a resource for TA leaders. But I think it starts with the mentality of let’s not meet, let’s not just meet current AI compliance regulations on a federal or state level or even on an international level. Let’s challenge ourselves to exceed those standards. What we found when digging into most of these laws is they’re rudiment. I mean it’s, they’re basic. I would actually love more regulation. I think there’s, we find we’re all the time like self policing ourselves above and beyond current regulations because we’re looking through them. We’re like, well this is interesting. We’re required to do a third party bias audit but it’s only on protected classes.
25:16
Diana Tsai
What about non protected classes? What if our AI’s bias. So we literally, that’s why we run another secondary bias audit that’s not required by law to focus on non protected classes and adverse impact because it’s like there are things that the law doesn’t require that I think we know we should be doing and especially TA leaders will know. Like that’s interesting. Like complying to be honest, like compliance. If you just go through the, go through it all, it’s pretty onerous but it’s not impossible. The harder bar to set is how you imagine how to exceed it. So essentially, let’s go beyond. I can give another example. So I think like regulation is one side of it. How do you exceed current regulations? You can set that as like baseline benchmark and go beyond.
25:54
Diana Tsai
The second is actually in the product design process, which translates from a TA leader perspective into really being attuned to what’s going on inside the products that you’re considering using. So specifically, like, here’s a good example. So our AI agents, we have an interviewing agent and analyst agent that does all the assessments. And so they’re actually completely separate agents. We did this intentionally because we actually end up redacting all of the PII from the interview before it goes through the assessment agent. So it’s stuff like that where you want to like, just dig deeper into the product. So if you see a product, you’re like, okay, it’s interesting. They’re doing interviewing and they’re anal. Are they, are they doing any redaction between like is all. Is how do we make sure that, that we sort of de bias in advance that analyst role? Right.
26:37
Diana Tsai
So that’s another side. The other thing I would look at is also really thinking about like, so one of the metrics that we track is the percentage of candidates that are completing their interviews after hours. So non traditional candidates that may actually just simply opp. So this is like little. They’re actually giant pots of bias, to be totally honest. But they’re just overlooked because it’s like, oh, well, we offer interviews when our recruiters are online 9 to 5. That that just means that everybody who can’t do 9 to 5 and you can sort of think through, okay, what are the scenarios of folks who are less likely to be able to nine to five, maybe they have another job, they have children, they have other gigs that they’re doing. Whatever it happens to be, it’s like those folks are now excluded from the process.
27:18
Diana Tsai
Do we include that right now as a set of like, oh, that’s by. No. But do we see that has adverse impact? Probably, yes. And then the other thing I’ve looked into is form matters. So the reason why we ended up like really leaning into chat is actually we did a lot of studies on video and essentially like video is very efficient in a lot of ways because you get to see personality, all these elements. And we really also are looking into how we bring video into the process. But what we’re very cognizant of also is Video oftentimes encourages certain candidates to participate and others to not.
27:48
Diana Tsai
And in several, not just studies, but also in conversations with TA leaders that are trying to either implement or remove video based on technologies, there’s a lot of like, oh my gosh, we noticed that women were opting out at a significant high, like significantly higher rate. They just opt themselves out. How do our diversity pipeline is decreasing. So it’s stuff like this where there’s a lot of nuances in form and how the product is designed that will inherently bake in or remove bias from your process. And the dial moved is much larger than simply like your bias audits or. Because that’s after the fact. You, you have the results tracking after the fact and you improve it in BAS implements like, oh, went from point A to point nine, whatever that is.
28:30
Diana Tsai
But really it’s a design question, like at the fundamental level, how did you design the system? Is it. Is there thought put into how the form includes more candidates, how the form is more inclusive of different. So there’s that. I think that’s the key thing here.
28:44
Matt Alder
But yeah, you kind of gave your sort of vision for the future role of recruiters and all that kind of thing. The kind of. The debate about the balance between AI and humans in recruitment is obviously, it’s obviously a very emotional discussion at the moment because, you know, radically affects people’s jobs and all that sort of stuff. And there are some very kind of polar opinions on it and what’s going to happen? Where do you think that balance is going to, is going to end up? So, you know, kind of realistically, what are the things that AI is never going to be able to do because. Or certainly that we can’t foresee that. We can’t foresee that at the moment because I think a lot of the debate gets into the nitty gritty about what’s possible now and people’s experience, that kind of stuff.
29:31
Matt Alder
But the looking at a different way, you know, fundamentally what are the things that AI, you don’t think AI will be able to do that humans do really well?
29:42
Diana Tsai
It’s the relationships. It literally is. It comes down to the fact that, I mean, I see the biggest gap honestly in just, let’s just call the human capital space, the people management space, the talent space. So when AI can do some of these initial automatable tasks, right, what it opens up is the opportunity to lean into areas that we don’t even have time touch right now and have Never had. So things like how do you provide white glove service to all of your inner internal applicants who have worked at your company for five years, have no idea a job exists that is open. In the meanwhile, you’re spending a ton on job ads trying to source a whole bunch of books externally.
30:21
Diana Tsai
And then if someone internal does find it, they’re thrown into a pile where they apply in the same process as an external candidate. They’re like, wait, I’ve been here five years, I already know this company. And there. So, so what does that look like? If you can literally have recruiters spending time saying, okay, my external pipeline is taken care of. Mostly I’m going to jump in and look at the highest fit candidates. They’ve been sorted for me, I’m going to go through those. But on the other hand, what’s opened up for me is I can go into my internal talent pools and I can actually not only, you know, use algorithms to find the best can but actually reach out, grab a coffee, be like, hey, I know you’ve been here for five years. Did you know that these opportunities exist?
30:53
Diana Tsai
Or I saw you applied for this job. I want to dig in with you because is you’re actually not a, like, let’s talk about skills fit. You’re actually a better fit for this other job within that I have open over here. Would you be open to relocation? How’s your family doing? Are you interested in this? How none of this. This is like crazy to me now. This is the stuff that’s super exciting. Like, I get chills thinking about like, okay, what happens when you actually have time to spend with people inside the business? And then I think on the other side, it’s really around. The best metaphor I have for this right now is in the. Okay, so in the AI engineering world right now, right, there’s this tool that’s sort of exploding and it’s called Lovable.
31:28
Diana Tsai
And so with Lovable, you can create your own websites with a prompt. And so people are freaking out about it. So I was playing with it the other day and what’s so interesting about this tool is that you can get to your 80% MVP very fast. And so at the 80%, you’re like, oh my gosh, this is so crazy. I’m using this tool and I’m not an engineer, I’m building this amazing product. But you hit 80% and you’re like, this is super annoying. Now I can’t get it the last 20% to be. This is actually, I believe, happening across all industries. And I bring it up because the moment this happened, we obviously work with a lot of our engineering team. The engineer’s role, because of AI tools and AI engineering tools like level has now shifted to the last 20%.
32:09
Diana Tsai
And it’s a very important role. But what that means is if you can open up the first 80% instead of having to spend 100% of your time to build one MVP or one prototype, you could build five prototypes for five different companies simultaneously in that time. Right. But only focusing on the last 20%. And also your clients could be coming to you saying, like, here’s my 80% prototype. Could you help me get it to the last mile? I don’t have to spend time going back and forth trying to figure out all the minutiae in the beginning. Super fascinating. Like, and that’s what I’m seeing also. It’s, it’s what we’re seeing in the recruiting world too. It’s like that last 20%.
32:40
Diana Tsai
You think it’s 20%, but let’s think about it like that last 20%, you can expand it so much into areas of detail we’ve never had time to go into. And that’s what I mean. It’s like the whole role of a recruiter can shift because now if you take off the first 80%, there’s this remaining 20% that you can expand into and reimagine what that looks like. Like, what does it look like if instead of a hiring manager intake form and a 30 minute process, I can sit down with this team? And it’s not. And I can actually understand not just the role, but the nuances of this specific team and what this team needs and the dynamics in the room. Because it’s not always the same person. That’s the thing. Like, you’re not always.
33:18
Diana Tsai
Even if it’s a single role, the same role I’m hiring for a designer, it’s not the same designer hiring for the same team. It’s team specific to and culture specific. It’s not even company specific. Sometimes it’s team specific and we do not have time for those conversations right now. Like, I rarely come across recruiting teams that are in that deep in the business to be able to have that level of nuance.
33:37
Matt Alder
But yeah, final question for you. So obviously there’s a huge amount going on. Different companies are at different stages with all of this. You know, it’s, it’s very, the landscape can look quite confusing when you sort of look at it. What are the. Going back to your sort of 10 year vision. What are the sort of significant signals or sort of waypoints that we’ve reached a certain stage or the next stage is coming as we go through. So it’s really kind of like how do people know that transformation is happening and things are moving forward.
34:12
Diana Tsai
So, so interesting. I think the one is the one most obvious metric or the most obvious signal is the adoption of AI agents. So you can almost do a count, right? It’s like how many AI agents are, is our team using how many AI agents are proliferating across our enterprise. That’s step one is like the initial. So I think of those as Bs. AI agents can be as simple or sophisticated as you want them to be. Essentially it’s like they can be almost a basic level. They’re doing a simple job to a very complex advanced behavioral interview like whatever it happens to be. So but they’re like bees in a beehive. And so once you have enough bees then you can essentially build a queen bee. And that’s what the talent manager is.
34:52
Diana Tsai
And so you actually need foundationally like enough AI agents to have network effects to then build a management level agent. So that’s where you’ll see the pro like the emergence of the queen bee talent manager type. AIs will essentially the prerequisite is enough Bs, enough agents that have been adopted. It’s going to be the same thing with unlocking the universal interviewer because otherwise the only data the universal interviewer has to build on is job applications. And we’re back at square zero where the interviewers referring people to apply for jobs and nobody gets responses. So, so you actually foundationally have to have the interviewers in place. And that’s why like everything is really about this sort of race right now to building the interviewers.
35:34
Diana Tsai
And the, the interesting thing is that when you actually have the bees, let’s call these like the interviewing agent bees in place. It’s not only you can not only build a universal interviewer for a job market, you can also do it for individual companies. So if a company has like 5,000 open recs, the same applies. Imagine a world where you’re not going to a career site trying to like figure out what the heck you’re applying to and it’s just like you’re talking to a universal interviewer that’s designed for X company and you can actually get to know all the roles. And it’s not a chatbot that like gets you to a dead end. So this is the key thing, I think that is really interesting. And so if you think about network effects, that’s how you can sort of benchmark what wave is coming.
36:10
Diana Tsai
And I think it could happen a lot faster than we think as well. A lot faster. But yeah.
36:16
Matt Alder
Diana, thank you very much for talking to me.
36:19
Diana Tsai
Thank you. This is great.
36:21
Matt Alder
My thanks to Dianna. You can follow this podcast on Apple Podcasts on Spotify or wherever you get your podcasts. You can search all the past episodes at recruitingfuture.com on that site. You can also subscribe to our weekly newsletter, Recruiting Future Feast, and get the inside track on everything that’s coming up on the show. Thanks very much for listening. I’ll be back next time, and I hope you’ll join me.
