Tower acquisition is going through an unprecedented transformation and many of you are likely in the middle of planning your strategies as we move through 2024 towards 2025. We all know that operating models, change management and aligning TA with corporate objectives are essential parts of any transformation strategy but with the market and AI technologies in particular evolving so rapidly, there's a real risk that your strategy could quickly become outdated. That's where strategic foresight comes in. It's a proven methodology that helps you build credible future scenarios, create agile strategies and most importantly have a proactive influence on what the future of TAL acquisition looks like both within your organisation and across the industry. I know you're busy, so I've created a concise online course that breaks down strategic foresight into easy to learn tools specifically designed for TA transformation. It's quick to implement and will keep you ahead of the curve. You can learn more by visiting mattalder.me/course that's mattalder.me/course. There's really never been a better time to shape the future of TAL acquisition, so don't miss this opportunity to make a lasting impact. There's been more of scientific discovery, more of technical advancement and material progress in your lifetime of the mind at all the ages of... Hi there, welcome to episode 653, a recruiting feature with me, Matt Alder. I know that many of you would agree that we're now beyond saturation point with the hype around AI and TA. It's clear that one way or another, JAI is going to become ubiquitous in the software that we use. So rather than talking about it as a topic on its own, we need to look carefully at the current and feature impact on TA in terms of use cases, innovation and the potential to rethink hiring. So what are the proven use cases for AI right now? Where are we missing opportunities and how are things going to develop in the short term? My guest this week is Mark Jaffe, co-founder and CEO at Hackajob. In our conversation we discuss the current use cases for AI and the lessons that TA can learn from the impact that JAI has had on software engineering. Hi, Mark, and welcome back to the podcast. Matt, thank you so much for having me on, always one of my favourite conversations, so looking forward to today. Well, always a pleasure to have you on the show. Just for anyone who may have missed your last appearance, could you introduce yourself and tell everyone what you do? Absolutely, my name's Mark, I'm the co-founder and CEO at Hackajob. Unbelievably, Hackajob has been going for 10 years this month. It is a crazy, crazy feeling. So yeah, I co-founded Hackajob with my co-founder, Razzbox, who is still at university and King's College London. 10 years later, here we are. And very briefly, Hackajob is a technical hiring platform that helps predominantly large enterprise organisations, sourcing gauge, hire technical talents. We've recently launched Hackajob in Pedagence, which is very exciting, like AI driven platform. And I'm sat in New York today, which has been a lot of fun as we've expanded out here over the last 18 to 24 months, particularly operating in the US, the UK and India. Awesome. This podcast is nearly 10 years old, actually, so probably started round about the same time. It doesn't feel like 10 years. It's been a very quick 10 years for time flies. So tell us about the current state of the market as you see it, because tech recruiting was always the kind of the bellwether of the good times. Obviously been suffering quite a lot in the last few years. What's happening out there? What's going on in the market at the moment? Yeah, I think it's even more fascinating for us and it has been because we have a much more global perspective and there's definitely a few macro themes that we're seeing. So offshoring and lower cost locations is back in a really, really big way. I was at a leadership summit this week with our investors, and nearly everyone I spoke to about tech hiring were exploring how do we find talent offshore in lower cost locations. And my hypothesis on this is twofold. Obviously, as interest rates go up, it's amazing how much you talk about interest rates in being in a world of recruitment. I never thought I would. But as interest rates go up, investors in the public markets want returns in the in the nearer term versus the longer term. So I think organizations are looking at how to become more profitable and therefore a lot of the cost based inside the organizations is people and therefore how does that impact talent strategy. But I also do think that there's potentially a bit of a backlash to remote work and hybrid work, and if you're the CFO looking at it's very, very expensive senior data scientist in California, we're saying actually could we do that job in Mexico and having that conversation. So undoubtedly, that's a big trend and a big theme that we're seeing. And I don't think it means the demise of major tech hubs at all. But I think as people think about their talent blend and their talent mix, they're certainly looking at lower cost and lower cost locations. I think that's a big, big trend. Obviously, we're going to talk a lot about AI today. What we're really interesting is probably the best or second best use case for AI since it's become popular with the generative AI waiver over the last two years. There's actually is a co-pilot to software engineering. And so as we think about the hiring for technical talent, it's going to be interesting to see how the companies approach their talent strategy when their existing engineers can be a lot more effective with these co-pilot tools. So Amazon recently released their course of the earnings. I think it was the prior quarter. And they talked around 50,000 Java applications that had been rewritten using a co-pilot tool, Amazon's co-pilot tool. And it saved hundreds of millions of developer hours in the process. And so how are we seeing that translate? Well, in the UK, there are a lot of these AI trained deploying models in the early career space. And I have a lot of fear right now for early careers in junior talent. There is a lot less hiring going on in the tech world in that space. And whether that is because of senior engineers leveraging these co-pilot tools and more other trends, that is definitely another kind of macro theme that we're seeing. So there's a couple of things that are happening in the market as we see them. Yeah, interesting times. And that sort of early careers thing, it appears to be the sort of the thing that people haven't really worked out about AI. How does someone start a career? How do they get to the level where they need that co-pilot? And I think we've seen from other industries in the past, where they've stopped focusing on early talent, they have these kind of huge issues, sort of five, 10 years down the track. So I'm presuming people think they can AI their way out of that. But it is an interesting, it's an interesting situation, definitely. As I was talking about AI, let's talk about AI in recruiting. So huge amounts of talk about what that means for the future, what it means for the TA team of the future, all of those kind of things. What about right now? What can AI effectively do in the recruiting process right now? Yeah, it's such a brilliant question. And there is so much noise out there. I guess, if I just take a quick step back and I will jump into what now, I do think that the fundamental technology shift is real. So I don't think this is blockchain and web-free, where bluntly, we looked at it and we could just never see an interesting use case for recruitment that would ever become mainstream. I do think that this time is different and this technology is going to become mainstream. There is so much height in the vendor space right now, we can get into that, around what is actually capable and what is it doing. I think if you think about the recruiter job, there is a number of very either manual or kind of data there tasks that are recruited us. And then there was a number of ratio in the tasks that a recruiter does. I think in the manual tasks, there was no doubt you could do a bunch of it in. It was hilarious, as I was at a round table in New York recently, with some great TA leaders and they were talking about the most, the thing they're most excited about for AI is interview scheduling and automating interview scheduling. And I think that's where there is a, the language used, maybe AI just gets chucked around, but that's some pretty basic automation that's been possible for quite a long time. And an interesting example where somebody might say, "Oh, now I'm managing AI to automate interviews." And it's like, well, you could kind of always do that. If we really get into this latest way of AI and what's called gen AI, it's these large language models, it's the main innovation. And so if we dig into this and like, where can large language models be applied? So large language models, well, are they? Well, basically a computer can now interact, engage, infer natural language. What was a programming language or what is a computer? A computer is called binary, right? It's zeros and ones. And programming languages were just layers of abstractions for ways for humans to communicate with a computer. And actually different programming languages have different layers of abstraction. Well, basically, with chat, GPT, we've reached the ultimate layer of abstraction. You can now speak to a computer in a natural language and fascinating, that was actually the Turing test, right? Adding Turing, created Turing test, which was good a computer to tend to be a human and deceive a human and thinking that it was a human. And we've basically got that, which is an amazing innovation. So let's then apply that to recruitment. Well, a computer could now read a candidate CV as well as a human could read a candidate CV and add a lot more context and inference around what actually is the candidate talking about. So if you think about online recruitment, it's always been incredibly keyword based. You know, I'm looking for x, y, and z and I'm just going to pull up those keywords in a candidate CV. Candidates get really good at gaming these systems and put all of the keywords in white text on their CV. So you can't even see the top today. Whereas now, actually, you don't need to use a keyword, a large-language model could read a candidate CV and actually explain back to the user, this is why they're a really good fit. So I think that's the first application that we're seeing work really, really well, is adding a lot more context into the matching. The second one I would give is how you actually gather requirements in the first place. And we've all got very used to writing impromptu a chat GPT or name your large-language model of choice and interacting in this very conversational way. And I think that, again, going back to events of recruitment, we've all had those amazing experiences where you work with a recruitment agent, right, that third party recruiter. And actually, I often think it's the dialogue. It's the back and forth that makes that such a brilliant experience. Well, now when you're searching and discovering talent, you can have that back and forth with a computer and have that same level of dialogue. So I think we should be really clear around what is actually AI. And so interview scheduling probably isn't, but amazing people are automating it. And then actually where we're using large-language models and ways we haven't before. I think that makes perfect sense. And it's interesting from that kind of contextual match situation, because I'm really interested as we move forward to see how much I think experience recruiters have always said, I can look at a CV, I can look at job descriptions sometimes, I just know, or I'm using my experience to know that this is a good match. And I'm interested to see how that perhaps gets deconstructed over the next few years in terms of what computers can do, what the humans are really doing. Yeah, it's kind of an interesting, it's definitely an interesting one. And I suppose that kind of does lead on to the next question. Other than having the interview scheduled for them much easily, much more easily, what does this mean for TATs? What do you think that the implications are? What should they be thinking about? Yeah, so there's two camps at the moment. There are people that are thinking around AI recruiters and basically we're moving the recruiter from the process. People think about this as just another tool, then a recruiter's toolkit. And my view would be much more on a latter camp. I don't think AI is going to replace recruiters, but I think a recruiter that knows how to leverage AI will replace a recruiter that doesn't. And a recruiter that knows how to leverage AI will be far more effective at their job. What will their job entail? Well, let's just use the example you talked about recruiters and say, I can read a job spec, read a candidate CV and just know. Well, artificial intelligence could read a job spec. Actually, there doesn't even need to be a job spec. It could just be a description of what you're looking for. Could then go and review half a million CVs and show 10 candidates are ready to interview in less than 10 seconds. So probably the human isn't maybe going to be doing that piece there. What is this human going to be? I'd always said the best internal recruiters are the best stakeholder managers. Actually, you'll notice the best TA leaders, their business partners in the tech division. In any division, we'll see them as a true expert on recruitment and go to them to be that, okay, how do we bring the best talent into this organization? And so, if you're an IC level recruiter, partner with a hiring manager, being able to shape their requirements using your industry knowledge, using your market insight, I think is absolutely critical. And then doing that on the back end with the candidate and delivering that exceptional candidate experience. And my hope, Matt, maybe on slightly naive, is that we will leverage these tools to actually get back to doing the thing that we want to do as humans. And delivering that incredible human-led candidate experience would be my goal and hope and future TA. Yeah, no, absolutely. I couldn't agree with that anymore. And I suppose it's interesting because going back to earlier in our conversation, we've all got a blueprint for this when you were talking about all the developer hours that Amazon has saved by using AI to do things. And it's some recruiting HRTA is always like a little bit behind the curve for the rest of the organization. But I think it's actually quite clear where we're going from what's happening, from what you're seeing in that market. Yeah, totally, totally agree. And I actually hope that we'll come onto the vendor stuff at some point. But I actually hope that recruitment in TA might not be as far behind the curve as it was previously. It's funny when we think about stuff like recruitment marketing and performance marketing and always felt like it was like five or 10 years behind just consumer marketing and consumer performance marketing. And it will be interesting to see how quickly does the TA industry adopt some of this technology and there are some brilliant solutions out there already versus how risk-averse or slower organizations and TA functions to adopt this tech. No, absolutely. And so my next question is like, what's it feel like to be a vendor developing products at the moment? I was at HR Tech last week in Vegas, thousands of people, hundreds of people exhibiting. I did about 16 briefings with different technology providers. And every single presentation, they had a slide that was called the AI bit or the equivalent of the AI bit. This is what we're doing. What's it like out there? What's the pressure like to innovate? How does it feel that your competition has potential access to the same kind of technologies? What's it like? Yeah, there's a few different things I'd say. Firstly, this is the most fun I've ever had in the 10 years of running Hackajaw. And a lot of that is because of the underlying technology. And like I said earlier, I do think this technology is real. I do think it's going to change the way we do things in the TA industry. And far beyond the TA industry. And how I know that is, gone 15 months ago now, we ran an internal hackathon playing around with some of these language models and decided to actually rebuild our core matching engine from scratch. And the amazing thing about that journey is what we now have live, what our customers are using today, was not feasible six to nine months ago, because it was too slow, the latency was too high, and it was too expensive. And that's what's so fun is that these models are developing and getting better so quickly. And that's like really, really exciting. So I think the first thing is just a really exciting time to be a vendor. Anytime you go for a hype cycle and we're in a hype cycle, the underlying technology can be real. You can still also be on a hype cycle. Those two things can be true at the same time. Everybody is obviously just added on that AI token features or say that AI, like the scheduling interview example out there that aren't. What's really interesting with our approach on product marketing is actually we've really lent away from AI. Because I hate quoting Steve Jobs. I really didn't. It's a site. Who am I to quote Steve Jobs? But he has this brilliant line like always from our team off, which is the end user does not care about the technology used to solve their problem. They just want their problem sold 10 times better than what it was before. And so actually it is irrelevant if you're using AI or not. How are you solving somebody's problem 10 times better than you were before? And if I was a prospect or company out there thinking about buying an AI tool, that is really one of the then desire to be thinking about it is ignore the two letters A and I and actually just how does this product work and how to look. The final thing I would say and I think what a lot of vendors are completely missing. I was on a call this morning with one of the large, well, the largest organization for second boys in the world. They are proactively not purchasing AI tools in recruitment right now. They've actively said we are not going to do this. And the thing, Matt, the thing we have to realize is recruitment really impacts people's lives. Like we are taking decisions that impacts people's lives. And we are a high risk category. If you have an AI system that is going to come in and take decisions on your behalf and that AI is somehow biased, which I've got a whole view on more broadly, that is going to really impact people. And there was an example in 2018 or 2019, one of the big tech companies I want to say who did automate some of the hiring process and did come out that it was actually inherently biased. So I think what vendors are missing is well, there's a lot of excitement around AI. There's also a lot of caution in the procurement process of actually purchasing AI products right now. And I actually think third is so given the industry that we plan. So you mentioned a couple of things there, but a great advice for anyone who's looking to buy a recruiting piece of recruiting technology with AI in it or an AI centred kind of solution. What other advice would you give to people? What sort of questions do they need to ask, or do they need to be aware of? Yeah, so much of this is around data governance. So how is your data getting shared and what are the models that the company is using? So are they using just an off the shelf open AI model? Are they self hosting an open source model like llama from meta? Are they fine tuning that own models? That may sound like technical jargon. And it kind of is technical jargon. There's a pretty simple question to ask. And then where does your data flow back through that stack? Because do you really want your candidate data powering some crazy model that's not even just being applied to this one product. So understanding that piece is really, really important. And respectfully, a lot of the products that have been built are just a very simple UI layer on top of a chat GPT. So that's the data governance piece is really clear. And if you do not know that, and maybe you're on a smaller organization, maybe they've reached some of your tech team type answer questions, if you're in a larger organizations, your procurement team are going to be all over the data governance piece. So that's kind of the, and the next piece I'd be looking at. From there, look, my general advice to any person is looking at vendors is there are a million and one tools out there, buying a tool does not solve your problem. Your users are going to have to use that tool and apply that tool to solve the problem. So how can you make that as easy as possible? I would always be looking for integrations like, can this plug into my main system? And as you're talking about ripping out your ATS or your CRM, actually everything else needs to plug in there. Because getting adoption of the technology is key. You know, technology itself isn't going to solve the problem, you need adoption of the tech. So if we go back to what I said earlier, take out the letters A and I, and actually just think, does this product solve a problem I had? Really be mindful of data governance and how is data flowing through from your systems into theirs and how are they using that data? And then does it integrate into my team's work? So they're actually going to use this technology in the first place. Final question for you. And it's question two parts. So what do you think things are going to look like in the future? And just to kind of frame that slightly, what do you think things will be like in one year's time? And what do you think they might be like in five years time? Oh, what an amazing question. So what's interesting is I think in one year's time will actually just start adopting the technology that is roughly available today. So I think actually what will be technically feasible will be far greater than where we are today. But from actually a TA industry perspective, a lot of that risk I talked about the largest company for tech people in the world that I spoke to this morning and a year's time, they will be leveraging AI in their sourcing team. There's no doubt about it. So I think in a year's time, we're not going to, the world isn't going to look too different from a TA perspective than what it does now. Because I think that's sort of the rate of adoption. However, having said that technically, what will be feasible, I think will be very different. My overarching view, which is where I'll get to on five years is AI is going to have a far bigger impact than we could all over imagine, but over a longer time horizon. And I think when you start looking at the five year horizon, you do start questioning, okay, well, then really, how has this embedded into our workflows at this point? So for instance, do I have a agent that is always on and me as a recruiter, the agent is watching everything that I do and just plugs in gaps on the way. So the agent sees that I'm on the core of a hiring manager and is listing in real time. When I say agent, effectively, like an AI agent is listing to the core of a hiring manager. And in real time, it's gone in source 20 candidates. And at the end of the call, it's gone, by the way, here are your 20 candidates and the students and all three people, I guess, they were having the call. That's the level I think you could end up getting to, right? And where my current mental models and this co-pilot, actually, I think you end up with like a co-agent. And all of us have these agents that are doing work for us in the background and coming back and the data sources they have access to our incredibly rich. So that, to me, is a very exciting future. I think it's going to be amazing to see how the world of work changes. I do think that this will be a fundamental technical and platform shift. But I think it will happen over a much longer period of time than we can all imagine. Mark, thank you very much for joining me. Nah, absolute pleasure. Thank you so much, dude. My thanks to Mark. 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. [MUSIC] This is my show. [MUSIC] (upbeat music)
I know that many of you would agree that we are now beyond saturation point with the hype around AI in TA. It's clear that one way or another, gen AI is going to become ubiquitous in the software that we use. So, rather than discussing it as a topic on its own, we need to look carefully at the current and future impact on TA in terms of use cases, innovation, and the potential to rethink hiring.
So what are the proven use cases for AI right now, where are we missing opportunities, and how will things develop in the short term?
My guest this week is Mark Chaffey, Co-Founder and CEO of Hackajob. We discuss current AI use cases and lessons TA can learn from Gen AI's impact on software engineering.
In the interview, we discuss:
The current state of the tech hiring market
Concerns over developing early career talent
The impact of AI on software engineering
AI isn't just about automation. It's about being ten times better.
Why AI will always match better than humans
The role of TA
The hype cycle
Critical questions to ask vendors
Data governance and integration
What will things be like in both the short and long-term future?
Follow this podcast on Apple Podcasts.
Follow this podcast on Spotify.
