The public launch of ChatGPT continues to drive enormous levels of hype about the impact of AI on life, the universe and everything. So what are the implications of generative AI on talent acquisition, how will it shape the development of recruiting technology, and what will candidates think of the changing balance between humans and machines in recruiting?
My guest this week is the perfect person to give us some answers. Adam Godson is Chief Product Officer at Paradox, and his unique combination of experience in the industry means he has a deep understanding of the potential of AI and the likely future direction of talent acquisition.
In the interview, we discuss:
• The current state of the market
• The difference between Generative AI and Conversational AI
• What changes will generative AI drive in recruiting technology?
• Why is there a rush of Large Language Models coming to the market at the same time?
• Decisions, Communication, and Automation
• Enriching interactive conversations
• Time reduction
• What is the art of the possible with personalization?
• The implications of “quick apply” on employer brand communication
• Polarising messages
• Calibration by the candidate reaction to AI
• What will the next two years look like, and where is technology heading next
Listen to this episode in Apple Podcasts.
Transcript:
Paradox (0s):
Support for this podcast is provided by Paradox, the Conversational AI company helping global talent acquisition teams at Unilever, McDonald’s, and CVS Health get recruiting work done faster. Let’s face it; talent acquisition is full of boring administrative tasks that drag the hiring process down and create frustrating experiences for everyone. Paradox is AI assistant. Olivia is shaking up that paradigm, automating things like applicant screening, interview scheduling, and candidate Q&A so recruiters can spend more time with people, not software.
Paradox (40s):
Curious how Olivia can work for your team? Then visit paradox.Ai to learn more.
Matt Alder (Intro) (1m 5s):
Hi there. This is Matt Alder. Welcome to Episode 515 of the Recruiting Future Podcast. The public launch of ChatGPT continues to drive enormous levels of hype about the impact of AI on life, the universe and everything. So what are the implications of generative AI on talent acquisition? How will it shape the development of recruiting technology, and what will candidates think of the changing balance between humans and machines in recruiting? My guest this week is the perfect person to give us some answers. Adam Godson is Chief Product Officer at Paradox, and his unique combination of experience in the industry means he has a deep understanding of the potential of AI and the likely future direction of talent acquisition.
Matt Alder (1m 56s):
Hi Adam, and welcome to the podcast.
Adam Godson (1m 58s):
Hi, Matt. Great to be with you.
Matt Alder (2m 0s):
An absolute pleasure to have you on the show. Please, could you introduce yourself and tell everyone what you do?
Adam Godson (2m 5s):
Hey, Matt. Yeah, absolutely. Great to be back. Thanks for the re-invite. I’m Adam Godson. I’m Chief Product Officer at Paradox. And so, my goal is to think about Conversational AI in recruiting and then to make those things happen.
Matt Alder (2m 18s):
Now, I want to talk to you about all things AI because it seems to be absolutely dominating the recruiting news cycle at the moment. It’s very difficult to get away from it. Before we do, though, by way of context, obviously Paradox has lots of clients, deals with lots of companies, and helps with the recruitment issues, all those kind of things. What’s going on in the market from your perspective at the moment? On the one hand, we’re hearing about economic downturn, recruiters losing their jobs, companies not recruiting. On the other hand, we’re hearing about talent shortages and companies not being able to get enough talent into their business. What’s it like from your perspective?
Adam Godson (2m 53s):
I think it’s some of each. It depends, and I think that’s the part that’s confusing. We’re not in a cycle that is clearly down or clearly up. It’s up in some sectors or regions or industries and down in other sectors, regions, or industries. And I think that’s making for a confusing overall narrative. And so, in many ways, sort of an array of different directions that’s happening in the market. But if you look at the macroeconomic picture, you see some unemployment that is still near record highs in the US — I’m sorry, near record lows in the US and in Europe. And yet companies are certainly starting to pull back, and think differently, and being more cautious.
Adam Godson (3m 36s):
And so I think there’s still a lot of anticipation about which direction things will head. And that’s creating some confusion in the markets, and I think sometimes some narratives that aren’t matching reality as well.
Matt Alder (3m 47s):
Now, the last time you were on the podcast, we were talking about Conversational AI and personalization. And it was probably about this time last year, I think. Now, since then, in the last few months, Generative AI has hit the headlines big time. And it seems to be the only thing that people want to talk about at the moment. I wanted to ask you this question because I thought you were the perfect person to answer it for everyone listening. Can you tell us the difference between Generative AI and Conversational AI?
Adam Godson (4m 18s):
Yeah, certainly, it’s captured the moment Generative AI has. So, Conversational AI is going to be an automated conversation. So, the method of interacting that feels like a conversational interface via voice, or chat, or video that is two-way between one or more people or machines. Where Generative AI essentially is talking about content that’s being generated for that particular purpose. And it can be conversational. That’s certainly what you see with ChatGPT, for example, and other large language models that are really good at that. That is generative and that it’s being generated so quite literally.
Adam Godson (4m 59s):
But generative also has a part that’s not conversational. So the generation of an image, for example, through stable diffusion, or Dolly, or another service like that. So if you draw your Venn diagram, the generative can be conversational, and certainly, Conversational AI can be generative or non-generative as well.
Matt Alder (5m 19s):
Companies have been working on these large language models for quite some time. But everything just seems to have launched all at the same time, and it really caught the imagination of everyone. And I think we’re already deciding to see some changes happening pretty quickly. What changes do you think that Generative AI will drive in the recruitment industry? I suppose particularly in recruiting technology products.
Adam Godson (5m 45s):
Yeah, it’s interesting. And you rightly point out that the Generative and Conversational parts mattered a lot. Where GPT-3, for example, was out for nearly two years. And we’d certainly used it in different ways just as a language model and experimentation before it really caught the public’s imagination. And what caught the imagination was the creation of ChatGPT into a conversational interface. So taking this language model that’s based on GPT-3 that you really could have asked most of those same questions before, and it did lots of similar things, but making it just a box on a page that could then quick carry on something that was very conversational made to go to million users in five days.
Adam Godson (6m 29s):
And so for us is sort of that story around conversational as a really important interface in a way that people want communicate and like to communicate in a different way. And then, of course, the next stage was everyone getting lots of pressure of, “Oh my gosh, we’ve working on this project for a couple of years. Get it to market. So we don’t like, we’re behind. So you saw Facebook and Google like what seems like an absolute barrage of things that were developed in two weeks were more about the release cycles coming up really quickly and people feeling like they had to get something, even an imperfect something out the door.
Adam Godson (7m 11s):
So, you seems like it’s all happened at once but lots of research that’s going into these. And from a recruiting standpoint, there are lots of things that are going to change. And you rightly mentioned that in technology, many folks will change the way that those interact. And for me as I boil down recruitment, it’s always really been about three things. And I’ve stared at the ceiling many nights thinking about what it is that we actually do. And it’s really about decisions. So decisions about a resume or effort an interview. It’s about conversations, and then it’s about automating lots of stuff.
Adam Godson (7m 51s):
We call the BS of Recruitment the boring stuff. But scheduling interviews, sending off letters, all those things. And the latter two really changed with this technology. So our ability to automate things, but it’s really that second one to be able to have interactive conversations in lots of ways that weren’t really possible before. And so, being able to have an automated chat. And being able to that’s really high quality conversation. Being able to get information from a hiring manager, get information about a labor market. All in a chat interface with a manager or a recruiter. That’s what is going to be substantially different going forward.
Matt Alder (8m 31s):
Paradox have recently announced that you’ve launched Conversational ATS based on the Conversational AI that you use. What is a Conversational ATS? How’s it work, and what kind of difference is it making?
Adam Godson (8m 45s):
Yeah, I think the difference in the way we have gone to market in the past and releasing a Conversational ATS was really stating what we are in that way. We, it’s a product that’s been in market for some time, where we for years sort of sat on top of existing ATS. And folks said, “Oh, can you make my ATS conversational?” And then, slowly, as people ask for certain markets. We built some of the underlying technology. And so, to be clear, we still can do both. So happy to sit on top of Workday or SAP or other systems that help make them conversational, but also can now stand alone and be able to do offer letters and some of the automation and the statuses and workflows.
Adam Godson (9m 29s):
And the things that you might find more traditionally in an ATS. So being able to do the whole job is the primary difference, but doing it conversationally. And being able to manage those conversations with candidates, and being able to gather information conversationally, and remove the friction that so many ATSs have in throughout the process. And do that through automated conversations, which leads to faster time to hire and hire candidate satisfaction.
Matt Alder (9m 57s):
Are they the kind of results that you’re seeing with the clients that you’re working with? Is there anything surprising that’s coming out?
Adam Godson (10m 3s):
You know, I think the amount of time reduction we can make through Conversational AI and mediums, like SMS and WhatsApp, for example, continues to surprise me. You know, we, of course, have done work with lots of restaurants and retailers, like McDonald’s and others for years and made substantial gains in those processes. But just last week saw the numbers from a client that went live about two months ago and to hire went from 31 days to nine, and just all the metrics with massive improvement.
Adam Godson (10m 44s):
And so, even after three and a half years of doing this, seeing those types of huge gains continues to surprise me in a positive way.
Matt Alder (10m 53s):
The last time we spoke, we were talking about personalization and how Conversational AI can drive personalization in the recruiting process. How have things developed since then? Are we seeing more personalization? Are there more things possible? What do you think could be personalized in the future?
Adam Godson (11m 13s):
Yeah, that’s where I think large language models and Generative AI are changing the game as we speak for that. So, being able to look at someone’s LinkedIn profile or their CV, looking at a specific job, and being able to create synergy to say why are you a good fit for this job given the experience you have, the job requirements, the culture of the company, and being able to automate some of those conversations is really exciting. To be able to really personalize in that way. I think the other thing that we’re seeing is some need for personalization and some more content in hiring processes as more companies use quick apply frameworks.
Adam Godson (12m 2s):
So certainly, this has lots of the US this year with Indeed change to pushing hard on Indeed apply and others ZipRecruiter and et cetera, to essentially have the candidate apply without ever visiting the career site. So simply sending all the candidate details through an API. And so what that has led to is lots of candidates that have applied to a job but have never visited the career site of the company. And so then, how do you do the employer branding activities that you might have done on a career site but do that throughout the process in a more linear fashion. So sending people things before the interview or after they apply, or write messages at the right time, and then being able to personalize those as candidates go through the funnel.
Adam Godson (12m 52s):
I think there’s tremendous opportunity to take what was a transactional experience on a career site and make that experiential throughout the life cycle of the recruitment process.
Matt Alder (13m 2s):
So do you think that’s people applying for the job then finding out more about what it entails? So what is it sending people video? Having conversations, letting people talk to hiring managers? How do you think that’ll evolve?
Adam Godson (13m 15s):
I do think it’s some content, and again, right content at the right time is the key principle of that. But I think the risk in quick apply frameworks, et cetera, is that we don’t get all the polarising messages in front of job seekers as we otherwise might, right? Employer branding done well is polarising. To say, you’re the type of person that fits really well here, or you’re not. And to be clear about that. And so I think what happens without some of that content at the front end of the process is you’ve got to build that into the middle of the process. And so letting people, before they come to the interview, understand your values.
Adam Godson (13m 57s):
Why you should work there? If this is the right place for you? And if not, opt out. And just be sure that we’re getting the right people for the job. And the the risk is that we just move the bottleneck down the funnel one step and make no actual improvement. The possibilities are that we can communicate well throughout, and we actually make some gains by removing some friction at the beginning.
Matt Alder (14m 20s):
Absolutely. And I think it’s that kind of transparent communication that people want from the recruitment process. If the job’s not right for them, they want to understand that and understand why. So I can only see that as being a really good thing for the candidate experience.
Adam Godson (14m 36s):
Yeah, absolutely. Transparency, it’s what people want. So, no one wants to waste time in a process that’s not right for them. And so if our job as professionals is to help remove friction, of course, but also help people make decisions. And it doesn’t do a company any good if they hire the wrong person for the job. That tends to go go poorly on all counts. And so being sure that we’re as clear about the expectations, the type of company to make it work for the long term as quickly as possible to avoid wasted time.
Matt Alder (15m 6s):
It feels like we’re on the edge of some massive changes in recruiting sort of being driven by technology and the sort of the continued disruption around the economy. People making decisions about their careers differently, [unintelligible] of things going on that are inevitably going to drive a lot of change to what we do. What do you think the next two or three years are going to look like for recruiting and talent acquisition?
Adam Godson (15m 35s):
A bit of cynic might say. It’s going to look like automation. Some overdoing of automation probably backlash against automation. And then repeating that cycle several times until we find the right amount. But I do think that the way that people perceive and feel about technology is going to matter a lot. And there’re going to be a lot of new technological possibilities for how we interact with Conversational AI and with companies. And how we use them and people’s reaction to them is going to matter a lot about how we use them in the future.
Adam Godson (16m 18s):
So, I can imagine a substantial backlash to mass outreach, for example, of companies looking to screen candidates from on passive sourcing for a particular role through really good Conversational AI. And candidates getting really frustrated and wanting to say talk to humans. I only talk to humans. And so I think we’re going to find out lots of wrong ways to do things as an industry. And then, we’ll focus in, adjust, and find the right ways. So, I think it’s going to be a tremendously interesting time as this technology evolves.
Matt Alder (16m 52s):
And to build on that as a final question. So you mentioned how long technology companies have been developing large language models and how you’d already been using that in the work that you do. Are there any other game changing technologies out there that everyone in the kind of technology industry is aware of but haven’t hit us yet in talent acquisition that might change the game again?
Adam Godson (17m 15s):
You know, I think the key will be like the application of this technology for language models. I think this is going to be a several years journey of really refining the application layer, and how this technology gets used. There are other interesting technologies being incubated in this generation, sort of text to video. You can imagine an experience that can be built today with a pretty lifelike person being able to interview someone that’s fully automated voice cloning, deep fakes. So, some of those experiences are really, really good where they haven’t been in the past.
Adam Godson (18m 2s):
And the conversation with many of those technologies is just shifted from not can we do these things but to, should we do these things? Are they something that people want? Are the experiences that feel good? I think I’ve continued to argue just with some friends conversations over a drink that so many times it’s actually the emotional element that someone cared to take the time to talk to me. More so than was the video smooth on something. And so, I think some of the emotional elements are yet to be worked out, and understanding and poking at those questions of, what role should technology play in recruitment process versus, you know, do we have humans?
Matt Alder (18m 45s):
Adam, Thank you very much for talking to me.
Adam Godson (18m 47s):
Thank you Matt. Always a great conversation. I appreciate it.
Matt Alder (18m 52s):
My thanks to Adam. You can subscribe to this podcast in Apple Podcasts, on Spotify, or via your podcasting app of choice. Please also follow the show on Instagram. You can find us by searching for Recruiting Future. You can search all the past episodes at recruitingfuture.com. On that site, you can also subscribe to our monthly newsletter, Recruiting Future Feast, and get the inside track about 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.
