Recap of #HRTX AI in Recruiting Event

All right, all right, I hope you like my fight song. That's my fight song. Thanks to Brooke putting that together. Got some interesting news for you today. I don't have a guest, so you're just going to have to put up with little old me for a little while. But I think I might still have something that could potentially be of interest. You can tell from my HRTX Camp shirt here, I just recently was at the HRTX Virtual and I'm going to give you a recap. So that's right, if you did not go, you are in luck because you're still going to get a quick summary— let's call it a crash course— on what happened at the HRTX event yesterday. And I'll show you some cool things on screen. Also, I'm going to announce winner, winner, winners! We have winners for the learning challenge and the expo, the two challenges, for those of you who did participate. And these winners are getting free access for one year to my new ChatGPT course for recruiters and sourcers. So if you hear your name, raise your hand and I'll, uh, I'll mail you the course. No, I'm just kidding. If you hear your name you'll be getting an email with access to the course. The winners for the learning challenge are Don Bozeman— hope I said that right— Winston Murray, and Takisha Williams. And for the expo, we have Shruti Chobe, Vanessa Strum, and Dana M. Not sure what the M stands for, but it's probably something really cool. So the 6 of you step right up and claim your code for the ChatGPT course. You're probably wondering, what is the ChatGPT course? Well, it is a new, recently released learning module, online learning module you can take. It's a self-paced course that you can take on the TSI University website. If you want to know where that website is, then all you need to do is follow this link that I'm going to put in the ticker right now, and that is tsiuniversity.com. That's the school, and you can see the LMS system and all the courses. There are 37 courses online right now, but there are a total of close to 100 different modules and topics, and they've all been, uh, you know, they're on the process of getting transferred in there. So check it out. The new one that you'll see there is the ChatGPT course, and I can take you directly to that one by pointing you to this website. Well, technically it's not a website, it's a link. Here you go, I'll put that in the chat. There, pretty simple to follow. It's just basically srcn.co gpt. It's an intensive introductory course on everything that you can do with GPT that is very practical. So this is not the fluff that you see on blog posts where people talk about, oh, you can rewrite your job descriptions. No, this is like hard-hitting stuff, including some sourcing techniques and things like that. Go to that link there and you'll be able to see the outline of the course. Um, right now, don't tell anybody, it's on sale for $49, uh, because I was giving it out as a prize and had it on sale yesterday and I haven't taken the price down yet. So if you go right now and take advantage of— it's probably going to take me most of the weekend to get around to going in there and changing the price back to the regular, the regular price, which is $129. But if you act now, you can get it for $49, except for those 6 people that I just mentioned. Who are going to get it for free. All right, it's enough of my shameless plug in my own show. I did promise you that I was going to be giving you some, uh, crash— a crash course, if you will. Not some, but actually just a full-on crash course on what happened at the event yesterday. So let's get to it. There were some pretty Let me just say, overall, the content was solid, like, throughout the entire event. Uh, very well put together. Brooke and Ashley from Recruiting Daily had a cool kind of theme to it. It was basically Hitchhiker's Guide to the Galaxy, so it was all kind of wrapped around that. But we had some really cool speakers that, you know, that, that contributed significant amounts of aha moments like Katrina Collier. We had Vanessa Rath. We had Rick Osborn and Justin Swisher. Myesha Cannon was there. Blair Fambrough. Oh my God, Blair, I like, I have a total man crush on this guy. I'll tell you a bit more about that. David Marr, you might know him from his LinkedIn sharing of jobs. So everybody out there that's looking, uh, Tyler Taldon and Kat Kibben. Um, I also have a crush on, on Kat. And Amit Bhatia, and the topics were pretty varied. The first one to come on stage was Katrina Collier, and she talked about balancing tech and touch, and it was about, um, the, the, the kind of use with caution statement which really permeated the entire event. I think a lot of the speakers had this one theme, maybe not by design but just because it just makes a lot of sense, in common, which was use wisely. You know, it's AI and generative language models, and these tools are tools that allow us to be more efficient in various ways, but they also require human supervision. So it's still very much important that you essentially pay attention, right? It's not something that you can just kind of put on autopilot. Just because it's a robot doesn't mean that you want it— like self-driving cars, right? You still want to be behind the steering wheel. You still want to pay attention. You can't have it— we're not at the point yet where we have self-driving cars where you can get in the back seat and nobody is behind the wheel. It's the same kind of concept, right? None of these robots are completely autonomous robots. They're, they're assistants. They help us do things like, like robots, uh, in a factory, you know, that are helping us with assembly. There's, there's still human involvement in all of that. So What did Katrina cover with balancing tech and touch? Well, her biggest takeaway, I think, was stop using so much technology that it creates barriers between you and the candidates. There's so much technology between us that people are just kind of getting non-responsive. Their response rates are dropping all over the place. Now, this is my own sidebar here. LinkedIn response rates are down, email response rates are down. Texting response rates are down. Why? Because there's a lot of this noise being created by automation, and that is reducing the response rates. Think about it, that the appropriate and, and considered, considered response, uh, to people getting spammed is to just ignore it because you can't stop it. You know, kind of like how we all got desensitized to advertising on TV, and that's why nobody pays attention to ads anymore because there were just so many of them, and now we like you know, we can't skip them, thanks TiVo, we fast forward them or we just buy our own television access which doesn't have advertising. It's the same kind of idea. So here's the other thing that Katrina said, besides the cautionary statement, is a lot of this automation makes you vanilla. It makes you the general flavor. I jested, not to her enjoyment, but I did jest that I like vanilla. However, my point was that, you know, vanilla is a great flavor. Lots of people like vanilla, but we like to mix it up. We like Napoleon, where we got the vanilla, the strawberry, and the chocolate mixed in there. I like vanilla with balsamic vinegar, so it's still vanilla. You know, vanilla is a great flavor, but you don't want to be just plain vanilla. It can be boring. And so what she was saying is the automation is vanilla-izing. I think I just made up a word. Vanilla-izing what you do. And people want to connect with you, not with this robot that is, sure, assisting both of us, right? It's assisting the recruiter, it's assisting the candidate, and we can use the technology to our advantage on both sides of the negotiating table, so to speak, but not at the expense of actually representing who we truly are. That was her takeaway. She used 3 words, I I guess, yeah, curiosity, self-worth, and fearlessness. She said, be curious, know your worth, and be fearless. Ramp up your self-worth. A big part of that she spent on the intake. Now, if you know anything about any of the courses that I teach and the presentations I've done all over the world for the last more than 25 years, you know that I have always maintained that the intake is the single largest point of Failure. You heard that right. Failure. Single largest point of failure in the entire recruiting ecosystem. Why? Because it's not done standardly. It's not done in a professional way. It's haphazard. It's done casually. It's not consistent. Sometimes it's skipped completely because of hubris, our own hubris. We already know that, right? I already know that hiring manager, so we don't do it. And we skip a lot of stuff that we really did need. So what she said is do a better intake. And by this she meant show up with market data. Know what the market is like for this role. Competitor intel— what are they doing better than your company? Apply for competitor jobs and see what they're doing, how they're doing differently. Get candidate data— how many people actually do this job? How many people do this job in that geography? How many people do this job with that job title or with that level of experience? Right? I mean, maybe you want people that have 10 years of experience in a technology that's only 10 years old, where there's going to be a lot fewer of those than there are people that have 9 years of experience. Right? And she also said, show up with internal data. I really like this. By that, what she meant was, when you come to the meeting, also be prepared to talk about who at the company could possibly do this. And I'm going to add to that, also who could do this or help you. Find people that who could do this. And what I mean by that is maybe there's somebody at the company that is knowledgeable in this technology, but they're not able to do the whole job, but they do have access to a community. Like for example, they're a member of a discussion group or they have a subscription to something or they just know what that population is interested in. So knowing that going into this conversation will help you open up channels. Another really big point that she made, and I think this is probably in my opinion, my biggest learning from that, it may not be yours, but it definitely was mine, Know your success data. Now, this is weird for me to say because I know my success data. I report on our team's success data, and I'm constantly talking about it because a big part of my job is analytics and reporting and strategy. That's my, my main job is strategy. So I know what we can do, what we have done, and what we need to do to get done what the leadership tells us we want to do, right? I'm not setting the strategy for the company. I'm setting the strategy for Talent acquisition. The CEO tells us where we want to go. We want to expand our market, we want to go international, we want to add product, whatever. And then I have to connect the dots between that and what TA can do to support that. And in order to do that, I have to know where we've been and where we can go and how fast it's going to take us to get there. But I never show up to these intake meetings with that. And the example that she alluded to, which really lit me up in a positive way, is if you're able to show them your results, your success history, and say, here's what we've done for other hiring managers, or here's what we've done for you in this, you know, this hiring manager in the past, they're a lot less likely to have recommendations that are useless or objections that are wasting your time. So I thought that was a really important point. And in particular, I thought to myself, okay, what if I show up to a hiring manager meeting where the hiring manager is insisting that we use an external firm First, I might be able to say, well, in the last 52 jobs that we've hired for your department, 3 of them have been through search firms and 17 of them have been through the sourcing team. Knowing that data, which I already know, is one thing, but bringing it to the meeting and showing up with that data is completely different. That hiring manager is going to have a whole other perception of what sourcing is capable of doing because I just told them what sourcing actually did. And so that was really big for me. Again, I knew the data, but I just didn't bring it to the intake meeting. I just knew it post-fact, right? Now, another one that I'm, I'm going to share with you from Katrina, and this took a lot of chutzpah, and it's going to take a lot of chutzpah for you to do it, and I applaud you if you can do it, but I also, I can understand if you're a little apprehensive. You know how sometimes you have a hiring manager that just doesn't really respond to your requests to set up an intake meeting or a kickoff call, or, you know, they're really, really hot on opening the rec, and then you open the rec, and then weeks go by before they even give you time of day. What she said was, "Close the req." OK, wow, that's, that's a big step. But you know what? I agree, because we do have the power if we, if we are assigned the req, we do have the power most of the time of opening and closing the req. And yeah, don't do it in a way that's going to get you fired. But if you're not getting any response from the hiring manager, close it or put it on hold. And you know what? That's going to get a response from them. And then you can say, oh, I'm sorry, we put it on hold because we didn't really know it was that important for you since you hadn't responded to our outreach to set up a meeting. We don't really know what you're looking for, so we couldn't really start on it. Now that I have you on the phone, let's go ahead and reopen it and get it going. I love that. Okay, um, 2 questions that she asked about the intake that I would love to add to my intake, and I probably will right away. The first of those is asking the hiring manager, how will you know if at the end of 12 months you hired the right person? So I ask it a little bit differently, but I like this way better. The way I usually ask is, pretend that this person has done a great job and it's time for their annual review. What are they being praised for? Well, that's a little bit more complicated and kind of a little bit roundabout way. So the way she says it is, how are you going to know at the end of 12 months that you hired the right person, which means you didn't make a hiring mistake? And that's going to cut right to the chase, and I love it. Second question is, What happens if we don't fill this role? Great, interesting, right? I mean, hey, failure is an option. It happens. So what happens if we don't fill this role? What's the downstream of that? There you go. So yeah, Paul, Paul says we put it on hold. That's right. If we cancel it, they have to— yeah, yeah, okay. Makes sense. Which means that that's now you're creating a whole lot of like, uh, ribbon, you know, bureaucratic ribbon for, for them to cut through. So I get that. But yeah, putting on a hold, same thing, same idea. Love it. All right, so moving on. Vanessa Roth, uh, live from South Africa. See how the Africa— I'm not doing a really good job of that. That's terrible. I apologize to everybody that doesn't speak South African. Um, that was really horrible. But anyway, she talked about harnessing AI for recruiting efficiencies, and she talked about using ChatGPT to empower and accelerate, not automate and replace, which is kind of what I started with. So in other words, you know, ChatGPT and Bard and all these other large language models are really only tools. They're, they're there to help you do things faster. So you already know how to do them. This was a great point by our own Brooke Allen at the very end during the recap. She said, you know, this, this technology really only works if you already know how to do your job. It's not here to do your job if you don't know how to do it. So that was Vanessa's point, and a beautiful point, right? It's here to help us accelerate. It's not here to, you know, take over or do it for us. It can just help us. And she actually gave a really good example of that, and I'll tell you in just a second here. But I also wanted to point out that she brought up a couple of the really negative side of these ChatGPT and language models, which I've pointed out a couple of times on this show before. I was really enjoying the fact that other people have noticed this. That is that these language models have really bad math, as in their math just doesn't add up. And sometimes they hallucinate. Now, we all know that most of them are out of date because it's just the nature of, you have to train the model and make sure it's not doing anything bizarre, even though sometimes that does get through. But she did point that out. And I think I used the example before, and I might have put it on here before, where I asked ChatGPT to tell me how many E's there are in the word between? And it said 2. How many E's— like, literally type in how many E's— I'm actually doing it right now— are in between? So I'm going to specify how many letter E are in between, in the word, but I'm gonna say the word between. Okay, uh, you know what, I'm gonna, I'm gonna show you. You're not gonna believe this. I'm gonna actually show you live. Check this out. I'm gonna share my entire screen with you and show you live what I said. I said, how many letter E's are in the word between? Okay, here you go, world. This is ChatGPT's answer to how many letter E's are in the word between. The word between contains 2 letter E's. Yeah, no, it doesn't. No, it does not. It contains 3 letter E's. Apologies for the confusion. You're correct, the word between contains 3 letter E's. Okay, how many letter E's are in the word between? Gonna ask it again. It still contains 2. Weird, right? I mean, like, super weird. So bad math is definitely a problem. That was one of them. And another big con is the hallucinations. So how do we use these kinds of models? Her advice is give it a framework and then use your brain to build up. So it's not like you can give it a single answer, or I mean, sorry, a single question, it gives you the perfect answer. It's iterative, right? So here's— I told you I was going to give you a tool for this. Basically comes down to what she said was AI is a starting point, not an ending point. I love that. What the question that is asked the most? That's the question which is asked them. Which one? How many E's are there in the letter between? Really? That is the most asked question? Okay, I did not know that. That's a— that's news to me. Sorry, so I digress. So here's what she said. She said AI is a starting point, not an ending point. So think of these tools like ChatGPT as a way to get from zero to 60%. So you need ideas, you need brainstorming. Use these tools to get you started, right? Like, give me a list of 10 things that I could do tomorrow morning for breakfast or whatever, but it's not going to get you all the way to cooking your breakfast and feeding it to you, right? So go from 0 to 60, or you have something and you just need to refine it. So she said you can take it from 75 to 100%. So kind of start— use it to start out and get some brainstorming, or use it to polish at the end. But either way, you have to be the one that's in charge of creating it. For example, um, don't just give it a piece of text and say, correct the grammar. Instead, say, check my grammar. And then what it'll do is it'll come back and give you advice on where the grammar might not be correct. Which allows you to do two things. It allows you to make educated decisions about whether or not you want to make those corrections rather than just making the corrections for you and then relying on its assumption, um, that might, I don't know, incorrectly put the— or remove the letter F from, um, a word, or put the letter F. I'm trying to think of a good example. So let's say that if you have a job title that is, uh, shift work, and it spell checks it and removes the letter F because the other word is also correct, then you have a job title that is a shit work, and that could be a bad thing, right? So it allows you to check the grammar and correct it yourself. And the second thing it does is it teaches you, because it's actually showing or pointing out what the, what the mistakes are, which allows you to learn from it, right? And I think that's important. You can also decide whether or not that's actually something you want to change. Every once in a while I use poor grammar as a pun. It's incorrect on purpose, right? So I don't want to lose that. So anyway, I thought that was pretty cool. We also had Rick Osmer and Justin Swisher, and I think the big takeaway from there was going back to this whole theme of it's really about leveraging tools to be more efficient. Um, the, the big takeaway I got from there is they said first you need to understand what the efficiencies are in your process and quantify them. And I really like this. I'm a data person, so, you know, I, I kind of cringe when people say we need more candidates. Uh, how many more? Like, what, what does more— first of all, what does more mean? And secondly, more compared to what? What, more per day, more per hour, more that make it through the first interview, more that make it to offer, more that are of a particular origin? What does more mean, right? And then how much more? So what they're saying is, first you have to understand where the bottlenecks are, where the inefficiencies are, then measure them, quantify them, and then you can look to AI because then you can actually measure if the solution had an impact, and that's brilliant in my opinion. I'm, I'm sort of adding to this, to what they said, but my thoughts are if you know that your process is inefficient in this area and you know that it's slowing you down by X amount, then when you put some solution in that speeds you up by X amount, you know that you've made X amount of an impact. That's called the delta, the difference between before and after. It used to take us 8 hours to do this, now it takes us 6 hours to do this. So we gained 2 hours per widget or job rec or whatever, right? Love that. After that came Maisha. Now I can't say enough good things about Maisha. She is really good on stage, um, and just kind of— she's got this slow burn. It sort of builds up. She's very gentle on the uptake, and it takes a little while to warm up, but when she gets going, man, she's a powerhouse. And so kind of ramped up, and then she just delivered this Massive chunk of wisdom. I love her SWAT. SWAT is something that I, I was a Peace Corps volunteer, and I was, as a Peace Corps volunteer, I was trained, it was hammered into us for the first 3 months. Like anybody else that's been in the Peace Corps and is a returned Peace Corps volunteer, you'll recognize this. It's called appropriate technology, okay? So something that we do when we go out and we work with people in, in communities that are either in a disaster area or they're impoverished or whatever, we don't go in there with millions of dollars and try to build some big electrical facility or whatever. We are taught to use appropriate technology. That means use what you have, use what's available. And so she makes this, this SWOT that basically start with available technology, which translates to me as appropriate technology, which means Why don't you take full advantage of what you already have before you start going out there and looking for new tools? I run into this a lot. I ran into this a lot when I was a consultant where companies would ask me what technologies I would recommend. And the first thing I would do is, well, what are you using right now? Because I'm not going to recommend something that is essentially a different— like it's an alternative but similar version of what you're already doing. That's not really going to do any good. If you have problems with what you're doing now, is it the technology or is it the part of the problems that are you? Are you the process, the problem, whatever, right? So start with available technology, then go look for a solution. She also gave a couple of, I don't know if these are trade secrets or not, but we talked about how response rates are diminishing. She quantified that. She said the average open rate for a 3-stage sequence, so you got email 1, follow-up, follow-up, right? The average open rate for a 3-stage sequence is 80%, which doesn't seem terrible. I'm used to a little bit higher, but she's saying that it's essentially gone down, which meets up with what I was talking about earlier. But the reply rate, the average reply rate for the same 3-stage sequence is 22%. Now, I would point out that if only 80% of the people open the email and 22% of them replied, that actually is closer to a 27% reply rate because 20% of the people never got it, so you're looking at 22% of the 80%, and it, you know, that's actually 27%. So don't want to confuse you with math, but basically it's a little bit better than what she said. However, it's still a very good point, and to me, 10 years ago, that would have been unacceptable. I mean, I would have been looking at my team and saying, why are you only getting 22% responses? Why aren't you getting 60, 70% responses? I remember getting 80, 85% responses 10 years ago. And before that, I remember getting 95. I had campaigns that got 95% response rate. If you target it correctly, that doesn't happen anymore. But I didn't realize it had dropped down to 22%. So, wow. Anyway, that was a big reveal. She also gave a couple of ideas kind of in a brainstorming way for things that you could possibly use AI for. And one I have never even thought of, which was brilliant, was multi-time zone scheduling. Wow, I wish I had known that before. I mean, I'd never thought of that before. I've had meetings with various countries in time zones that are very dispersed. So for example, I have to coordinate a meeting with California, Atlanta, London, India, somewhere in India, like say Chennai or Pune, and Tokyo, right? Now in India they have a half-hour differential, so that's like off-kilter a little bit. And then in some parts in the United States, like for example Arizona, they don't observe daylight savings, so they may be in Pacific time or they may be in Mountain time depending on what time of year. And so this can get really confusing, but if you go to one of these language models and you say, I need to schedule a meeting between these 6 cities. What is the— what are the best time options? It gives you a list of times that you can suggest to everybody, and it helps expedite that. I've had to do this using the world time clock and matching it up, and it's a nightmare. It's a visual nightmare. You got 2 cities, it's not a problem. You got 7 cities, it's a problem. And I've done that. So that was a huge light bulb for me. I'm definitely stealing that one and not giving her any credit. What am I talking about? I just gave her credit. So Anyway, got a couple of other ideas. Um, one of them was VR office tours. I'm not exactly sure how you would do that, so I'm gonna have to ask her about that. But maybe if you know, uh, comment in the chat what, what a chatbot could, could do with VR office tours. I'm not exactly— what, maybe chat with them about the— I don't know. Anyway, so, um, next up, Blair FamBro, man, this guy was just dropping the bombs of knowledge one after the other, and I was feverishly writing stuff down. Mind you, it's my job to do all this so I can analyze the content and then kind of summarize at the end of the day, so I was still writing down a lot of stuff. But as I'm writing this stuff down, I'm like, I can't catch up because I'm thinking about these ideas already, and I'm writing them, and I'm thinking about it. I'm like, wow, I'm going to go do this right now. So one idea that he, that he gave us was, uh, he hires people out of the military. Now, not everybody does this, so if it's not you, then it's not going to apply, no big deal, all right? But it's still a really cool thing to use this technology for. He gives it a resume that's a military resume and says, convert this military resume into a civilian resume, which is kind of like what the Crosswalk does. Now, the, the online Crosswalk helps you translate the codes and, and the terms and all that, but it doesn't do that on a resume. This will actually straight up convert it and give you back a resume that is, uh, well, that you can understand if you don't have direct experience with that particular branch of service. You may know a lot about military, but if you haven't been in that branch of service, you may not know what some of these abbreviations are, some of these codes are, job codes, etc. So this will help translate that. And even if you do have this experience, it'll still help translate it so you can show it to somebody that doesn't know what that experience is, right? So huge, huge kudos on that one. He gave an example of Bard, which is google.bard.com, and he said, give me images of women engineers at Microsoft. I was like, no, that's— no, that's not gonna work. I was so skeptical, but I— again, I'm trying to write notes and I just had to try it. So I go over to Bard I literally just type in, give me images of women engineers at Microsoft. It did. It gave me 2 images just like he said. I was like, no, this is a fluke or there's an article somewhere, whatever. Then the shallowness was, all right, give me 10 more and boom, a list of 10 more images and little brief bios. Mind you, this is a Bard technique because ChatGPT doesn't do images. This might work with Microsoft's Bing AI, but I'm not sure. It definitely worked on Bard, which is what I checked. And then I was like, okay, give me 10 more. And it gave me another. So now I had 22. It really did work. So it's, it's a way of doing a diversity search using essentially reverse image search. Now, you, to be fair, you'd have to do this for all the other classes as well, right? Not just stop there, depending on what your project might be. But this is a brilliant use of the technology. Third one that just really lit me up in a positive way was ChatGPT. And he said, give me a list of search keywords to find people with disabilities. Boom, folks, you guys know I'm really passionate about that. I have a nonprofit. I've been working with people with disabilities for many years, particularly blind and visually impaired. And I've, I've been doing— I literally wrote the book on diversity sourcing in 2000. 3. And I have a— when I say the book, I have a giant book full of all kinds of, you know, techniques and search strings and stuff like that. But the chapter on disabilities is like 2 pages long, if that, because there are so many different ways that people describe their various abilities or allude to, you know, being in that class that there's really not— there's no— nobody puts on their resume, I am disabled, right? But there are ways that they, that they express that, like being neurodivergent and all that. There's just too many. But when he did this and it created a nice long list, I was like, wow, that's a really good starting point that I can expand from. Again, 0 to 60%, but that's— that was better. Just that one thing was better than the little, you know, couple of page excerpts that I had that I've been putting together manually by clipping words from resumes for the last I don't know, gazillion trillion years. So another bam moment. And then the last one from Blair— not the last one, but the last one that I really liked and wrote down— he asked, this one was, I think it was, uh, I think it was Bard: what emojis do software engineers use on their profiles? Okay, um, I never thought of that. I, I'm not much of an emoji user. Um, so it was like, okay, I have done emoji search and I teach that in my sourcing classes, but it never really occurred to me to like, I don't know, research it. Um, so I was just thinking about the emojis that I know, like there's an emoji for a software developer, right? You've seen that one. And there's like an emoji for, um, the medicine for doctors, there's one for nurses, there's one for scientists, you know. But I never like really It never occurred to me to go to a large language model that has a whole bunch of data and say, when you have a bunch of software engineers put together, what are the emojis that they use? That was brilliant. Now, I think that you can do this with ChatGPT because it's not image-based, but emojis are ASCII, so therefore it would come up. Now, so I went and I tried that and I said, what? 'What emojis do software engineers use on their bios and profiles?' And it came back with computer, desktop, mobile phone, keyboard, stuff like that was pretty bland. But then it suggested woman technologist or man technologist. And I thought to myself, 'Oh, that's interesting.' So I did one better. And I took his idea and I said, 'Okay, what emojis female software engineers use on their bios and profiles? What emojis do female software engineers use on their bios and profiles? And right after the kind of blah, blah generic, just, you know, disclaimer, emoji usage varies greatly amongst individuals, but then it comes back and it gives me a couple of emojis like the woman technologist, but there's also different variations of that based on ethnicities. So if you probably already knew this, but you can click on an emoji and change the color of it. So it gave me those and it gave me a couple of other ideas, um, and I was like, wow, okay, that really does work. So there, there's a pretty cool example of something that large language models are really inherently capable of doing because there's so many profiles they can look at 100,000 or 100 million, you know, profiles and see what the commonalities are, which is not something that we can easily do as humans. You might be able to trust but verify, which is how I had arrived at my list of emojis before. So I was like, all right, well, there's a, you know, programmer, so I'm going to use that and try it in search, and it worked. Okay, but that was one. That's not the list of 30 that, that I got here. It's a pretty brilliant way of doing it. Um, all right, sorry, I'm rambling on a little bit. Tyler, Tyler Taldone dropped a couple of, uh, doozies about LinkedIn that, um, I, I really didn't know. I thought I knew, but he said 70% of LinkedIn users don't log in daily. I kind of knew that people— I, I'd heard numbers like people spend an average of 17 minutes a day on LinkedIn or something like that. But, but that's people, like everybody spends an average of, right? Which means that there are people that are on LinkedIn, if you're a recruiter, they're on LinkedIn 8 hours a day, and there are people that are on LinkedIn 0 hours a day, and that's how you get an average. But this is different. This is 70% of LinkedIn users don't use LinkedIn daily. So that means only 30% of them are active. But wait, it gets worse. A lot worse. 5%. I don't know where he got this. 5% of LinkedIn candidates receive 50% of inmail. Whoa, so 5%. Okay, 30% of them are on there daily and only 5% are getting half of the inmail. That's why your response rates are so low. Why aren't you inmailing the other 95%? What's going on here? Anyway, he also talked about email response rates. And he said— this is a gentleman from Gem, so he was using aggregate data from the Gem CRM. He said that after the first email, typically you get an average of 7% response. Then after the second, you get an average of 15%. At the third, it goes to 18%, and at the fourth, it gets to 24%, 25%. Essentially what this means— and by the way, this was over a 2-week span. Essentially what he's saying is that if you really wanna get above 20% response rate, you need 4 messages in your sequence. Now, Maisha had said 22% at 3, so that's compared to this, Maisha's way above the 18%. But if you wanna get to that 24% and, you know, you wanna compare yourself to what the standards are, That takes 4 sequences. You're probably gonna ask me to repeat that, so you know what I'm gonna do? I'm just gonna put that here in the slider. There we go. Hopefully that helps. Pretty interesting, right? All right, so, uh, then we had Kat Kibben. Kat was talking about the human touch and crafting authentic job postings. And let me tell you, she knows her job postings like I know sourcing. Brilliant. And she gave one of the most complete thorough, detailed, and actionable presentations on any topic I've ever seen. But on job postings, it was the most actionable and authentic content on job postings and, and getting them to increase your candidate quality. So essentially focusing on quality, not focusing on quantity. A couple of big takeaways: I encourage you to go check out her website. There's actually this gender decoder that she provided that you can copy and paste the job into. And when you do that, it will, it'll like pull out any gender bias that might be on there, which is pretty cool. It worked actually. So let me put that up here as well on the slider. Might be a little bit harder to write down, but see if you can copy that down. And as I, as that scrolls by, I'll tell you a little bit about what she did. So she talked about a couple of really important things, like for example, don't make your job posting too long. It needs to be brief. In fact, she specifically said the first and second paragraph, which are where you should pitch the job, should be about 250 words or less. So that's, that's brief. That's the first and second paragraph. Then after that, you want to have a section of required experiences. And don't get creative, she said. People get too creative. You don't want to get too creative with this. You want to be very, very, very clear, straight to the point, and say, here's what we're looking for you to do, here's the experiences that we expect you to have, right? Like, be very bold about that. And then after that, you have the About Us, which should be boilerplate so you can be consistent. And the legalese, which should be boilerplate and consistent and come from an attorney. And those two things as a template will help you avoid introducing any bias in the future because you only are going to be writing the job pitch that's specific about this job and the required experiences. And that's where the trick comes in on making those experiences be, you know, inclusive, right? So that's the four sections. And that's brief. That's brief. Another thing she said is don't guess on job titles. If you try to get fancy with job titles and get creative and, you know, try to put like, I remember technical Yahoo over at Yahoo was the job title for software engineers. It's going to be hard for people to apply because if people can't relate to that job title, they don't recognize it, what's going to happen is people that aren't qualified will apply because they don't know what it is. And people that that are qualified aren't going to know what it is, so they won't apply. In other words, you'll have a bunch of unqualified applicants because they don't know what you're talking about. So simplify it and don't guess. I think that's a pretty good piece of advice, all in all. So anyway, um, there's a couple of other sections I would really encourage you to— oh, actually, you know what, I think this is a better URL here. Let me edit this. Might be easier. There, that's probably better. Yeah, check that out. Page.co, capital K, lowercase j, capital Z, capital Z, lowercase z. Oh, I get it, she's trying to say job postings are boring. Probably not, she didn't mean that. Anyways, check that out. The last person, um, before we did the Q&A and wrap up and all that was Amit Bhatia. And Amit actually kind of gave me, um, a bit of a stroke. He said that 1 in 3 RECs fail. I don't know where he got that information. I wish I'd challenged him on that because he said 1 in 3 regs fail, which means— what does that, what does that mean exactly? That 1 in 3 jobs never get filled, or, or what? Are they— the job gets canceled? Is that what they— I mean, does, does fail mean that they never hired anybody for that job? So what happened? Did they not need anybody? It was kind of really a weird number. It was a shocking number. I wish I knew a little bit more about that, but it was shocking that 1 in 3 regs fail. And he said that this was based on data that they had gathered from job postings, um, over a period of time. And so they're seeing the same job posting and over and over and over and over again, which indicates it's the same job that never get— got filled, and then it just gets taken down, um, for whatever reason. So anyway, that was a little shocking. So hope you enjoyed my, my recap. Don't forget that we have the course available now at TSI University. So it's the ChatGPT course. There's a URL for that, on sale for $49 at tsiuniversity.com, which has a whole bunch of other courses you might be interested in. And act now on that ChatGPT course because normal price is $129. I'm going to be raising that price after the weekend. So hope you enjoyed the recap. If you Did, then please send thank yous to HRTX and Recruiting Daily. If you did not, please send your complaints directly to me. All right, thanks everybody. Have a great weekend and be kind to each other.