Latest / Recruiting Evolution / FDE+ | From Prompting to Programming: Elevating Recruiting in the Age of AI with Mike Wolford, CEO of Lex Duo

Transcript

FDE+ | From Prompting to Programming: Elevating Recruiting in the Age of AI with Mike Wolford, CEO of Lex Duo · Recruiting Evolution

• Mike Wolford [00:00:00]: There's a quote by Albert Einstein we're going to hear a lot of in the age of AI and that is imagination is more important than knowledge. So just remember that imagination and being creative is now your advantage. Use it. A lot of us have come to recruiting from other fields, but for the moment, it looks like English lit majors are going to inherit the earth because they'll be able to get the best outputs from the system because they know what to ask it for. But this is what's going to make you stand out from the crowd that writes a million emails that hopes it finds you well, and follow ups that ask if you were abducted. Kortney Harmon [00:00:39]: Hey guys. Kortney Harmon, host of fde. We're bringing you a special series of episodes called fde. Those are going to be highlights from our recent virtual conference where hundreds of you joined us for an incredible event focused on boosting revenue for 2025. Each of these sess is packed with valuable insights, expert discussions and actionable strategies to help you drive growth in your business. Whether you miss the live event or want to revisit each session, we've got you covered. We're going to drop each of these 10 live events to wrap up our year and kick off the new year. Right? So let's dive into today's session and uncover the key takeaways that will propel your success in the coming year. Kortney Harmon [00:01:28]: Stay tuned and let's dive in. We are so excited to have you back here for a day two of our second half day virtual conference where we're deep diving into AI and how it's completely transforming the recruiting game. This isn't just another high level webinar where you go and you don't walk away with something to do. I am so excited for today's first session. We had an amazing lineup of speakers yesterday from Ben Mena, Matt Strain, Greg Bandida, Chris Hessen. And today the fun continues. Today we're going to have Mike Wolford. He is amazing. Kortney Harmon [00:02:10]: I have gotten to see some of his fun that he's already put in his PowerPoint, but he is the author of nearly 100 published articles, four books. He speaks around the world about recruiting and the impact of AI in our industry and he's been in the seat where you have been. He's been a recruiter, a sourcer, a recruiting manager, a sourcing manager, talent, intelligence, professional. So many things across the board in our industry. So Michael, I am so excited to hand it over to you for our audience as we continue to talk more about AI today. Mike Wolford [00:02:42]: Good morning, good afternoon, Good evening. Depending on where you're hearing me. For those of you who don't know me, my name is Mike Wolford. I'm the CEO of Lex Duo. Our job is to train recruiters on how to use AI for their day to day work. We also build AI and agentic AI powered tools. My background is in recruiting. Like it was mentioned, I've been in the industry almost 20 years now. Mike Wolford [00:03:05]: I started an agency, I was an agency recruiter. I've been a sourcer, been a recruiting manager, I've been a sourcing manager, I've been a talent intelligence professional. I'm closing in on just over a hundred articles published in various websites over the years and I just published my fourth book. I've worked at npr, Twitter, Allstate and Capital One. I speak around the world about this topic of how AI is going to impact our industry. Just for fun. This is supposed to be a picture of me. It took the AI, took the two pictures of me on the bottom right and created the picture on the left. Mike Wolford [00:03:39]: I did ask for the purple squirrel and me to be looking at each other, but such as the state of AI today. And here we go, off to the races. For some reason, Lex Duo is a company. For those of you who haven't heard of us, we're actually owned by talent acquisition professionals and that makes us unique among companies in our space. We are a company who are full of professionals who know what you've been through, what you're going through. We have everyone from a staffing agency owner to HR business partner. We have recruiters and sourcers all on our team and that gives me an additional advantage because I hear from them what's going on in the market and what do recruiters and sourcers need today for the agenda, I'm just going to really quickly talk about AI and the three levels of AI. So we're going to kind of talk about what AI is and then I'm going to go through step by step instructions to building some basic GPTs so that you can walk away with potentially some tools to make your life easier. Mike Wolford [00:04:40]: We're going to show also if there's time, I'm going to show you how to create your own interactive avatar. And if there's not time, then I'll show you where to go to create your own interactive avatar. Full disclosure, I am going to be demonstrating on a couple of tools that I did pay to use. I know a lot of us are on a budget. Hopefully $20 a month is affordable because that's really the entry point into AI as far as pricing goes. So, yeah, that's what I'm going to talk about. Before we dive into LLMs too much and AI, I do want to say basically, as we're talking just so eye level set here, I don't know what other speakers do, but in my opinion there's basically three different levels of using AI at this point. There's level one, which is the most common, which is prompting. Mike Wolford [00:05:28]: A lot of people are on GPT and they're learning prompt engineering. That is a skill that was valid and relevant two years ago. Today we are past prompting. Now we're into programming. That's level two of AI programming. Sounds intimidating and it can be, and it has been in the past because that required you to speak a programming language like Python or Java, and most of us don't know those languages. However, given the advances in AI, the new programming language is whatever language you happen to speak when you're programming AI, you can do so almost exclusively these days without needing to code. So what you need to do to program basic AIs and GPTs is basically understand how to document a step by step. Mike Wolford [00:06:17]: So step one, step two, step three, step four, and then we move into true AI agents and workflows. And that's level three. And there are very few of us in the profession today who are at level three. Probably 90% are at level one, maybe 8% are at level two. And there's very, very small 1 or 2% of our industry is finally at level three where they're able to actually build AI workflows and AI agents. There's a video here I'm not going to share because we're short on time, but when they share out the projects. If you want to watch this, this is a very interesting and educational video. If you're looking for somebody to follow who's out of our field, who is just an expert on AI and LLMs in general, I recommend Dr. Mike Wolford [00:07:06]: Thompson. He's the person I started following first as an expert. His background is unique. As far as I understand, he has conducted over 50 interviews with GPT. This is how I became aware of it. He recorded himself asking it questions and then he animated the answers from the AI so it looked like it was a question and answer session like a normal interview. His area of expertise is human intelligence. He was the head of Mensa International for a long time and he worked with extremely gifted students. Mike Wolford [00:07:40]: Over the last few years he has focused on artificial intelligence. His website's called TheMemo AI or Memo AI. And it is a really good overview of all of the different models from all over the world. How big are they? How powerful? What kind of tests have they been put through? How did they score? How did they compare against each other? So if you're looking for somebody who can just give you an overview of AI and what it's doing in the market, I recommend following Dr. Thompson. You can see there, he's got a jar in his hand. And in that jar are colored crystals. Those are jello crystals. Mike Wolford [00:08:17]: And each one of them is meant to represent a different type of data. So the blue. Each one of those little blue crystals, which are finer than grains of sand, each one of those little blue crystals represents a website. So that's the general Internet, which has been indexed, which I used to remember joking as a kid, there would be commercials that you found. The end of the Internet, like that was supposed to be impossible to do, but not anymore. We can actually index the entire Internet and all the information or websites on it. So that's where the blue information comes from. The orange is curated web. Mike Wolford [00:08:52]: So Facebook, Instagram, X Reddit, these sort of things where content has been voted on by humans. It's a little bit different than the content you're going to find on the general web. So that's orange. The purple is the next above it. And that is all of the books and articles. So it's everything from Agatha Christie to the American Journal of Medicine, right? Scientific articles, fiction books, nonfiction books. Each one of those little crystals represents one of those. Then on the green on the top is Wikipedia. Mike Wolford [00:09:27]: Each one of those is a Wikipedia page. And I know people have mixed feelings about that, but actually it's become a really good source of data. I dare anybody, for fun, after this call to go to Wikipedia, update any random page you would like with factually incorrect information, and then set the clock till how long the community eats you for breakfast. There's a lot of fact checking that goes on in Wikipedia. So it's actually a really good source. Now what the AI does is once it has all that information, it starts to read. And what it's doing is not reading for comprehension, it's reading for connection. What I mean by that is I took this second bullet point here and said, former head of Mensa, and I picked the word of. Mike Wolford [00:10:12]: What the AI is going to do in an LLM is it's going to say, what's the word before of its head? What's the word after Mensa? What's two words before former and then international. Right. And it will map out the connection between words. So that's what GPT3 originally was. Simply, it gave the AI all the data and then it had IT run for 250 plus years of compute time across 1000 computers. And it mapped out the connections between the words and all this data. So the books aren't there anymore, the articles aren't there anymore. Only the thing is the words and what words they're connected to in kind of a 3D vector map. Mike Wolford [00:11:01]: So when you type your question or you speak your question, the AI takes the words that you've put in and it says it maps how connected those are to other words. And then based on that, it tries to predict the most likely next word that is a base level. What the technology is, how that works and why it works is still a mystery. We're something like cavemen who've discovered how to make fire, but know nothing of organic chemistry or releasing chemical bonds, right? We do not understand how it actually works. We know how to make it and we know it's useful. That's kind of where we are with AI. So in that sense, it's kind of like digital fire. What can it do for you? Actually, quite a lot. Mike Wolford [00:11:48]: I know a lot of things have been hyped and people have talked about AI and its capabilities and can it do this and can't it do that? It's actually quite capable. For those of you who would like to understand the volume that you're dealing with, if you take everything that humanity has ever written in our sum total of existence, not only has AI written more, AI has written about 10 to 15 times more since its inception in 2022 or 23. So it is a phenomenal writer. 90% of everything in human history that has been written has been written by an AI. I find it personally a little bit funny when people want me to build them something that detects AI writing. I think it would be easier to build something that detects human writing, given the volume. But I also want you to understand that for context, if you're hitting your AI and it's saying, I hope this message finds you well, you're going to lose, because there's 10 million messages like that now, because that's how every email starts. And also it helps you to understand why the AI does what it does. Mike Wolford [00:13:04]: And to me, it's one of the great ironies in the technology is that it's designed to have all of this information about writing and really everything all of humanity has ever written. And yet by design, it will be the most average writer humanity has ever produced. So unless you tweak it and give it very specific instructions, if you just let it write, understand that it's the most average writer in the universe and ask yourself if that's the tool that you want writing for you. If not, there are ways to modify its writing style and how it does that, but out of the box, this is the default. Also, it helps calm kind of the fears I think people have about AI trying to copy itself or take over the world. It does not have any ambition. It doesn't have emotion. It is a very unusual thing for us as humans to deal with. Mike Wolford [00:13:56]: Every time we deal with something that has intellect or intelligence, it also comes with emotions, right? Humans are intelligent and emotional. Our dogs and cats, they're intelligent, but also emotional. AI is intelligent but not emotional. And that is a very hard thing for us to wrap our heads around. But when you type in words, it runs a process, tries to figure out what the most likely next words are to respond to you. And then it turns off. It doesn't wonder what you're doing. It doesn't think about taking over the world. Mike Wolford [00:14:28]: It runs what you asked it to run. Once it completes this task, it shuts itself off. And there's a lot of reasons for that. But if you can believe nothing else, believe that nobody wants to pay for the AI to be running when it's not doing anything. So I mentioned it's a great writer. I am adhd. Full disclosure, it takes me two to four weeks on average to write an article. When I got a hold of GPT4 back in April of 2023, I was able to write an entire book on recruiting in 1617 days. Mike Wolford [00:15:00]: So from blank page to sending it to my publisher in a record amount of time for me, what normally would take me to write a article? It's nine chapters, 74,000 words. It has all of the prompts that I use to create the content in the book. So it's not just, hey, AI can write a job description for you. Here's a few sample prompts. Here's a prompt cheat sheet. You can also create one of these on your own. It's a nice little bonus if you get stuck doing something, tell it, hey, I am a recruiter. I'm trying to write a, whatever, can you explain to me what my options are in table format, one in column A, give me my options and B, give me the variations and C, give me an example and D, give me a sample prompt and E, give me a sample outcome. Mike Wolford [00:15:53]: Right. Any type of writing that you want to have is now available to you. So AI is incredibly fast and it allows you to create content, good content at speed. Just to make sure it wasn't a fluke, I did it again and I wrote a second book called the AI Analyst. And my conclusion in dealing with this AI and writing books with it and creating content with it is that this is the New Gutenberg Press. It not only can print books, it can write them. And this is novel in human history. The AI recruiter was designed for recruiting tasks. Mike Wolford [00:16:31]: The AI analyst is designed to teach recruiters and sourcers how to use AI for analytics. So as I mentioned, I do have about two years of experience in talent intelligence. I do believe that as a profession AI gets all the attention. But I have said for a long time we are not data driven enough. As a profession, AI gives us the ability to actually build our own analyst. So we don't normally have access to a data scientist or data analyst. But now with AI we can build one who can do the types of analytics we need for our work. So if you want to talk about labor market conditions or you want a competitors report or you want a skills report, that's all talent intelligence. Mike Wolford [00:17:12]: In this book I used AI to write a guide on how to use AI for these tasks. Over time it has moved incredibly quickly. I know we feel like we're on burnout when we say AI today. We're not talking about the same thing we were talking about two and a half years ago. I did an interview in December of 22 with GPT3 about the future of AI and talent acquisition. It was interesting, it was novel. Kind of felt like I was writing to Tom Riddle's diary, for those of you who understand that reference, you type something in, it disappears and then magically an answer comes back. It was novel, it was interesting, it was engaging, it wasn't necessarily earth shattering and it could kind of be rude to you. Mike Wolford [00:17:57]: Sometimes they tell you to take a long walk off a short pier. In the last two and a half years, AI has evolved considerably. And now we're to the point in the last few months where something called agentic AI has really come into the scene. And agentic AI is two things. Agentic AI, the most popular version of it is creating an agent that can do a task. And that is the manifestation that gets the most attention in our business and honestly drives the most fear. Because we're afraid that you're going to say go recruit. This position of the AI agent is going to go be able to do that. Mike Wolford [00:18:35]: We're not quite there yet, but we are to a point where agentic AI in that form can be told what the outcome you want is given a brain, which is basically selecting the LLM model you want it to use, giving it tools and then giving it an outcome and letting the system figure out, based on my tools and based on my instructions, how do I get to this outcome that is going to become the most popular way to use agents in front of the scenes, behind the scenes. Where the real impact on the economy is going to come in in our industry is agentic AI is a software engineer. It can build product. So if you're a SaaS company and think of Amazon and Google and Microsoft, a lot of Microsoft products are LinkedIn, these are SaaS products, right? You go to the website, you log in and then you use the system that's software as a service. That model is going to be replaced. And the reason why is companies like mine are going to be able to build a version of LinkedIn, a version of Career Builder, a version of a CRM for about 10% the cost and about 5% the development time. This is the huge difference and the real impact that agentic AI is going to make. Agentic AI is the world's most accessible software engineer. Mike Wolford [00:20:05]: I don't want to say best because that's subjective, but it is a very good software engineer. You're going to hear the term vibe coding a lot and this is what people are using. They're using the agentic AI to help them code, but ultimately it's the AI that's writing the code and not the human. And that's the point that we're getting to. And so this is also a cautionary tale for us. If you recruit in tech, you've seen this firsthand. If you have not, it's coming to an industry near you. Agentic AI is now fully capable of building very complex products and very complicated software engineering in a fraction of the time and in a fraction of the cost. Mike Wolford [00:20:47]: So if you want to know what the new companies are going to be, Google may be Google, they may stay Google, they may look like Google, but they're no longer going to be SaaS company. They're going to be an agentic AI company. And the big fang companies have already started to let their higher level developers walk out the door without trying to stop them. And that's because an agentic AI can replace a software engineer. And so, spoiler alert, if it can replace a software engineer, it can also replace A recruiter. A word of caution Before I dive into actually teaching you how to build some second level AI GPTs, I do want to have a word of caution here. As we know most of us on this call Workday is currently being sued in a class action lawsuit for age discrimination for their AI. Now, I have no idea on whether that's true or not. Mike Wolford [00:21:39]: I'm not in position to say. However, my best guess based on my experience will be that Workday will settle and that will open the floodgate to every ATS vendor being sued. So this is definitely something that we should be thinking about and talking about. And a word of caution. Just because you can doesn't mean you should. Right? An example of what I mean by this. As we discussed before, AI is looking for the connection between things. So if you're doing matching and you give the AI a candidate name, it's going to try and associate that name with an outcome. Mike Wolford [00:22:19]: Not because it's evil or it's biased, because that's the way it works. It's trying to map out the connections between pieces of data so that it can give you the best answer. So remember, just because you can doesn't mean you should. When you're creating these tools, think about it from the candidate's perspective. Think about it from the hiring team's perspective. Think about it from a legal perspective before you deploy these tools because you can get yourself into trouble. They are amazing and very powerful if used correctly, but they can get you into trouble just as quickly if you're not careful and purposeful about how you use them. Okay, so first things for those of you following along at home, I am using ChatGPT, the paid version, in order to get this here. Mike Wolford [00:23:11]: But before we dive in, I do want to just show you three things that I do when I'm building my tools and my agents. The first thing I do is I teach it how to write. I give it a set of general writing rules. As we discussed before, it's a very average writer out of the box and that's not helpful. So tell it, show, don't tell, Use strong verbs, be specific. Right? Avoid hyperbole, stick to facts. Whatever your rules of writing are, jot them down or use your AI to come up with them. This is really going to be part one of programming your AI. Mike Wolford [00:23:47]: The other thing, like I mentioned, if you get stuck, ask it to create a table for you and explain what your options are. When anything's an option and you're staring at a blank screen, it can be kind of Overwhelming. So ask the AI. And this is what I did here. I said, hey, I'm trying to create some writing directions, but I really don't know what my options are. Can you create a table for me? Tell me what my options are, the variations, explain it to me like I'm five. Give me an example prompt and what I can expect as an output. And here we go. Mike Wolford [00:24:17]: And now when I'm programming it, I can say, all right, I want the tone to be enthusiastic in the voice of Yoda in iambic pentameter. Right now, the beautiful thing about this is I don't actually have to write that. I don't have to be creative enough to actually sit down and write that email. There's a quote by Albert Einstein we're going to hear a lot of in the age of AI, and that is, imagination is more important than knowledge. So just remember that imagination and being creative is now your advantage. Use it. A lot of us have come to recruiting from other fields, but for the moment, it looks like English lit majors are going to inherit the earth because they'll be able to get the best outputs from the system because they know what to ask it for. Most people don't know even what iambic pentameter is, much less to ask the AI to write an iambic contaminator. Mike Wolford [00:25:13]: But this is what's going to make you stand out from the crowd that writes a million emails, that hopes it finds you well, and follow ups that ask if you were abducted by Part two is actually giving it the directions. So the difference between prompting and programming is really just one layer. In prompting, you're saying, I want you to do this this way. In programming, you're saying, I want you to write emails, but here's how I want you to go about doing it. First, I want you to ask who the email is going to. Then I want you to ask the user what's the subject? And then I want you to ask the user for any special instructions. Then I'm going to give it the instructions like I would in a prompt. Professional voice, enthusiastic tone, make it like a Shakespearean sonnet, make it like a country song. Mike Wolford [00:26:01]: Whatever else I want, I could put in those directions after I've asked the user for input. So instead of having to prompt every single time and remember all the options, effectively programming allows you to save your prompts and move to what a push button functionality should be. Instead of saying, I want to write an email and writing out this prompt, you push a button, says, I want to write an Email. It's already programmed with all the prompt information. It just asks you who is it going to, what's the topic and do you have anything special you'd like it to do for that? That is where you start to get the real efficiency gains from AI. When you programmed it to be push button for the things that you do on a regular basis and you've pre trained the AI on how to you want that task accomplished in a step by step direction. Now there's another really interesting thing about AI that kind of gets overlooked and that is you can give it examples. So if you go on YouTube and you try and learn this yourself, you're going to find a lot of software engineers who are programming in Python and then letting it run. Mike Wolford [00:27:05]: Very few of them actually get to the point where you can give it examples. And you can do that in two ways. You can give it to it in the prompting screen where you're programming it. An example of that would be when I was recruiter, somebody would reply back to me and say, hey, thanks for your note. Now it's not really a good time. I would reply, thank you for your note. Forgive me for not introducing myself fully the first time. My middle name is Opportunity. Mike Wolford [00:27:30]: Do you have 15 minutes to talk to me? And that would be an example of a good reply. So you can do that, you can build that right in, or you can actually program it and upload it on the back end. So again, when you're programming, the thing that's going to be different is you're actually going to be telling it how to complete this task versus prompting. You're using very specific things. So what you're going to find when we get into AI is you're going to come into the edit screen and this is going to be blank. And so this is the AI recruiter. It's available on the GPT store. It's totally free to use. Mike Wolford [00:28:11]: So for those of you who are looking for a free tool to walk away today with, this is 1 of 2 I'm going to give you. But when you're building your own, you simply name it, you give it a description. The first part is where I showed you before, where these are the writing directions and I tell it what Persona it's going to have these act like a recruiter and then I tell it what it's going to be doing. And then here are my rules, right? Also, just a little tip for those of you who decide to do this on your own, if it refuses to obey your instructions, this is a Little nuance I learned from some software engineers, right? If I hope this finds you well, replace it with greetings. Now this lets me know that it's not doing what I want it to do, but it doesn't let the user know. And then you put the previous task as being mandatory. So if, when you're writing it, get it to write your emails, it keeps doing this. This is a workaround to get it to actually follow your directions if it isn't. Mike Wolford [00:29:07]: And sometimes no matter how you prompt it, program it, it's never going to. So when you're asked to write an email, follow the directions below. Step one, ask the user for who the email is going to. Step two, this is the step by step instructions. So the new skill to me is being like a business analyst, right? You actually have to describe your requirements and then tell it exactly what you want to do. Now I have this push button functionality over here I had the conversation starters, but down here there's a little power upload files. And this is where I've uploaded my book, the AI recruiter. Now, the thing is, AI is so fast, when it is asked to do a task, it's going to reread the entire book before it does. Mike Wolford [00:29:53]: And this is the other part of training your AI. You can upload your knowledge into the back end here. This is where you would upload. If you've written articles or you've written books or you've got emails, you can upload this to the back end so that it gets your voice and then it gets your tone. You also have options as to let it search the web, generate images. Down here is the other little button that is magic. So you. Most people skip this, they don't even make it this far down. Mike Wolford [00:30:23]: If you create new action and you go here, some of you are gasping. I can hear on the other side, this is an API key. What this means for you as a recruiter is potentially you have access to your ATS and CRM, all of the candidates for the very first time. If you as a recruiter search your database. And I'm just going from tech knowledge because I was a tech recruiter, but if I have 10,000 Java developers in my database and my system limits me to 500 search results, I'm never actually going to be able to see more than 600, 700 of those resumes because the computer is going to rank the results basically in the same way based on the keywords, right? There's very little I'm going to be able to do in order to see all 10,000 resumes. However, if I teach my AI recruiter how I read resumes, right, start at the bottom, read the education, look at their first job, read from bottom to the top for their jobs, see if they've made career progression, right, Read the bullet points, see what they've been responsible for. Have they been responsible for budget. You can step by step explain to an AI how to read a resume. Mike Wolford [00:31:29]: Once you do that, you can connect that AI to your ATS or your CRM with an API key. And for the very first time, you're going to have access to every candidate in the database, because this is not limited by your search results. This can actually go in and look at every single resume in your database and then do it in 2, 3 seconds. And so you're going to bring back results based on your criteria that is fundamentally different than what we have the option to do today. So if you're looking to get the full power and ability from your CRM and ATS systems that you have implemented, API key, ask for it at work, see if you're allowed to use it. GPT, the professional version, the one you pay 20 bucks a month for. Now, OpenAI says they don't train on that data, which is normally the concern your security teams are going to have. But I ask you a question. Mike Wolford [00:32:21]: With these new AI agencies and agencies being built with AI from the ground first, which recruiter is going to win? The recruiter who's manually searching the database and manually searching LinkedIn or the recruiter who's programmed their AI the way that they search? And it's going to search LinkedIn for them, and it's going to search the ATS for them, and it's going to search the CRM for them consistently over time. Which recruiter do you think is going to perform better? I leave that for you to decide. Oauth, for those of you who aren't familiar, this is your email and your calendar. So also, if you want to start scheduling, handling emails, outbound emails and inbound emails, this is how you would hook up your AI to those systems. Again, this is incredibly powerful. You're talking about a tool that's 20 bucks a month, and it not only can help you write emails faster and job descriptions of candidate summaries, it can also hook in to your API and help you search and help you in your mail and manage your time for 20 bucks a month. I can't imagine something providing more value. So this is level two. Mike Wolford [00:33:29]: Level one is prompting. Level two is creating your GPTs and effectively saving your prompts. Also, just a word here. Since I'm among my fellow recruiters. Don't just automate, elevate. You are empowered for the first time ever to do recruiting the way you specifically think it should be done in your specific job, in your specific instance. Think about the best way that it could be done, not just how to automate it. And let me give you an example here from my own examples. Mike Wolford [00:33:58]: And again, you guys should get the deck after this. I hope so. No need to take notes here, but in my candidate summaries, one of the things I tell my AI to do is the standard stuff about pitching a candidate. However, I also, as you see step four here, at the end of it, I have the AI create a table and in column A are all of the job requirements must haves. In column B is the candidate experience that matches based on the resume and my interview notes. And then column C is the percent match. So you can use this not only to speed up and automate your candidate pitch process, which you absolutely should be doing, but you can actually make it better than it was before. Because if you send a candidate to a hiring team and they're an 80% match or better based on your ability to filter at this point in the process, they're going to have to have a really good reason to say no to you. Mike Wolford [00:34:51]: Especially in the US and Europe when there's very specific laws about rejecting candidates. So not only can it make you faster, it can improve your submit to interview ratio. And this is where AI can really shine. It not only can make you faster, it can make you technically and performance wise better. Because everybody would love to be able to do that. But who, who has the time? Especially if you're trying to get five or six submits out the door and an offer and schedule interviews and follow up with the hiring managers for feedback, right? We just don't have the time. But if you program it from the very beginning to say this is how I want this task done, then you've elevated your game and automated it at the same time. It's a double win. Mike Wolford [00:35:38]: And so for those of you who are on the phone or on the call watching, my message to you is, you need to learn how to use this so you can become efficient enough to stay in recruiting. If you don't want to be a recruiter, now's a good time to find the door. There's nobody who has five years of experience in AI, right? There's tons of opportunities in automating things and building Products in teaching. So if recruiting is not for you, this also might be a good time to exit the profession and find something else to do. I know we normally don't like talking about that, but real, real, over the next several years, there's going to be fewer jobs in recruiting, which means some of us are going to leave voluntarily and some of us are not. But the future is fewer jobs, not more jobs. There will be fewer recruiters. Now the upside to that is the recruiters who stay. Mike Wolford [00:36:32]: I can't speak for you, but for me as a recruiter, my favorite part was getting on the phone at the end of the process, calling my candidate and just saying, congratulations, you got the job. That was the best part of it for me. And so if you're a recruiter three years from now, four years from now, and you're still in this field, the upside is you're going to be extending a lot more offers than you do today because your process is going to be so much more efficient. You're going to get to spend more time with your candidates and your hiring teams, less time doing paperwork and more time saying congratulations. So I get it, it's a trade off. There'll be fewer of us, but we'll be more highly skilled, more highly paid, and we get to talk and interact with candidates more. So it's a flip. Now this is level two. Mike Wolford [00:37:17]: Now we go into level three. This is NAN for those of you who haven't seen or heard of it. And this is where the technology is today and why it's so powerful. As you can see, there's no programming here. All of this is click and drag and drop and drive. I have been working with N8N for the past couple of weeks now. For me, the hardest part has been getting my logins for the different tools to work. But on your left here, this is what an AI agent is going to look like on paper. Mike Wolford [00:37:47]: You're going to have an agent from whatever system you're building. You're going to connect it to a brain, you're going to connect it to resources, and you're going to give it a task. And so this is a trigger, this is what's going to start the agent. The agent is going to start whatever brain LLM model you're using is going to be over here. And then whatever tool you're giving it, these are the tools over here. And this is the memory. So you give it a brain, give it a memory, you give it tools, and you give it a task to accomplish. And it will take this and create an output independently. Mike Wolford [00:38:23]: And so that is incredibly powerful. AI agents are going to outnumber humans in the not too distant future because they're very easy to create and they're very powerful. Now I advise, generally speaking, for recruiters not to use agents because when it comes to things like hiring, you have to be able to explain how and why you made a decision. And if you're using an agent for that, it's going to be very difficult to explain. AI workflows, however, are something that I think everyone on this call eventually is going to have to learn and master. This is going to become as part of our skills as Boolean. This is going to be synonymous with us. And this is a workflow. Mike Wolford [00:39:05]: And this is very simply, here's where you start. I want you to do these things, take these actions with these tools and then give me the result. So, so it's a little bit different. Whereas an agent, you're telling it what result you want to come up with and you're letting it go. Here you're designing a process. So for those of you who are an HR tech or workflow process designers, this is your golden age. This is really a time to step up and learn a tool like lovable.dev or N8N and really figure out how to build these. Because once you can do this, you can literally build anything. Mike Wolford [00:39:41]: You could build yourself a sourcer, you could build yourself a lead gen tool, you could build yourself an automated email follow up, you could build yourself an interview scheduler, a CRM, an ats, anything in recruiting that's software based. If you can master level three of agentic AI and workflows, there's literally nothing in the world that you can't build. And that's really empowering because most of us are not software engineers yet in the next two to three years, a significant portion of us are going to be able to do this blindfolded. So this is where the technology is going and this is what it looks like. I'm going to stop right here for questions before I go into animating our avatar. Do we have any questions? Kortney Harmon [00:40:25]: There was a question so far. The question he asked, what do we need to connect our CRM with? How do we connect your CRRM to our ChatGPT? Mike Wolford [00:40:34]: Yeah, so that's going to be via API. So if you go to your software team and say I need an API key for our ATS or our CRM, they're going to know what that is and what that means. They're going to ask you why you want it and what you want to use it for. And then you're going to have to start the discussion with them, right? Like hey, I want to use it to sort resumes or I want to use it so that I can have access to every candidate in the database. And I'm going to use this paid program over here that doesn't train on our data. So that's how you would do it. Like so if you go actually to chatgpt.com you're going to see here that's going to ask you down here you can actually just go to GPTs and there's actually a store here. But create your own and then you configure and then this is where we fill out all the information we showed before sneak down here to create new action, the little gear. Mike Wolford [00:41:29]: And this is your API key. And this is just going to be a long set of letters and numbers and that's just going to allow this system to talk to the data in that system. But that's how you'd actually hook it up. Good question. Any other questions? Kortney Harmon [00:41:47]: That is it so far. I don't see more, but continue to write them in the chat. Mike Wolford [00:41:53]: We're coming down the home stretch here. The other thing to get done and there's two systems I use. I'm happy to put those in the reports. I've noticed that I have a hard time getting my hiring teams to actually read reports, especially before the call. So there are two tools you can use. There's synesthesia, which is this one, and there's Heygin. They both generate avatars. And so what I did was I went to our talent intelligence agent in GPT and I asked it, I gave it some basic data and then I asked it to create a labor market report for me. Mike Wolford [00:42:35]: And after it did that I had it thin, right? So I gave it some basic data. I asked it to create a labor market report based on the data and its ability to access information on the web. So I also had it read, for example, the BLS Jolts report. It gave me this really nice report, right, that nobody's going to read. And so what I did is now can I turn this into a nightly news report? And now here it is. And now I just copy and paste those directions into a system like heygen or Synesthesia. And I create this as a knowledge base. And then all of a sudden the AI has all of that information in it and it can create and give a labor market report to my hiring team. Mike Wolford [00:43:24]: So I send them the Link and it says, you know, good evening, Dave. Here's your labor market report. Here are the top three companies that are hiring in our space. Here are the skills, here's the pay, right, that they actually listen to. So if you want to get the attention of people in your industry and you want to come across as the expert in AI, I recommend paying the 20, 30 bucks it is to get one of these tools. Take the Talent Intelligence agent again, it's free, it's on the GPT store. Anybody here can use it. It is programmed by me and it does have the Talent Intelligence book I've written on the back end, for example. Mike Wolford [00:44:02]: So totally free to use. Get it to create the labor market report for you. Get it to create it as a nightly news report. And then animate that in your avatars, send that to your hiring team, watch their head explode. Then you'll know you're on the right track when they start calling you and saying, okay, when is Skynet taking over? You know, when's the AI going to take over? And you can calmly reassure them because you understand how it works works. The real threat isn't from the AI. It's from people misusing it and getting it to do things that we normally wouldn't want it to do. The danger is the people, not the technology. Mike Wolford [00:44:43]: I coming close to time. I wanted to save some time for questions. What questions do you guys have for me about AI? The future of AI, how is it going to impact our work? What can I answer for you? Kortney Harmon [00:44:57]: I know there has to be questions. I currently just went and got your little talent Intelligence Agent. I was in doing the things while you were talking. I was excited. So thank you, Michael. That's amazing. Mike Wolford [00:45:08]: Absolutely. Let me put those links here in the chat too for people. Kortney Harmon [00:45:12]: The API you have that. I have a message out to our support team. But we have open APIs, so you can easily do that to connect your GPT to your relates. Mike Wolford [00:45:25]: Yeah. And this is the thing, right? Like after class, I watch students go out on their own and they created a rubric, an interview rubric. One of them did. One of them created an AI that does a skills gap analysis for their company. So they just downloaded everybody's resume who's at the company and says, okay, where are we strong and where do we need improvement? Given that this is our industry and this is our business. Right. Boom, done. I had another student come through class, totally unrelated. Mike Wolford [00:45:54]: They get together with all their friends and do Royal Caribbean cruises. That's their thing. And they live all over the country, they have all different schedules. So all that information gets put into the AI along with all the information about the cruises, where they go, right? So they have information about everybody's schedule, what they like to do, kind of in the information section. And then over here they have all the information about the cruises. The AI is like, okay, here's when everybody's available. And this is a good cruise because it has these sort of things and these people like doing that. So this will be good for them. Mike Wolford [00:46:26]: Right. Once you turn it loose, it gets out of recruiting and into the rest of your life because you learn not only is this helpful for your work and day to day stuff, this can be helpful for managing your own affairs and doing things. You know, people use it for stock trading, retirement planning. You could create a code. The interesting thing, one of the takeaways I hope everybody here gets is if you have a book of expertise on any given topic and it's digital, you upload it to the back end of one of these agents or these GPTs and you now have an expert in that field. And so that is incredibly empowering and extremely useful. So once you get started, it's really hard to stop. Kortney Harmon [00:47:06]: Well, I love it. So there's many more questions now. Number one, Tanya wants to know, do you have a link or a website? Because we've already saw people talking about your course offerings, so we need to know more. Where do we go to your bootcamp? Mike Wolford [00:47:17]: So lexduo.net is our website. You can sign up for our bootcamp. The next one will be on the fifth, and then another one toward the middle of our end of September. And that's where we'll learn how to do what I built, what I've showed you. We'll build a recruiter, we'll build a talent intelligence agent, and then we'll build an interactive avatar. But the thing we're really excited about is to announce that we now have an advanced agentic AI class. So we've hired an instructor who has a million downloads on udemy teaching agentic AI and he's going to come teach a class for recruiters on things like N8N and how to actually build your own tools. So this is kind of level two training is what we offer at first, and then the new agentic AI is what I consider level three training. Mike Wolford [00:48:02]: Candidly, the people who make it through the 500, 600,000 of us who have a job in recruiting in 2030 are all going to know this. Just like every recruiter can bool you today, this is going to be the way that we go forward. We're going to be expected to know how to design systems and build systems because that's what the recruiter is going to do. The recruiter is going to be the person who builds the system and then runs the candidates through it. So when something isn't wrong or isn't working right, they're going to turn to us actually, because there's not going to be software engineers in the not too distant future, my little crystal ball prediction. Or there'll be very highly specialized people who are very expensive, which probably won't also be available to us in recruiting because we never have a budget. So we're going to have to be able to build these systems, understand how they work, manage them. That'll be the other half of the job, besides talking to the candidates and the hiring manager. Mike Wolford [00:48:55]: The other job actually managing the process, which to me is great. Because who better to run a recruiting process than somebody who's been a recruiter? Honestly, it's empowering for us in my mind because we're going to not only be the ones who have to do it, we're going to be the only ones who can do it. And honestly, I think recruiters are going to do a better job in general than HR professionals at designing a process because recruiters have to live with it. Recruiters have to take the candidates who aren't happy as a result of the process and we're much closer to it and we're much more highly invested in it. And it's kind of the skill set we're going to need to learn. Kortney Harmon [00:49:32]: Haresh had a really good question. He says, as a CEO, I play different roles. How can I compartmentalize the various roles using ChatGPT, I. E? If you're thinking sales, marketing, recruiting, operations, all of the different areas. Mike Wolford [00:49:46]: So sales is the one that's getting automated first at scale across industries. And that would be an N8N automation you could create. Almost all of sales is going to be done by AI in the not too distant future. And it's simply because of speed. No human can compete with the volume that an AI can get out the door. And an AI can customize things all day long and not get tired, whereas I will go cross eyed after about 30, 45 minutes of that. So this is the number one thing you can do is automate your sales pipeline using an NAN automation. Some of the other things you're talking about automating will be agents that you can create. Mike Wolford [00:50:26]: So you will create an agent that has very specific tasks like an onboarding agent. Well, it's going to have all the benefits information, it'll have the legal requirements uploaded to the back end. It'll have your process and then what you want, the outcome you want is you want the candidate successfully onboarded to the company. That's the outcome. You don't really care how as long as it's, you know, generally speaking, professional and nice. The details are unimportant. So that's where agents become really powerful because it doesn't really matter how I get this task done, I just need to get this task done. That's where agents will come in. Mike Wolford [00:51:01]: So if you're the CEO and you're looking for that, I would definitely look at and you know, take our class, don't take our class. But that's the class you're going to need. You're going to need to learn N8N which is for non software engineers. There are other tools out there, but most of them are still going to require you to program. N8N for example, is drag and drop. You don't ever have to drop single line of code. You can, it's capable of reading it, but you don't need it to build anymore. So that's what I would do. Mike Wolford [00:51:32]: I would learn that particular system, master it and then design my workflows to automate the things I'd like to our sales pipeline, our social media posts. Right. This sort of thing can all be automated now in workflows and then to do tasks like onboarding, you create an agent. Kortney Harmon [00:51:50]: Amazing question. I love it. Melissa has another question. How do you see this impacting internal TA teams versus companies that often work with recruiters? Mike Wolford [00:52:01]: So internal recruiting is going to go the way of software engineering. Sorry, this is what's going to happen. What happened to software engineers is going to happen to us and we're not unique. It's going to happen to doctors, it's going to happen to lawyers, it's going to happen to accountants. Right. We're not unique in that this is going to happen, but this is what the big deal is. And so companies will no longer have a team of software engineers, they'll have a software engineer. Likewise, they won't have a team of recruiters, they'll have a recruiter. Mike Wolford [00:52:30]: Now the upside is if you that corporate, corporate recruiter, you get to say congratulations quite a bit more. But if you don't believe me, you can certainly look at the data. The data is very clear about what's happening to our profession and what's happening to professions around the world and it's fewer, fewer people. Now that said, staffing is going to go through a golden age. Like if I was a recruiter who'd been at this 15, 20 years and I wasn't running Lex Duo, I'd be running a staffing firm because staffing is great in times of uncertainty. Also, staffing is where the technological adaptation is going to happen first in our industry. But again, look at Meta, right? They had 3,300 recruiters, now they have 800. They're not going to ever go back to 3,300. Mike Wolford [00:53:13]: That's never going to happen. Amazon stopped hiring recruiters for a year. They just didn't hire any. So this is what's going to happen. That's what's happening at corporate already. This is going to continue until they get to a baseline where the 300 recruiters at Meta are at full capacity right there, but there's only going to be 300 of them. So corporate will shrink and a lot of places be replaced with an rpo. You can look for companies like IBM and Deloitte to come out with their own service offerings. Mike Wolford [00:53:43]: They're hurting too, consulting businesses. McKinsey is shaking in its boots and McKinsey has been around for forever. They're well respected analysts all over the world and they're afraid that their analysts are going to be replaced by AI. So they're looking for new streams of revenue. And there's an old saying, nobody got fired for hiring IBM, right? So if IBM comes to your company's door and they talk to your chro, they're going to say, we have an AI recruiter, an AI sourcer, an AI scheduler, an AI phone screener, right? And if you don't already have that or you don't have a roadmap to get that into your company, they're going to hire IBM and fire you. I hate to be the bearer of bad news here, but this is just how businesses work. If they can do something 80% as well, for 20% the cost, bet your bottom dollar that the company's going to do it. I know because I worked at Twitter and they were very bring yourself to work, right? Open culture. Mike Wolford [00:54:36]: The second money was on the table, somebody tried to buy us fiduciary responsibility, right? Like all of a sudden they learn these terms. If you're in a for profit company, understand that that's what it is. And again, this is our opportunity to inoculate ourselves. We have the chance to go, you don't need I.B.M. you got me. And I'm going to show you how to build a new process based on AI from the ground up that works for us, not based on what the engineers at IBM think. And that is actually a really powerful way to save your job. So I get that it's frightening and I get that it's a major change. Mike Wolford [00:55:09]: But, but for those of us who are interested and engaged, it's also the biggest opportunity we've ever had. Because if you are one of those corporate recruiters in one of these big companies, you will be able to command 300, $400,000 salary because you're able to deliver hundreds of hires and that's actually valuable. So yeah, it's going to be tightening, it's going to get harder, it's going to become more technical as a profession. But for those of you who endure and flip through the other side, to me, I have always hated feeling like an order taker on the other side of this, there's none of that. We are true, dyed in the wool professionals who have a set of skills that are incredibly valuable to a company. And to me that's actually a nicer future for us. Even though there'll be less of us, we'll have more security, more pay and more options. Kortney Harmon [00:55:59]: Part two to that question, as AI matures, what would compel a client in the future to work with us as a vendor opposed to just using some AI agent that they can buy off the shelf? Mike Wolford [00:56:11]: Yeah, good question. I don't know. A lot of companies won't. You know, if you have a Walmart model where you are a low cost provider, that's absolutely what you will do. The only difference is not everybody buys the economy car or the cheap, you know, car. There are people who buy trucks that don't need them. SUVs, Mercedes, Cadillac, they sell cars too, right? So that'll be the difference. It'll be the economy model, will be the cheap AI that does everything. Mike Wolford [00:56:41]: And those who want a better, more novel experience, the Cadillac type of experience, the Lexus type of experience, they're going to hire a human. And that'll be actually a differentiator. Like here you talk to an actual human being. That'll become a competitive difference. Kortney Harmon [00:56:56]: You know how like you're trying to call somebody, you could press zero to get to the human. You could press zero. Mike Wolford [00:57:01]: Yeah, press zero to talk to a human. Exactly. Yeah. But it's a fair point. And it's not just us, right? Lawyers are going to do the same thing. You're going to go, right, you Cannot afford a lawyer, one will be appointed to you. Chat GPT just became your public defender. And honestly, you probably do a better job than the public defender you'd get. Mike Wolford [00:57:21]: GPT passed the bar at 90%. I don't know a lot of lawyers who got 90% on the bar. If they did, they're probably working at a white shoe law firm, you know, and you can't afford them. So it won't be long until GPT or something like it, it's a public defender. This is going to be everything. Not only that, I think in the not too distant future, it will be illegal and unethical to see a human doctor alone without an AI to diagnose. It'll be unethical, unthinkable. And it's simply because a human doctor takes 20, 30 years and they'll see a disease, a 1 in 100,000 disease once in their career. Mike Wolford [00:57:58]: Well, the AI can be programmed with the sum of human medical knowledge. It's seen whatever rare case a million times is far better at diagnosing than a human will be. And if you talk to a doctor today, anyone will tell you that in a few years that this is going to be true. They need it now. They rely on it now for diagnosing diseases. And that will only get better. Also, AI was used to solve the protein problem in nature. There's 200 million proteins. Mike Wolford [00:58:26]: Long story short, it would take 800 million years for human PhDs to figure those out. AI did it in three months. It's the biggest gift of medical knowledge in human history. It's available publicly to anybody in the world who to wants wants it. So there are huge advances going on behind the scenes. It's just AI is so prevalent in every field, it's hard to get the noise out of it. But it's doing and making incredible progress and doing things that humans could not possibly do without it. And that is exciting. Kortney Harmon [00:58:56]: It is exciting. I have two more questions and I'm going to cut the questions off because I think we could ask you questions all day long. Michael, Chris asked a question. What about companies that say any tech you develop, even as a recruiter, is company property? He's curious how you feel about that. Mike Wolford [00:59:12]: Yeah, good luck enforcing that. I don't know. It depends on the employment agreement you have for my employees. I specifically say anything you invent with the company's resources and doing the company's work belongs to the company. But if you're off building something to trade stocks or plan vacations, like that's yours. Right. So anything you're doing on your time, with your resources that's not related to this work is yours. I don't know how your employment agreement is set up, but I'd also be really careful of that in the future because you don't want to create a really profitable agent or really good agent and then have the company claim it's their property. Mike Wolford [00:59:50]: So that's an employment law thing. Kortney Harmon [00:59:52]: Good advice. Last it was asked, can you give some more examples about the legal industry? I was very vague. I don't necessarily know more details about that. Mike Wolford [01:00:03]: All the research, right, paralegals, that's their job is to do research. Well, OpenAI now and Gemini have a research database. So all you really have to do as a law firm is attach that researcher to your legal database, your digital legal database database. Now, they all have the books, right, to look impressive, but actually it's all on digital. They never actually open those. They breathe from the Internet like everybody else. So that's what you would do. You would upload the library to one of your agents. Mike Wolford [01:00:35]: It would be complicated because it's a huge data set and it takes a minute, but it basically can hold the sum of human knowledge as it is. So it could definitely hold all the legal knowledge. So what you do is you actually are getting rid of a bunch of hay, right? Like, it doesn't need to know about accounting, it doesn't need to know about marketing. It only needs to know about legal. So it's more focused on that. It doesn't have to sort through a bunch of extraneous data. It's just focused on the stuff that's legal. But paralegals will go away and like I said, at some point and AI, you know, there's already an AI CEO, there will be the first AI lawyer that is coming. Mike Wolford [01:01:12]: That is a product that you will see, the first AI doctor, the first AI therapist, which I think will be great benefit to mental health everywhere. You know, if it's three in the morning and you're having a panic attack, you really can't get a hold of your therapist then. But if you really needed it, you could turn on your AI and it's going to have your whole medical history. You've read every book on psychology, every article on psychology and know you and have remember all your conversations and it will get to the point at which you can prescribe your medicine. That's going to be coming. So you're going to get up in the middle of the night, you're going to have this panic attack. It's going to say, okay, you need this, you know, hook to your wrist, right, your watch, and it's got your heart beating, your blood pressure, and it's going to say, you need this. You know, go get a glass of warm milk. Mike Wolford [01:01:57]: Breathe in, breathe out. I've ordered Amazon Instant medicine or whatever it is, right? The drone delivers your stuff at 3:30 in the morning. You're treated, you go back to bed. It's actually a better world, I think, especially in things like healthcare and mental health care. We'll see a huge improvement because humans will have access to good medical advice and knowledge anytime that they need it. That's something we as humans can't provide each other as much as we'd like to. So all kinds of things are going to change. Love it. Kortney Harmon [01:02:28]: Okay, Michael, I'm sure there's more questions. If you want to sign up for Michael's boot camp, it's lexduo.net sounds like there's one coming up. I will be registering myself at this point, Michael, so stay on the lookout. Very excited and he put his information here. Feel free. He's on LinkedIn. I've tagged him in all my posts. Reach out, learn more. Kortney Harmon [01:02:52]: Continue to follow Michael, thank you so much for your time today. You're amazing and I continue to look forward to continuing to learn from you in the future. Mike Wolford [01:02:59]: Thank you so much for having me. Me. I hope it was helpful. Kortney Harmon [01:03:02]: It was wonderful. Thank you very much. Have a wonderful day. We hope you found today's session insightful and inspiring. Remember to stay tuned in the upcoming weeks as we'll be sharing all of this amazing content of our virtual conference. If you missed any part of it, don't forget to subscribe to our show so you don't miss anything upcoming. And if you like this valuable content, if you enjoyed this episode, please feel free to share it with your network and leave us a review on your favorite podcast platform. We'd love to hear your thoughts. Kortney Harmon [01:03:33]: Together we're building a community of growth and learning. Until next time.