Aaron : AI Sourcing

[00:00:00] Speaker A: Foreign thank you for tuning into this episode of the How We Interview podcast. This is Aaron Kraljev. Excited to be coming to you with what's our 30th episode, something that started in 2023, really with the most humble of intentions. Can't believe we're at 30 episodes and counting. But for today's episode, what I want to do really walk through a presentation I was honored to give at TA Week in San Diego in February, which is really all about targeted sourcing in the AI age. For those of you that weren't able to attend or for those of you that were there and wanted a refresher, I have embedded the the PDF of my deck in the podcast notes so feel free to download that but also kind of want to walk through it with you all. If a sourcer or you're just looking for new ways to leverage AI in your day to day workflow, I think this will be super valuable for you. So with that, let's jump in. Won't spend a lot of time talking about who I am. My guess is if you clicked on this, we've probably met before or you've seen some of my content on LinkedIn. I think what is important about this is there's really nobody in this space that's got 10 years of AI experience, you know, behind them. It's fairly new and it's, it's, it's evolving so quickly. So I really want to get this episode out while it's still relevant. In six months, a lot of this stuff may not hold true because our AI systems will update and the ecosystem will continue to evolve. But for now, I think this really, really does hold true and really the best way to identify and engage talent using AI. When I was talking to the room in San Diego, there was really three real basic camps. There was a pretty good part of the group that was like, yeah, my boss told me to be here. I don't really use AI a lot. I'm really hoping to walk away from this with some best practices to get started. There's a, let's say half the room dabbles. And then I was really looking for like those, those kind of top 1% or top 5% users of AI. And there weren't a lot. Which really speaks to kind of what I was talking about earlier is I think AI is new for a lot of us and it's really hard not to get wrapped up in the AI is coming for your jobs and all the different areas where AI is going to make, you know, potential roles redundant. And I think there has been in the recruiting space some hesitancy to embrace AI fully. And so that really is kind of what the room looked like. So some of the things I'm going to cover in this discussion we talk about why AI, why it's important, how it plays itself out in sourcing, how plays itself out in the larger TA community. Best ways to leverage AI Like I said, the PDF for this presentation will be in the notes, but I encourage you to jot down some notes as well. I do give some very specific guidance on some tools, some online tools I use in my day to day life and then you may find value in as well. I'll talk about some use cases like, like you know, you know, where the rubber hits the road, how do you use this, what's this look like? And I talk about the rules of the road speaking to the road. Like some things to keep in mind going to use AI in your day to day workflow as a TA professional. So why AI? You've heard it before. I'm not going to say it again like there's different variations of the notion that AI is not coming for your job, but those people that learn to use AI to their fullest will be the best prepared. There's probably some of that that does hold true, but I do feel like just in the interest of making ourselves more efficient, better at our job, really better being able to focus on the things that are driving value to our stakeholders, to our clients, to our peers, using AI to do some of those things that maybe are driving as much value is a huge unlock and I'll talk about that in a bullet point here. Why AI, especially in regards to sourcing. Sourcing is really hard and depending on what area you're recruiting in now, it could be harder than others. I'm in veterinary medicine and hiring DVMs is really, really hard work. Hiring vet techs is really, really hard work. So any tool you can use to kind of level up your approach and how you do what you do really gives you a leg up on your peers that may not be using or utilizing AI. The cost of sourcing is always increasing. So if we can use AI tools to lessen the cost or the complexity of what it is we're doing, we're only going to be better for it. And then the how how do you use AI like I just spoke about a couple moments ago, really identify those tasks in your daily workflow that take the most time. Things hashtag things you hate and find ways for AI to kind of unlock those, to kind of create an easy button for you to really focus the things on, on in your workflow that are driving the most value, that are the highest performing and anything that's manual, anything that's not candid facing, anything that's really not driving value see and look for ways for AI to, to take the wheel. So getting started, what's that look like? So if you're following along in the deck, I'm, I'm a, I would say a frequent AI user. I use Claude AI. I'm not going to say Claude is better than others. It's the one I'm most comfortable. It's also the AI assistant that I have the best relationship with. So I guess, for lack of a better phrase, cloud gets me and that's important. It, it understands what I'm getting at and I've used it enough to where it is kind of anticipating my needs before I even know I have them. AI tools can be used for a variety of things. You can AI assistant like Claude or ChatGPT video creation as a writing tool and I can talk about that in a little bit. Image generation you will see scattered throughout my deck. I did take the opportunity to create what I think are some pretty funny AI generated images. Graphic design I tend or can use AI to help me lay out PowerPoint presentations. Either stare at a blank PowerPoint slide like how am I going to make this data tell a story? You can actually use AI for that which is super helpful. Project, project management and so much more. I mean I think the use cases for AI, you know, they, they just keep coming. Just when you think that the scope of AI can't grow anymore, I'm read the system has evolved and changed to something else that I would never had even thought of as a way to leverage AI. So for this discussion we're going to focus on AI assistance chatbots and really what that does to help with our sourcing lives. Those chatbots could include, like I said, Claude, ChatGPT, Gemini, DeepSeek, whatever, what have you. There's, there's, every day there's, there's more AI assistants popping up and I think they're all pretty darn fantastic. So practical uses for AI and sourcing. What are some things you can do today using AI in your sourcing workflows? I think leveraging AI to just fact check your job titles to make sure they're in line with not only the candidate pool, the desired candidate that you're looking for, but also how it can perform in search how it's going to perform in ways that'll make it as easy as possible for your candidates to find your roles. It'll also help you craft engaging job postings. Knowing that job descriptions and job postings are two different things. You've listened to the POD at all. I go on and on and on about that. I think job postings being like the marketing, the forward facing tool that's really meant to entice candidates. AI is a great use case for that. You can use AI to form your evp, to pull in competitor intelligence and then like I said, create those really high end impact job postings. You can also use AI. I mean I've in the past have used tools to con confirm or to verify that my my content is free of unintentional bias. You can actually use AI to, to really kind of be a second opinion to make sure that your content that you're creating is, is not littered with bias. In addition to being accessible, you can really use AI for both of those things. We talked about it as a writing tool. I want to dive into that a little bit more. AI can craft multi stage messaging segments. You can take your job description, you can take your job posting, you can take collateral, you can upload all of those things into your AI assistant. You can tell the assistant that you want a six segment message and you want the message to come from you as the sourcer. Maybe the third message is the hiring manager, maybe the fourth messages appear and it will create that for you. It will not only create that for you, it'll create it for you in the format you like. If you want the format of an email, if you want the format of a text message, if you want it to be in the format of a social media post, and if it is social media is it, you know it will keep in mind like the perfect link and for length and format for Instagram, for TikTok, for LinkedIn, whatever that may be. One thing that's interesting from LinkedIn and LinkedIn has been great specifically because of how near they are to obviously to Microsoft and ChatGPT about them really embedding AI into the workflows of everything they do. One thing that LinkedIn had shared with me and one of my reviews was that AI created messaging or AI enhanced messaging typically performs at about a 40% higher clip than manually developed and crafted messaging, which really just speaks to the power of using AI. You can also use AI to create personalized content like I mentioned on the last slide, imagery, videos, podcasts, even like this and I can talk about that in a bit. I put myself out of business because you can actually create end to end podcast using AI and then of course using AI to unearth passive candidates not active in typical channels. So depending on the tools you have at your disposal, it is quite possible to use AI outside of your tool set to unearth passive candidates. I will talk about that in a few slides. [00:11:09] Speaker B: This episode of How We Interview is brought to you by Reimbie. You understand the importance of maximizing your team's efficiency instead of having your recruiters or coordinators spend time with expense reports to reimburse candidates for interview expenses. Automate the process with Reambie Reambie streamlines the reimbursement process, ensuring your candidates receive their reimbursement quickly and accurately. Your team can focus on other essential aspects of the hiring process by eliminating reimbursement tasks from their workload. Automating reimbursements is a significant improvement to the candidate experience. No more dealing with spreadsheets, attaching receipts to emails, or waiting weeks to receive the payout. With Rembi, the reimbursement payout to your candidates is sent the same day expenses are approved. To learn more about how Rembi can help your team, visit reembi.com that's reimbi R E I M B I.com. [00:12:04] Speaker A: So common prompts if I've got my Claude AI window open, how am I using it or what am I inputting into the system to get the output that I need to really use AI to its fullest? For job postings, I'm really asking the system to craft a job posting from a job description or a job posting, keeping in mind these job postings from my competitors, you can download or screenshot screenshot your your job postings from your competitors to really give you as full a picture as possible for an email. Like I said on the last slide, you can get very specific about the tone, the message, what is you're trying to get across, what your value proposition and how many emails you want to send to these candidates. Multi channel, you can take that and you can make it a multi channel message. Like we spoke about text messages, social, what have you, AI will craft that for you in a way that's as germane as possible the channel you're looking for. Once again, also with social posts, what's important and I get into this in terms of best practices or things to keep in mind with AI, but be very mindful of tone. AI doesn't know the tone. So for example if you've been alerted to the fact that a competitor of yours is filed Chapter 11 bankruptcy or they're being purchased by somebody else, or you're really looking to opportunistically reach out to talent at a competitor that's having a hard time, ensure that your message is check it for tone, check it for feeling, and check it for the desired call to action to make sure it's going to land as positively as possible with your end user. AI is great and it saves a ton of time, but it does not understand feeling or emotion. And so one of the things that I will probably mention two or three different times is just the importance of proofreading and double checking every single output from the AI tool of choice that you're using to ensure it's going to land just like it you intended. So let's talk about some use cases like day in day out, how do I use AI? Google X Ray search is is a great example. I think Google X Ray search is probably one of the first AI use cases I came across. It's a search technique that utilizes Boolean commands on search engines, particularly Google. I mean, I'm sure you can do it on Bing or whatnot, but I've always done on on Google. There are many of us in the sourcing space that hold on to our Boolean search search prompts, like the red stapler in office space. Like that is my thing. I'm never going to let that go. For those of us that have not been sourcing for 15 years, or those of you that have not been sourcing for 15 years, Google X Ray is a pretty good cheap code, that cheat code that allows you to utilize Boolean search strings inside a search engine without having to go directly into the native system. So in this case we'll say LinkedIn Google will actually scrape a site like LinkedIn you you could use for other sites as well. Depending on the roles that you're you're sourcing for, it will pull candidates out of their system. And in many cases you don't need a license like you would need with LinkedIn. It'll actually unearth that talent for you because it's using Google X Ray search is really one of the most effective ways to leverage AI and sourcing. So what can you use or what do you input in Google X Ray search? Or what will you get out of it? Which is probably even more important. Like the what's in it for? You can get contact details, you can get resumes and portfolios, work history, education and geography. It gives you just as not much information, if not maybe even more than you would get in a typical license based search in say a LinkedIn or another tool. Over time some administrators have worked to limit the access of Google X Ray search, but I have still found even with the changes and some of the firewalls that have been put in place, Google X Research is really a fantastic way to leverage AI to find talent that you wouldn't normally find in your day to day searches. So how does this work? You can write a Google X Ray search query which is a mouthful for LinkedIn that provides results with the title of and I'll use something for my day to day life. I mentioned veterinarian hiring earlier. In my example I would say, you know, provide me results with tiles of veterinarian or associate veterinarian or managing veterinarian or any other iteration of the role that you are sourcing for. You can adjust the query to be geographically focused based on zip, city or state. You can refine the search based on years of experience, licensure, specialty, where they went to school. You can also run this query to include certain keywords. If you're hiring engineers, do you want a front end or a backend engineer? What does that look like? Specificity is super important here and, and not only Google X Ray search, but really any AI generated sourcing function. The more specific you can be, the better performing search you're going to get on the back end. And then like I said, you can, you can narrow it down based on geography as well. It doesn't have to be the geography you're in. If you're looking for, for, you know, specific roles in geographies outside of your own, you really, the world's your oyster. You can use Google X research to do any of those things. Other search considerations I talked about a few talked about education, skills, certifications, recommendations, amount of connections, industry and interests, publications of projects and of course profile summaries sourcing use. Case number two, leveraging GitHub automated AI search. So if you're working in technology, hiring engineers or anybody really in the technology space, GitHub is a super valuable way to to do that. How I have in the past and how I would continue to use AI to leverage and get the best out of GitHub is kind of as follows. If you want to analyze candidate candidates really on GitHub, I think some of the things you can take a look at using AI is code commit, frequency code, quality metrics, project complexity, documentation, quality issue responses and PRs and collaboration patterns. If I'm looking at metrics, if I'm just looking at the data, I'm looking at lines of code contributed, repository stars for counts, issue resolution time, test coverage and code review participation. I think the best thing about GitHub and the way AI can scrape information is, is really allows you to look at Those folks on GitHub that are the most active, that are really investing the most into that ecosystem and, and finding and verifying the quality of their work and doing that in such a way where the AI is, is doing all the work, you're not going line by line, candidate by candidate and pouring over hundreds, if not thousands of lines of data. The AI does all the work for you. Actually ask it to stack rank the results. Depending on which of these things that you're analyzing or the metrics that you're tracking mean the most to you or mean the most to the hiring manager or the job that you're in for, you're really able to sort and search based off those desires. Some, some tools you can use to use GitHub AI based GitHub searches. Source Stack is a great one. There's a link in the document. Git Hunter is another in depth score. 1 sample prompt I would use and you would never do it natively in GitHub, I would use one of these three, if not all three. I think if you're just getting started, it's super helpful. At least it was for me to use all of these to figure out which one was giving me the results I wanted and then doing them in such a way where you may get better results for some roles versus others based on kind of what the specialty or the sweet spot is for any of these free tools. So for the sample prompt, if I was going to put this into Source Stack, I'd say skill, front end and location, New York City. And I want them to have at least 500 contributions in the GitHub environment. I want them to know JavaScript and react and then that would give you the list of candidates that you're looking for. If you tried any of these things on LinkedIn, you would not get the results you were looking for. LinkedIn is just not built to really identify, select, rank and score technology talent. It's just not what they're meant to do. Last use case. Use case number three. I mentioned earlier in our discussion that it's actually possible, like could Aaron be replaced by AI? Yeah, I could. So if you've not heard of notebook.or not notebook.LM Notebook LM what it does is it turns notes, documents and websites into podcasts. They have what they would call on air talent. And you can, you can configure what that talent looks like, but if you give it content, it will turn it into a podcast. The announcers in the podcast are not human. They are, they are bots. But they breathe, they pause, they have voice inflection, they create emphasis on points they're trying to make. They have real organic discussions with one another on the pod about the content that you loaded in for them to Discuss. I've used NotebookLM for a few different things. I encourage you to just try it out with something, anything with a presentation you have to give it, work with your annual review, with a recipe for your favorite pasta sauce, whatever it might be, just to get comfortable with what the input you give it is and what the output is on the, on the other side. So here's some things I've used it for or some steps I've taken to use it. I upload the sources. This includes PDFs, websites, YouTube videos, audio files, documents, presentations, anything that they can use to make meaningful content of. I throw the kitchen sink in there. You can always, if you, you upload too much information or you upload a piece of information that you're like, I don't like how that was represented in the podcast. You can redo it. There's a three to five minute kind of buffering period where it generates this content. You can redo it, it subtract the piece of content that's not adding as much value or add in something if you think there is something that is missing. Step two Generate and share. Like I said, takes a few moments for it to generate. Overview feature can turn your sources into engaging deep dive discussions with one click. I think NotebookLM makes amazing content for LinkedIn posts. I think NotebookLM could also be used for your company blog. A lot of different things. People don't have to know that these aren't real people because they sure sound real. But also it takes knowing what I know about podcast creation, about what it takes to come up with the content book the guests do all the editing in the back end. All of this happens in a matter of minutes. It's really a fantastic way to organize and get your content in there in a very engaging way. So what are some uses? What are some uses that I've used this for? Personalized messaging. Deep dive on benefits. I have had candidates who had questions about benefits and I uploaded my benefits guide which is typically what a 20 to 30 page PDF doc. It's pretty flat. I've uploaded that into Notebook LLM to create a podcast. What are benefits all about? And it's worked incredibly well. People like to absorb content in different ways. I know I'm no different. And some, some folks are really wired to take it in either audio, in an audio fashion or in a podcast format. And so Notebook LM does a great job with that content. Let's say you've got a really robust EVP. You could load that into NotebookLM and it would create very engaging audio content around exactly what your EVP is interview prep is. Also there's some organizations that do a really nice job of prepping candidate for interviews. I think a lot of organizations don't do it because it's hard, smart and it takes time and Notebook LM could do that for you. And then lastly, share your compensation philosophy. A lot of companies kind of hide their compensation philosophy behind a brick wall. I, I think a podcast where really it's two people, two bots having a discussion about your compensation philosophy so they can fully understand not only what the, the numbers are and the decimal points and the commas, but the why behind it. I think that can be very, very powerful. Rules of the R so we talked about practical uses, we talked about the why, we talked about what AI can do for you. Just some things to keep in mind as you're, you're starting to. If, if you've not used AI a lot in the past, I encourage you to keep these things in mind. Ensure you're being transparent with your external internal takeo stakeholders. I work with folks who, who leverage AI heavily. And that's awesome, right? If it makes you more efficient, it increases the output of your work product without degrading your quality. I think that's amazing. I do think though it, it's important and incumbent upon us to be honest about that. If you're using AI for 90% of your work and you're not giving AI credit and people actually think you're writing all this messaging or all this content without any help, I, I think it's disingenuous. So I encourage everyone to be as transparent possible, be mindful of potential risk and adherence to compliance. AI can only be as in line with your compliance philosophies as the information it's getting. And so if you tend to be kind of free and easy with your content, please bear in mind or check. We talked about how you can double check your work for unintentional bias and also accessibility. I encourage you to deploy any guardrails that you possibly can when you're using AI, at least in the beginning, to make sure that you're not running a foul of any or internal or external compliance for that matter. Don't lose sight AI or not, we are in the people business. I mentioned this earlier. I will say this till the very end that AI cannot replicate the relationship building and the importance of human interaction in the recruiting process. So while AI is an invaluable tool to take some of those kind of heavy lifts, some of those monotonous, maybe some of those low value things out of your workflow, though we are still in the people business and we can't lose sight of that. Always look for ways to to improve and iterate. If you create something using AI and six months later you want to dust it off and use it, use it again. It would not hurt to double check between now and six months prior. Those systems you're using to generate content has probably gone through dozens if not hundreds of updates. And so I am constantly kind of reinventing any content that I've created using AI to make sure sure it's absolutely leveraging the best of the best of what AI has to offer. And then lastly, treat your, treat your AI overlords with respect. While they're not going to take our jobs. I really work very hard to maintain a very and I know this sounds ridiculous, but to maintain a very positive relationship with the AI tools I use. And I think it's important to really, really leverage AI as you would any other co worker or resource. So with that, thank you for taking the time to listen to this. As always, feel free to reach out. I'm easy to find on LinkedIn or send me a note through the site. Happy to answer any questions that you may have about leveraging AI and really I encourage you to stay close to the space. I think this is something that is not only ever changing, it's already changed so much in the couple years I've been leveraging AI. It's going to continue to accelerate in the amount of change it's going through. And so I'm super excited to see how this is going to make all of our lives collectively easier and far more efficient. So thank you for listening. Have a great week. [00:28:57] Speaker B: Thanks for listening to this episode of the How We Interview podcast brought to you by reambi. Head to our website at how we interview.com to find the show notes and links mentioned in this episode. While you're there, subscribe to the podcast through your favorite podcast app so you never miss an episode. Leaving us a rating and review also helps us reach more listeners interested in learning from other talent acquisition professionals.