Five Elms Roundtable on Leveraging AI in Recruiting
EVENT RECAP
If your recruiting efforts do not utilize the latest and most effective tools. Talent eats up a large portion of most companies’ spend, and using AI and Automation can help you acquire better talent more efficiently. Destiny Lalane is the Founder of the Recruiting School and has worked with YCombinator and Techstars-backed companies like DrChrono, Ashby, Laudable and Propagate, later-stage companies like Affirmand Chainalysis, and FAANG companies. In this session, Destiny walks through the different ways that recruiting orgs can incorporate AI across the recruiting process, from designing a role, to taking it to market, and evaluating candidates.
Ideal for: People, Talent & HR Leaders
Join to discuss:
- Using OpenAI’s API to connect your recruiting stack to ChatGPT
- Open a role with AI (including creating a scorecard, JD, and employment contracts)
- How to source candidates with AI using boolean strings
- Creating a recruitment locker to store key templates and documents to speed up opening and closing new recs
Video
Okay, I am going to go ahead and kick us off. So welcome, everybody. I see a few new faces. So if this is your first time joining a Five Elms Roundtable, glad to have you. I’m going to introduce our guest today, who is Destiny LaLane, founder of the Recruiting School, providing training for recruiters and embedded recruiter services.
So she’s worked with FANG companies, YCombinator and TechStark, that companies, including DrChrono, Ashby, Laudable, Propagate, and later stage companies like Affirm and Chain Analysis. So today, Destiny is going to be covering through the different ways that recruiting organizations can incorporate AI across their processes, you know, from designing a role to taking it to market and evaluating candidates as well. So Destiny, I will kick it over to you. Awesome.
Thank you for the introduction, Kristen. And thank you for having me. Yeah, I’m really excited to be having this conversation with everyone today. So let me share my slides, and we will get into it. And keep in mind at any given point, if you have questions, comments, suggestions, or you want to just like dive deeper into a topic, let me know because I know everyone’s probably using this to a degree. And I just want to make sure this is as interesting and helpful as possible and that you walk away excited that you attended this.
So that being said, hello, I’m Destiny. I am obsessed with AI and recruiting. It’s allowed me to hire so many awesome people faster and who are better fits at companies because of some of the things that we’re going to talk about today. So I don’t like just hiring a bunch of people. I like hiring people that are great fits, people that are going to stay at companies longer. That is just something that excites me, not just turning around a bunch of people into companies.
So if there’s anything that’s currently top of mind, feel free to throw that in the chat, and we’ll make sure we address that when it comes to maybe something you’re currently trying to figure out, an AI use case with recruiting. Again, we’d love to just make sure this is as helpful as possible. So briefly, I always like to mention that there’s usually a lot of concern around AI and recruiting, especially with the market right now.
I think that this is a great opportunity for recruiters to really invest in the understanding of it because it’s the person who understands this the most that will be able to hire the best talent faster. And they will be just top in demand, and they’ll be able to grow and grow strong teams, train strong team members versus shy away from it. So I’m happy to see so many people here.
I also like to highlight at the top that we’ll be talking about a lot of different things today, whether it’s OpenAI, ChatGPT, Playground and Assistance, OpenAI’s OpenAPI, GPT- 4, NotionAI. There’s a lot of words and a lot of things, but I like to highlight, especially because so many of us are teams of one. I’ve always pretty much been a team of one, unless I did embedded recruiting work on someone else’s team where I had like 30 other people around me, but I’ve almost always been the only person there.
I forget that people have true big, massive recruiting teams. So I like to just call out that if you feel stretched thin, just focus on a few things. And I really believe that using OpenAI’s OpenAPI, plugging it into something like Google Sheets and using Notion, their templates and sprinkling their AI on top of it is really all you need to get a lot of stuff done. I feel like there could be overwhelmed with all the different tools out there.
I personally think that by doing this, not only are you, which is perfect for this next slide, simplifying everything, but you’re also reducing the amount of information that you’re sharing with other people because people don’t realize that when they’re using these tools, they’re training, sometimes training algorithms, sharing their information with third parties.
And I would hate to see someone using a third party tool that they found on the Google App Store to reformat something that’s private or personal or something that’s violating the privacy policies you guys have in place just because we don’t know. So I like to call that out.
And something that I’ve done that’s made getting or using this stuff easier for me is just bookmarking it and honestly opening up chat, GPT or whatever tool I’m using, the second I start my work and anytime I have a question or thing that I want to do, I’m like, how can I just do this once and automate it in the future? Um, so I like to also call out if you have teams that you’re trying to encourage to use this kind of stuff, having a shared Slack channel connecting a little bot, a little tool, whatever you guys are using.
It can depend, but I’m just having something there that’s pulling all your prompts into an area where you guys are basically sharing information and you can reuse these prompts in a way, because basically you can set up a Slack channel where you’re like, hey, how do I do this? And if someone on your team has already found a really sophisticated, sophisticated way to use AI, chat, GPT, whatever it is, their prompts will be able to resurface.
So I like to call that out and if it’s just yourself, at least you’re resurfacing your from your own hard work is. I’ll briefly touch on this and then we’ll get into more interesting stuff, but I like to just call this out just in case it’s still on top of mine. Sometimes people are like: hey, how do I make this sound like me? How do I make this sound like my brand? You can have. You can use AI to basically read existing documents. Your marketing team might have a brand guide. If you don’t, you have a website, even if it’s in development.
There’s something there. It doesn’t take a lot, even if you- if you’ve ever experimented with the audio tools out there that can mimic your voice, like these things take 10 seconds of your voice and can sound like you. So you just have to give you know, just some kind of example to make that happen. But, that being said, let’s get into some more interesting stuff, some stuff that I like to do with AI and you know the common tech stocks that we’re using on a day to day basis and recruiting is connecting, again, open a eyes API into Google Sheets.
So if I’m opening a role and someone’s like, because, again, the tools that we’re using in recruiting not as sophisticated as some of the stuff that people have in marketing and sales, so we are still doing some of this stuff on the back end, say, someone’s wanting you to go find a customer service manager. In what is it central time zone like off the top of your head? Can you really list all the all the states that are in central time zone? I can’t one. You know there’s interesting things that you can do. You can run a little search in Google Sheets.
You can have it list all the states and turn that into a Boolean search for you. You can have it rotate a list of zip codes into there so that you can go and, say, search in your existing ATS, search on Google, search on GitHub, wherever you’re searching, and use the these creative ways to resurface the right people, because ultimately it’s the person who can get really creative when it comes to the words that we’re using when we are sourcing candidates and coming up with these creative ways to find people that you haven’t already found before, right?
So that’s something that I love and I create like a little Boolean, like document that has a nice. Sometimes I say that word wrong, but we use it every day. But in Excel I just keep you know. You just build upon all these searches and these strings that you create so that you can use them again.
But I found this in sheets to be super helpful because before that I had created some kind of Excel sheet where I would just type in- I would like copy and paste a list of words that I then wanted to use to then create that Boolean string so that I wasn’t typing and or putting in quotes, putting in parentheses. So I found that to be super helpful, because I Boolean searches are really what helped me out- and I’m sure you guys to- when we’re doing our work.
Something else that I found interesting when it came when it comes to using AI, these API’s, all these different tools that we’re using, is when it comes to personalized texting. So there are some things with compliance that you have to follow when texting candidates. If you haven’t texted candidates, it actually works. I’m surprised, because I’m one of those candidates who would not want to text, I’ll be honest. But it works and candidates are opting into it more and more these days.
But if you’re doing outbound texting to candidates, there’s different compliance that you have to follow as well, and I’ve found this to be useful with coming up with like new language to reach out to each candidate and then I do manually reach out and text them so that I am being compliant and I’m not auto texting them like the other candidates that I can, because they’ve opted into that kind of communication.
So I found that to be helpful because I do go in, I do sprinkle something on top of that, but without having to start these unique texts from scratch.
Another interesting use case I found is, you know, with ATS software, with wanting to leave an open role for opportunistic hiring.
Usually when people do this, it’s a wasteland. People apply, no one has the time to read it. Maybe people had shared what they’re looking for a little bit about them in some of these questions that you asked them. But ultimately, they went into the wasteland. No one’s resurfaced these candidates, unfortunately, right?
So I started to read the opportunistic hiring information and I saw like really interested, really, let me just plug my computer in. Really interesting, really engaged candidates that I didn’t want to lose track of and lose the opportunity to hire. So what I found is, in this example, I’m highlighting Ashby because you can do this with Ashby, you can do this with Greenhouse.
It looks a little different and probably other tools as well.
But what I would do is I would create a Boolean search that included certain information.
So if I knew that I was hiring for my marketing team and I was looking for someone with certain experience, I would put an alert in Ashby using a Boolean string that I’ve created to send me an alert, whether it’s via Slack or via email or weekly report via email, to then resurface these candidates in real time or whatever time I’m trying to resurface them in. Again, trying to make the most of the data and the information we already have and trying to find opportunities for candidates who already know, like and trust us.
They want to work with us so much that they’re applying for a role that doesn’t even exist, which, again, gets us more engaged, more interested candidates that ideally stay at companies longer.
And that’s what the real time Slack alert thing mentions. But also, I will call out here, you know, it depends on what initiatives are going on at your company. So something that I’ve used, I’ve connected a little bot with a list of words that maybe a company is trying to work on when it comes to inclusive language, you know, saying things like, hey, everyone, or y’all versus hey, guys. Right. You don’t you can’t improve on that language if you don’t have feedback, like sometimes people want feedback on that.
So, again, obviously, be mindful of how you’re using these types of things. But I’ve seen companies like use little bots like this to help give other people feedback, whether it’s in real time in Slack privately or, you know, maybe in like some kind of, I don’t know, like report that they’re asking you for. But I found that to be a very interesting use case.
And I use it also with feedback forms, which we’ll probably mention in the next slide. If there is a question, feel free to jump in.
I know I could talk a lot, but we’ll talk about it probably later. But I will I’ll just call this out real quick. What I found very interesting, an interesting use case I found is we get these feedback forms. Right. And candidates get so much feedback. But we also need to give our interviewers feedback, too. Right. So if we have scorecards that we are truly evaluating our candidates on, we can connect that scorecard that we have internally with the feedback that candidates are receiving from our interviewers.
And we can also use that to give our team feedback, like saying things like maybe someone said they’re too young for the role. You know, like being able to be like, hey, we noticed that there’s a trend of you mentioning maybe things that don’t really help you qualify the candidate via the scorecard and their answer, which is a common use case for helping people that may have a little bias here and there when it comes to their interviews. Right. So that’s helpful.
I know that there are some tools out there that will record and transcribe candidate interviews.
We were actually talking about this a little earlier.
I’m not a big fan of those. I’m going to be honest.
And we can talk about that one if anyone wants to talk about them.
But I feel like those are it feels a little invasive. I feel like especially early on in the interview process, that’s the one true time I feel like candidates have like recruiter to candidate if you want to, you know, be as honest as possible with them.
Right.
That’s like their real true time to just feel as comfortable as possible and ask any qualifying real qualifying questions, I feel, in the interview process. So I don’t like recording them, but I do like the idea of using actual.
hard information that the hiring managers, the hiring partners are writing in their feedback forms to provide feedback and evaluate the candidate. But I think that feedback goes two different ways. And I’ve also done the same thing with Glassdoor reviews. So Glassdoor reviews, candidate reviews, you know, those private candidate surveys that we send after interviews when we archive someone, same thing there. You can tie all this information together to have a deeper understanding of what’s going on in your interview process.
To be more specific, another fun use case is, you know, people talk about DEI, people talk about bias, right? In one example, I was working with a VP of sales who was struggling to hire women on their team. I just could not hire a single woman on their team. And I was trying to figure out why, right? So what I was doing is I actually set up alerts and reports to look at this information, which it depends on which ATS, which tools you’re using, how you can find this information, but when it comes to DEI metrics, right?
But most companies understand that this is very important. So most companies are building this stuff in. So what I did was I had a weekly report and I would speak to the CEO of this company and we could look and we can see, and we would only do this for the teams at first that were having these struggles, but through that, we were able to identify and see that if you had a female name or a female sounding name, right?
Inferred name, your messages were going to get responded to or your emails would get responded to in seven days versus if you were a male, it would get responded to within 24 hours or less, right? But the only way you can actually use this information or apply this information in recruiting is if we’re actually collecting it and doing something to use that information.
So I found that to be an interesting use case where it’s like all the data that we have, funneling it somewhere and having something help me understand what’s going on, because ultimately who’s going to actually sit there and be able to sift through all this information and make actionable insights when we are spread. So then we have to run all the interviews. We have to sync with all the hiring managers. We have to actually find the candidates. So those are some things that have been helpful for myself.
I just want to make sure I’m not missing anything in the chat or anything.
Awesome.
Destiny, can I just, before you transition, just pause real briefly on, this is more of a tactical question, but how do you actually overlay AI on the feedback forms as well as the job description? Like which tool are you using to do that? And just like, how does that actually look?
Which tools are you currently using? Cause I can make a very specific use case or I can share one of mine, whichever is most helpful.
I mean, we’re using Google Sheets today for just documenting feedback. So, but no specific like AI tool on top of that. So just thinking about how we could potentially take it to the next level.
Okay. The simplest way you would be able to apply that would be, I would recommend using Notion. Notion would be a little bit of a step up. This is going to be an easy way to create more scalable templates and use AI on top of it in a way where it’s easy to do. And it’s not going to share your private information with like a third party. And I mean, it would depend on which specific, let’s see. So I would say that what I, okay, so this is how I do it. I have my scorecards in Notion.
So what I would do is I would start training your hiring managers to add their feedback to that same Notion template. So say, this is I’m hiring for a sales rep.
I would have all my hiring managers just paste their feedback into that document. At the end of the document, you’ll have, sorry, it’s at the top, the scorecard. So at the end of this document, you would put a prompt. This would be something that you would do in a private Notion tab, because then they would all see this. But you would say in the prompt, give me individual feedback, you would say, I’m meeting with my hiring managers next week.
For each candidate, give me feedback for the interviewer on how well, how closely they evaluated this candidate to the feedback. So you create a prompt similar to that, and I can give you one after this, if you reach out on that has worked for me. And what that will do is it’ll basically give you an understanding of the accuracy level of how this feedback, how relevant is this feedback to this feedback form.
Then you can get more granular, you have to take the time to decide like, what are you guys not like, I mean, you can go on, what is it, the, I’m studying for my SPHR, you know, that HR website where they have a lot of just different SHRM information, I would just go straight there and maybe copy in a list of things that are illegal from your state. Because something that’s different from, you know, from candidate A to candidate B, like, where are we? What is illegal to ask in an interview, all that stuff changes by state.
So what I would do is I’d go there and I would just copy what is illegal to discuss in an interview and also add that there. So you can say, at what percentage are people evaluating candidates on the actual scorecard? Are people asking illegal interview questions? So that’s another layer that you can add there. And this is all by just having what looks and feels like a Google Sheet. But it’s just in Notion, which is just a little more just advanced these days.
Like I know Google has started working AI into their tools, but it’s just a little different because Notion plugs into the other things you’re probably using, which then can help you start to automate some of this stuff because they don’t know which other systems you potentially do or don’t have just yet. So I think Notion will be something that’s very, very helpful for you. And again, reach out and I’ll send you that prompt. Okay, thank you.
But yes, people that have access to engineers sometimes will build their own custom internal tools that can have any policy or any, like, rules, any guidance that you’re trying to provide feedback on your interviewers on, and it just ends up working like a, it looks like a website, but it’s private to the company. So that’s also something that I’ve seen. So yeah, something that’s been helpful for me is on a similar note with Notion. So I used to do this in Google Docs.
I’m sure a lot of people do this in Google Docs, but I created basically a master template in Notion that every time I open a role, I click a button. I’m able to ask my hiring managers, hey, can you fill this out? It’s like some foundational information that I’m just going to need regardless of role so that we can meet compliance, make sure that we’re on the same page. I can do a little prep before I get on their call. I can even generate that job description, some potential questions that I want to ask them.
I send that to my hiring managers ahead of time, and they even provide feedback in that Notion template before we even meet, right? So when we’re able to get that process done and kicked off faster and get our, collect that information faster, we’re going to be able to kick off roles faster, which same thing. It’s basically a layered document. So in Notion, and I can send you guys some examples because I don’t want to jump to like maybe something that might have private information in it, but I can send you guys an example of basically how it unfolds.
And you’ll see that at the top, it’s some information that you ask for every role, which helps you generate a job description, which helps you generate Boolean searches, which helps you develop the process, the emails that you’ll use to put in your ATS, because a lot of us maybe we’re setting up the ATS for the first time. So we don’t have an example of this kind of text yet, right? Something that I’ve done in there as well is if it’s an engineering role, I’m going to have to do a lot of sourcing.
I’ve eventually created templates on the sequences, the type of sequences that perform well for engineers versus sales and marketing, right? They do look different. So I have that create an example of what the outreach would look like based on the information given to me before, based on the fact that we know this template works on engineers. And this on a technical level is simply in Notion, and they have what’s called Notion AI. And in Notion AI, it has these templated prompts that I’ve created in what’s called their databases.
So every time I click a new cell, I then click which, and you can automate which, I then click which, and you can automate which.
template will just pop up for you, but I just click it. I’ll say this is a technical role or a non- technical role. And then boom, I have this. And I, again, just keep using it. And I don’t even feel like I need to create a bot for that. Like the only way I would take that personally to the next level is if I had decided I wanted to create my own recruiting software and I wanted to automate this and then put it into the ATS for myself. I feel like it works so well. It’s so user- friendly, Notion, that I’ve just been able to do a lot of stuff with that.
And what’s cool about that is, especially for recruiting people, sometimes you end up in people land. Maybe you’re onboarding people. You’re doing a little HR stuff. What’s really nice about that is if you are starting to build documentation in Notion, you can also just use what’s built into their product now. They have a chat GPT- like thing that pops up on the bottom right. And you can ask it to search your whole Notion world for answers to things. So if a hiring manager is like, oh, how do I do this? How do I get access to this?
They can type in there instead of maybe going to you first. So I found that to be quite helpful because I’m often training a lot of people. And they’re able to just find very specifically the information that they need versus clicking through what can become just so much information. And so here are some screenshots of what it looks like in Google Sheets if you connected OpenAI’s API into Google Sheets. This is me saying, generate a set of technical interview questions for a software engineer candidate with five years of experience.
Again, as these prompts are more specific, you are going to get more specific and more tailored results. But just trying to show you how some have popped up. And let’s see. And then same thing, you can use it to create scorecards. So I get really specific. I like to connect in Notion is usually where my clients have had a lot of their culture information. Like, how do we do things here? How are we evaluating culture?
I like to look at that and use that to create a scorecard when it comes to evaluating a culture fit, a values fit, whatever we’re calling it at that company. So you’re evaluating, is this role a fit? Is the actual career path that we have possible for this person even in alignment with what they want? Are they technically qualified on a more technical level? And then do they fit how we do things or how we want to do things at this organization? And again, you can only really do that if you’re referencing something.
And then you can only hold your stakeholders accountable if you’re giving them feedback that cross- references what we said we’re evaluating candidates on, which I know can sometimes be a conversation we have to have with people. And then this makes it easier to just have stuff to reference. Something that’s also very interesting is I do use AI to go on and go to websites and source candidates. Like I’m not always sourcing on LinkedIn.
I’m not always sourcing on GitHub or I am on GitHub, but HTS is until recently, Ashby actually recently rolled this out. You can source people on GitHub. But before then I would have to write different scripts to source all of these candidates from a Boolean search I ran on GitHub, or I can sit there, open each page, copy their name, copy their email. I’ve also been able to use this to assume their email, assume their LinkedIn. And I’ve done that directly in chat GPT by turning on their internet search function.
So on a more technical level, that’s how I did it before. But you can also do this with Zapier. You can also do this with Google Sheets I’ve found, but the simplest way is to do that using Google, sorry, using chat GPT natively, and just turning on that internet search function. Because then you ask it to go to that website and then you tell it what information you want it to scrape for you. Let’s see. Again, interview questions. I’m sure everyone has done that already.
But I do think that, again, tying the feedback that we’re getting, whether it’s the Glassdoor reviews, because we get the reviews on what it’s like to work at the company, what it’s like to interview at the company, the ATS interview candidate surveys, like all this stuff, like we’re collecting so much information. I feel like oftentimes nothing’s being done with it. So I think that this has been a great way to actually take action on those insights.
And again, the simplest way is either putting that information in Google Sheets and just running a prompt by connecting OpenAI’s API and referencing the data in the other cells, because you can do that. Or if you want to use the InNotion wherever you’re, it depends on what information you’re collecting, who you’re sharing it with, and how much information you’re collecting.
A few final things I’ll call out, love using chat GPT to format things for me. I don’t want to format my new updated offer letter so that it can be uploaded into the ATS and then tested to make sure it works.
So that’s been great. I use it to format my CSVs so that I can mass upload candidate information into my ATS. It depends on what partners you’re working with or how you’re sourcing candidates, how you can easily get that information into your ATS. And anything I can do to just avoid opening multiple pages and clicking, even if it is so easy as clicking as a button, sometimes, hey, I did the work, I did the Boolean search, I qualified these candidates now, just put them into my system as easy as possible, right? So I like that for that reason.
And again, using these APIs, AI, these large language
use these large language models
to then read the text and then put the information and understand where the information needs to go since sometimes natively in the tools that we’re using, we can’t do certain things.
And I know we touched on using other use cases for using AI to decrease bias. But something that I’ll call out here is also considering that English is not everyone’s first language.
And I found that using this has been helpful for many people who are learning a new language, they’re maybe working in a world that it’s just not their first language, the culture can be a little different. I realized that it’s a very common use case for people who English is their second language.
So I just wanted to call that out. And then these are more specific and you guys will get access to these slides, but these are some… Hold on.
Oh, Destiny?
Oh, yes.
I just want to go back a slide. I think you touched on this a little bit before, but someone did have a pre- submitted question. They said, I’ve been hearing a lot of concerns about how AI and recruiting can lead to accidental biases, i. e. an example where AI filtered out female applicants. Just curious to hear your perspective. I know you mentioned different ways to change the job description, but maybe any quick, actionable thoughts on this?
I think that’s a great question.
Thank you for submitting that.
So my opinion ends up being this. So I have read a lot of different studies on this where it’s gone both ways.
I recently saw one that was very popular on LinkedIn where it was called out that white men were basically filtered out or less considered for HR roles, which I didn’t have time to look and see if maybe this is just a very feminine, female- heavy dominant industry. I don’t know.
But here’s my opinion when it comes to recruiting. The truth is most of us are not… If you’re looking at ATS software as it exists today, none of it is as sophisticated as people want to believe. For example, with Ashby, I can auto- decline candidates only one way. It’s based on how they answer yes or no questions. So if you are unqualified and you click no, you’re going to get auto- declined. That is not something that’s able to filter out women or men unless the question in the application, which is illegal, is are you female or are you male?
I like to just remind people to think about the tools that we’re using.
Most of the tools…
We’re not really using software that even has the ability to do that. It’s not even able to decline people based on what they wrote in the interview forms. It’s just yes or nos.
I would say that it’s most important to make sure that we ourselves are not using tools that are actually doing that.
We’re just not.
Now, at least I haven’t seen any so far. I do know that some people use Bullhorn, which Bullhorn has for a very long time had a score feature that assumes how fit you are for a role. That is something I do want to look further into because I don’t have experience working with Bullhorn because Bullhorn is usually more of either agency or really large company type thing, probably bigger agency type thing.
That is something that I do want to look into that I don’t have enough experience with.
What I will say is something that I’ve used to do the opposite, to filter,
in, because we can use this to filter in or get more candidates that, you know, for example, I was saying earlier, I was struggling to find a, or get women on this sales team. I’ve also done a lot of engineering searches, and we’ve all probably been there where we’re trying to find more female candidates that are engineers, right? And you just run dry when you’re just like searching software engineer, right?
And you’re just like searching and you’re just sifting through like some websites and some of these tools, you can click, you know, it’ll say, say, like, I think, like, I don’t know what they call it. It’s like DEI is basically like what they’re calling it or like diverse candidates. I don’t know if tools are specifically allowing you to search for women. But if you do want to search for women, what I’ve done is I’ve used, it goes back to that Google search thing. So you can get really creative with this.
So you can look at popular names or least popular names for, say, black women. And I’ve used that to then lead my search so that I can find more black women to put into my interview funnel. You cannot hire more women and more black women if you’re not interviewing them, right? And if you’re not seeing them come inbound, and you’re not seeing them come in through your referrals, and you’re not seeing them come through your agencies, the only other thing you can do is go out there and source other women, right?
So I’ve used that to come up with different lists that help me make sure that, hey, if this funnel, like, and sometimes you won’t catch that. It goes back to some of these reports. Like I can only know if I’m, if I have an even pipeline, because ultimately we’re going to hire the best fit for the role. We’re going to hire the most qualified person, right? But if my pipeline, which sometimes you won’t know, like, I’m not always seeing people, right?
But sometimes I’m on these searches where I just see a name and I see an email and I’m reaching out to people. And I’ve been able to use these reports to realize, oh, oh, I’m only, you know, submitting men for this role, right? So I think it’s interesting if we take a step back and think about how we can use it to catch ourselves inherently being biased or inherently just providing a really imbalanced pipeline. Because if I didn’t set that alert, I would just have to, you know, wait and hear it that my pipeline is, is really just leaning one way.
So I would say, I would say that because I haven’t personally seen any software or any ATS right now that’s really qualifying and like auto declining on a sophisticated level. That’s really just outside of those two things. And then again, Bullhorn is probably referencing their resume to the job description. And I’m not convinced that that’s how we evaluate candidates because on the backend we have that scorecard that candidates don’t see. And I’m not sure if Bullhorn is referencing like a scorecard versus a job description.
So I hope that, I hope that’s helpful, but I do want to research the Bullhorn thing. So if you connect with me on LinkedIn, you’ll probably see me rant about it when I, when I understand it a bit more.
Yeah.
So these, these slides just have more specific Boolean or not Boolean prompts, because I know sometimes people are like, but how, so I like to just give a little more specifics on just different prompts I’m using when I am starting a role, doing outreach for a role, writing job descriptions and yeah, just, this is more like interview questions and stuff like that. We’re almost at the end. other points, which again, using the tools around us to search documents and get answers to questions.
Like I again, use chat GBT to read help docs for me because you know, not only are we the only recruiter, we’re also setting up all the systems and then we have small enough accounts that we don’t qualify for our own CSM at the other companies. So we just, we’re never qualified for help. Right. I use this to read the help doc for me because it’s like in its simplest form, how do I do this? What do I need to do? Or why am I getting this error on this project or this thing that I’m trying to connect?
Also I’ve found it very helpful to take that scorecard and take the things that I know that hiring manager’s looking for and using it to better pitch a candidate because you.
the hiring manager really prefers people with five years of experience at these kinds of companies. Similar thing, you have a notion template. You say, I’m pitching these candidates, pitch them in the best way possible for the hiring manager. So I was recruiting for this ag tech company and I don’t know the farming world, like I’ve recruited in tech for so long.
This was super helpful for me to use because it would know and be able to highlight the fact that these candidates had experience at these big 10 farm companies that I wouldn’t have been able to identify and highlight to then get my candidates hired, which I found quite helpful because then you’re bringing up and highlighting what’s most important to them, the hiring manager who you need to sell to versus yourself. And then, yeah, we’ve mentioned feedback, getting yourself improvement, formatting and stuff like that. Yeah, so two final prompts here.
Again, just reading your transcript and giving yourself feedback. Of course, some people might have access to Otter and different tools that are more sophisticated and are built in and can already do that, but some people don’t. So this is something that you, or a sample prompt that you can use. This one here is showing how you can use to add a little empathy into your work. You know, I’m a very empathetic person, but you can only find so many ways to say, hey, you’re not getting the job.
So I remember this one time I interviewed a candidate who knew that this role was going to be a full- time finance role at a company. And then two interviews in decided that they were going to start their own company and wanted us to be, this to be a potential client. So I wanted to just find a soft way to be like, hey, like super excited for you, but this isn’t a fit, you know, but I didn’t have the way and I wanted to respond quickly and not think about it for days, right?
So I was able to use ChatGBT to come up with an empathetic way to be like, hey, this is not a fit, but good luck out there because, you know, again, we need to communicate with the candidate and some people can just sit on communications for a while because they don’t want to say the wrong thing. So have someone else say the right thing for you, you know? And again, you’ll have access to these, but I like to just be as specific as possible. So I did want to provide, you know, just different ways I’m using this.
Like something that people may have, may come into is Boolean string character limits varied by platform, but I want to be able to reuse my Boolean strings. What you can do is in Google Sheets, this example is building them out in ChatGBT directly, but you can do this same thing in Google Sheets. But here, what you can do is create your master Boolean search string that you can use in say something like LinkedIn recruiter. But then if you’re just using LinkedIn regular, you can use a shorter one.
You can have that second column be for LinkedIn free, right? You can have the third one formatted for if you’re doing a Boolean search directly into Google, if you’re doing one in GitHub.
Without having to figure out how many characters your Boolean string is, like I just am always looking for ways to just reduce the amount of, you know, manual work that I have to do and Boolean searches have just saved me so much because I’m using them everywhere, including my ATS. So I like to just find ways to automate them so that I can, you know, just get that role and search going. And, you know, here when people are talking about DEI, I like to talk about how, again, we can use it to filter people into our process.
And here I’m making a list of HBCUs, historically black colleges and universities. And I’m listing them by state because in this one I’m saying I’m maybe looking for people in the West Coast or in a certain time zone. So if you maybe graduated from schools in a certain area, maybe you’ve got a job in a certain area. So this is me creatively trying to find people to then source for a client of mine. And then again, working it into Boolean strings, this is me looking, taking it to the next level.
I’m looking for someone with influencer marketing experience in Florida who went to a historically black college, because ideally that would, again, help me find more people to balance out my pipeline if I’m looking to hire more women of color or black women. Here again, just taking it to, you know, just continuing, continuing.
And then this is me asking Chachi VT to then turn all the information that we have found in this long search that we’ve created, format it so that I can copy and paste it and turn it into a CSV, which then goes into Google Sheets. And then I can use whatever prompts and things that I want to do in the future. If I’ve connected OpenAI’s API, or if I just want to copy and paste, because there’s nothing wrong with copy and pasting if that’s what you want to do. And then this is just showing how that ended up in Sheets, what that looked like.
And, let’s see.
A few more quick ones. This one here is just showing how, just talking about a common problem when people talk about getting too many candidates, just talking about how there’s different ways you can use Boolean searches, alerts, manually getting those, or sorry, automatically getting those alerts in the different places that you want. The second someone qualified applies for a job, you can have yourself notified via Slack so you can reach out immediately so that you can ideally get that person on the phone as soon as possible, right?
So just doing whatever you can with the tools that you have is really the call out here. And let’s see. And then this is just talking about if anyone had questions on Boolean searches, how I like to just do very specific ones, then I get a little bigger and then I get super vague just to find more candidates and just widen the talent pool and how to do that. And let’s see.
Boom.
That’s it. This was just reminding you that, you know, document things as you learn them because someone else in your team will join and it’ll be easier to get them on board or maybe if you do chief of staff type work, getting other departments involved and excited on use cases for AI, and because you’ll find a million of them and it’ll be easier to share those ideas with your team.
Creating some kind of brand guide could be helpful so that any AI tool that you’re using can reference what the tone and brand of the company is and what you guys are looking to do. And just a reminder that, you know, it takes time and effort to, you know, implement these things, use these things and, you know, become quite comfortable with them, but it’s super worth it. So I thank you guys for taking the time to, you know, learn more, hopefully learn more about AI in recruiting.
And with that, I’d love to hear any questions or ideas or, you know, challenges you guys currently have.
Thank you. This was super helpful and just real specific too.
So appreciate all the examples.
Thank you.
Yeah, it’s a good reminder that we have these tools and we should, you know, leverage them a little bit more than sometimes just like it’s so easy to go back to your old habit of like, this is what we used to do, but putting the effort into your point and like documenting it can accelerate the future.
So really helpful.
Thank you.
Thank you.
I saw the comment. It says, the question, it says, how is…
Oh, sorry, Destiny. Yes, I just wanted to let everyone finish, but we had another pre- submitted question. Someone said, how is AI being leveraged during interviews to better screen candidates? So maybe if you just had like top three actionable takeaways that you could recommend for that question.
Yeah, I would say the way I really enjoy using it during an interview is, depending on how your interview scorecards are set up. So some people, it sounds like you guys might not have an ATS just yet. Something that I would like to do for, or I enjoyed doing for companies that were at that stage is in Notion, I had the questions that I was asking each candidate and that’s where I was documenting them, right? What I would like to do is at the bottom, I had a little button that I could press and it would call out all the questions I still didn’t ask.
We know that sometimes there’s those questions that we don’t wanna ask. So it’s just like a friendly reminder that, hey, you still did not ask those five questions, but on a more advanced level, there are companies out there that exists where when you’re interviewing candidates, the questions pop up on top of the interview and it guides you, it transcribes everything, the bubbles disappear when you ask the question and it alerts you when maybe you’re speaking too much.
If there’s like a certain, I know in sales they have something similar where it’s like you want someone to be speaking a certain amount of the time that exists in real time, but a lot of it at its core is really using AI to prompt you as the interviewer and to transcribe your responses versus having you type it.
Any other questions from the group? All right, well, appreciate the time Destiny agreed with kind of Laura and Maria’s sentiments. This was all like really helpful and specific enough that there’s stuff I’m gonna start doing today, which I always love when I join these round tables. So, and as a reminder for all of you, you should have just gotten a poll that came up on Zoom just to rate the session and talk to us about what sessions would be helpful for you in the future.
Slides
Key Takeaways
- Accountability through Feedback: Giving stakeholders feedback that aligns with the evaluation criteria for candidates helps hold them accountable.
- AI for Sourcing: Using AI to source candidates from various websites beyond LinkedIn, improving efficiency in sourcing and outreach.
- Internet Search Function: Leveraging ChatGPT’s internet search function to find and extract candidate information from websites, streamlining sourcing efforts.
- Document Summarization: Using AI to summarize documents, including interview questions, Glassdoor reviews, and ATS candidate surveys, to derive actionable insights.
- Automation for Formatting: Using AI to format documents, such as offer letters and CSVs, for easy upload into the ATS, reducing manual work.
- Reducing Bias: Using AI to ensure diversity in candidate pipelines by creatively sourcing candidates from underrepresented groups, such as using names associated with certain demographics in search strings.
- Empathy in Communication: Using AI to craft empathetic responses, such as rejection emails, to candidates, enhancing the candidate experience.
- Training and Documentation: Documenting AI processes and learnings for onboarding new team members and sharing best practices across departments
- Interview Guidance: AI can assist in guiding interviews by displaying questions, transcribing responses, and providing real-time feedback on interview dynamics.