The Full Desk Experience

Matt Strain [00:00:00]:
One of the first things I see for managers and leaders to be successful in AI is really around curiosity. Not just about AI, but are they curious about new business models, new ways of working? Are they willing to dive into AI and experiment with AI in the various parts of the recruiting chain that we just talked about? And that's probably the number one characteristic that I think is going to be important moving forward for the vast majority of employees.

Kortney Harmon [00:00:27]:
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 sessions is packed with valuable insights, expert discussions and actionable strategies to help you drive growth in your business. Whether you missed 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. Stay tuned and let's dive in.

Kortney Harmon [00:01:21]:
I am so super pumped that all of you have taken your time to be a part of this half day virtual conference. It's like two half days where we're really deep diving into how AI is completely transforming the recruiting game. This isn't just one of those other boring webinars. It's going beyond the basics today and according to Matt's presentation, you're going to be on your phones. So we have an amazing lineup of speakers and sessions that are really hands on, really talking about cutting edge, what's new, what you need to know, maybe some additional tools you could cover. Today we have Matt Strain talking, Benjamin Mena, Greg Benedita. Greg, I hope I said your last name right. We also have Chris Hessen today from Crelate going to show our agents that are built in.

Kortney Harmon [00:02:05]:
Tomorrow we have Mike Wolford myself. I'm going to be talking about KPI Hamster wheel in the post AI world and what I'm most excited about. Not that I'm not excited about everything. We're going to have a panel of owners and leaders that are currently running recruiting desk now and what they are doing with the world at large when it comes to AI. So be sure to make sure you're registered for those sessions as well. I would love for everyone to continue to use this chat function, the comments box. Feel free to chat the entire Time. We want to hear your questions, your thoughts, your hot takes.

Kortney Harmon [00:02:41]:
So this is interactive. Do not be shy. We're here to learn from each other. So this is actually our FDE Plus. We are all about AI recruiting evolution. And so excited. Our first speaker, Matt Strain, he's a seasoned technology leader, AI educator. I actually got to see Matt speak at AESC a few months ago.

Kortney Harmon [00:03:05]:
I don't even remember what month that was, Matt. Was it April?

Matt Strain [00:03:07]:
I think it was April.

Kortney Harmon [00:03:08]:
Yeah, it was April. And he is the founder of the Prompt AI, former head of innovation at Adobe. Matt is amazing. So he is going to be talking to us today. From AI curiosity to competitive advantage and how executive search and CEOs can thrive in the age of intelligence. So, Matt, with that being said, I'm going to give it to you and I'm going to stop talking because nobody's here to see me today. They're here to see you.

Matt Strain [00:03:34]:
That's great. Thank you, Kortney. It's a real pleasure. Many thanks to you and the team at crelate. Some of what you'll see today is I'm going to show how we can use a number of different tools, including some of the LLMs. One of the things I love that the team at crelate has done is sort of made it easy to access the power of these LLMs. As we go through today, I'll talk about why I think accessing this sort of the core LLMs is important, but also services like Creelate just make it so much easier to access in a consistent way that underlying functionality. So I'm super impressed with what the team has done.

Matt Strain [00:04:04]:
So, as Kortney mentioned, a couple of things I want to do today when we go through this is I'm going to be talking about AI, but I also want to do that in a way that gets you using AI. So we're going to work this at two levels. One is sort of to talk about some concepts that I think are going to be important for you to know as you're going forward, but also to have some fun with it because I think that's a key part of learning AI. But I'm going to point out some tools and techniques that I think will be interesting for you to work with. So this is kind of sums up, you know, from the word cloud and so forth. There are the people who I kind of liken it to, you know, the high school kid who you put in a Ferrari for the first time, and they're either going to jump in and race off and maybe do all kinds of irresponsible, dangerous things with it, or the, you know, the kid who sits there and gets worried and concerned doesn't know how to drive the car. Right. So there's a whole mix of ways that people are reacting to AI and I think having a balance of those things is the important thing.

Matt Strain [00:04:57]:
But realizing that these tools are incredibly powerful. But there are also a lot of ways we could get ourselves in trouble or misrepresent things or go to go down a path that either makes others uncomfortable or, or presents us in a way that we don't necessarily want to be presented. And in this picture, all the pictures in here, I did with AI. But it's worth noting the door on this car is not exactly what it should look like. So even though we have some incredibly powerful models right now, GPT5 is still doing funny things with the graphics many times. So let me tell you a little bit about who it is you're talking to today. As Kortney said, my name is Matt. My background, I have 10 years at Apple, a range of operating positions there.

Matt Strain [00:05:36]:
AltaVista I bet on the wrong search company early on. I did spend about 17 years at Adobe in a number of roles from marketing and product strategy and then I spent the last eight years in the research lab working on gen AI products. About three years ago I started a company called the Prompt where really my goal with that is to make AI actionable and accessible for what I call real people, right? People like yourselves who have operating roles who are trying to get things done and aren't necessarily technologists who either care or need to know a lot of the details around what's happening underneath some of these tools. What I do is I do a lot of workshops, strategy and advisory session. We've also just started rolling out an AI Readiness Index, which helps companies see where they are with AI and what things they need to do to move forward. My clients, a lot of them are in the financial space, the HR space, the security space within these companies. It's a combination of people who are in operating roles and CEOs and HR teams. As I said, I've also been moving a lot into the executive search and recruiting world, AESC Coaching foundation work with the Conference Board.

Matt Strain [00:06:44]:
My claim to fame and micro fame is I did make it on the front page of the Digital New York Times for a book that I did around how to Make Cocktails with AI and based on Chinese Medicine. So sort of a long quirky story, but that's actually what one of the things that showed me how to make AI more accessible for people. And for any of you who have those roles internally, where you're trying to get the rest of the company on board, one of the things I found most useful is starting with things that are outside of where their core work is and things that are kind of fun and seemingly frivolous in some ways because it allows people to exercise their creativity, to start experimenting in a way that doesn't threaten their immediate work function. So we can talk more about that later. So today we're going to run through three things. Three main areas. One is an AI framework for thinking about AI. We're then going to talk about some workflows within the HR or not the hr, the recruiting and staffing context.

Matt Strain [00:07:44]:
And we're going to wrap up with a discussion on some of the new skills and competencies that are needed to manage and lead in the AI area. And I think as we go through this, I suggest you think about it on two levels. One is what it means for you and the companies you're working for, but also if you are in a staffing role, what it also means for those companies and your clients that you're working for. And this could be whether you're working at the CEO level or whether you're working anywhere throughout a company, there's really a new set of competencies and skills that are going to be new, needed or added to to be successful in the future. So the first one will be called the AI framework. I want to start with a little framing on where AI has come from, because to a lot of people it feels like AI is a brand new thing. But really, when we think about it, you know, AI started off in like the 1950s and 1960s. So in, you know, the early 1960s, out of MIT, there was an Eliza bot that was developed that was really the first chat bot that helped people with emotional issues.

Matt Strain [00:08:42]:
And it was basically an active listening chatbot. But it was. So the people who were in the test for this early chatbot loved it so much they went back to the developers and asked if they could keep using it. So all that to say we had some of these tools working with us and being experimented with for the last several decades. And we've watched over time AI and sort of autocomplete, if you will, as an early version of AI telling us autocomplete and spell check for a word. We then saw it move to sentences and then to paragraph. So moving from paragraphs to autocomplete on pages. I'll show you example in this of some things where I've done, where I've going from text to a website and text to a full app.

Matt Strain [00:09:25]:
And we're now into discussions where there's basically autocomplete for a company. So if you think about where AI agents are going to go in the future, it's a number of years, who knows exactly how far out in the future that we're going to be able to spin up AI agents that do most of the functions that are in a company today. And just for context, the things that enable this are really the advancements in compute. Right. A lot of what's driven Nvidia over the last couple of years, we have algorithms, the science and the programming and the research that's behind these models in terms of both inference and training and the algorithms that go behind it. And probably for this group, the area that's most relevant is when we look at data. So when we look about data, about people and their skills and their competency, mapping and ways of extracting data from what people do or what they're saying or what they write or just what's out there on the web, the ways of using data more creatively and more efficiently are going to be really important for those of us who are trying to identify needs and talent and staffing and skills in the future. So more specific to what you're doing, right.

Matt Strain [00:10:31]:
Why AI is really well suited for this audience. If you look at both executive search and staffing, the research shows that probably 80% of what is done is formulaic. It's operational work or research work that can be well done. With AI, there's 20%, plus or minus. We can argue about what the percentage is that is really that premium work. If we dig into some of these things, you're all very familiar with these, but on the left are some of the things where we still need people, and we will for a while. But there are areas in here where AI can strongly assist you in what you're doing today with these things. And I've talked with companies where they say things like developing candidate characteristics, creating these documents, candidate outreach.

Matt Strain [00:11:13]:
They're compressing this by maybe two or three weeks in some cases what might be done in three to five hours. So there's some work that needs to be done to get there, but there's a lot of opportunity to condense that. And again, I would say, because I had AI help me with this graphic as well, you can see on scheduling where it's still, it's not quite perfect. So the point being, anytime you're using AI, you really need to need to check it. But having said that, I did leave that in there on purpose. And then on the right, the things that are going to be the domain of people, right? So the judgment calls, the persuasion when you're trying to talk to a candidate or a client, the whole idea of building trust and assessing risk AI can help you with a risk assessment, but at the end of the day, because you're going to want somebody, a person who's held responsible, you're going to want a person who is on the front end of that. And reputation for yourself and for the firm is something that people will continue to focus on. So a lot of the direction that is happening is for the 80% to figure out how you can get AI to assist, which is where Creelate comes in.

Matt Strain [00:12:12]:
And then on the right hand side, what are those areas that only people can do and how can we double down on that area? The word that's coming up more and more in these discussions is trust. So both building trust and understanding with the tools we're using, but also looking at the roles of all of us on this call, how can we go from just doing the things on the left to actually becoming a trusted advisor and sort of moving up the chain a little bit to being more of a partner to our clients than just providing some of the things on the left. So let me give you a quick example just to kind of show you what I'm talking about here. And from the initial part, the initial survey, it looks like many of you may know this, but this is a quick example of what I did with ChatGPT5. Came up with a fictitious scenario where I said, let's do a market scan. If Mayo Clinic is looking for a Chief AI and Innovation Officer, what I'm going to do now is click on this and this is going to pull up the report that was generated. And this was with deep research on ChatGPT. So I won't read through all this in detail, but my point being is how we can use this to do a first pass in some of the materials that may have taken us a lot longer to do.

Matt Strain [00:13:20]:
So some of the things that you'll see here, and this is a market scan from Mayo Clinics Chief AI and Innovation Officer gives us an executive summary. Some of the things that it also does now is it does give you source and citation. So if you're coming in here, we worry about hallucination and AI making things up. There's now more citation and sourcing in these. It Comes back with a set of the initial search parameters. So even if we don't tell it what the parameters are, it will come back with an initial set. And I look at this as not making something up, but if you look at this as sort of a prompt or guidance. So what you would then do is go through and say this might not be perfect, but it's pretty close.

Matt Strain [00:13:59]:
So it kind of takes you down this path. Nice to have off limits. I didn't ask it for a talent heat map, but it comes up with a draft of one. So again, this is one where this probably isn't perfect, but it might get you to think about things in a different way or give you a starting point to lead from competitive movements. What's happening in the space compensation snapshot. And again, these numbers may not, they may or may not be accurate, but at least it gives you an idea of what the framing is for looking at a compensation package, right? And then it comes back and it gives you a pre qualified list of candidates. Scroll down through that anyways, remainder pools, risks and opportunities. So I won't go through all this.

Matt Strain [00:14:40]:
My point being is with the end citations, rather than having an intern or someone junior go through and pull some of these things together, starting with this could save you considerable amount of times that you could then go through and edit this and hand it off to someone else to continue on. So one of the things that we, that we talk about with this is sort of three parts of the framework, right? One is we've got to have a mindset for AI. And that's really the beliefs and attitudes, which I think increasingly is one of the most important parts to using AI, right? The tools and the skills you can learn as you have with other things. But it's really that mindset of how you approach it, which you'll see me talking a lot about curiosity throughout this. So that I think is one of the most important things the skill set. We'll talk about that next, which are some of the ways that you interact with the tools. And you'll hear a lot of people talk about prompt engineering and context engineering. And then on the right are the tools.

Matt Strain [00:15:32]:
And those are the tools that frankly are changing every week or every couple weeks. So the tools are important, but you know, I think we will learn how to use new tools if we understand what the mindset is and have a basic skill set. So let me go through that in a little more detail.

Kortney Harmon [00:15:45]:
Matt, I do have a question. I'm going to throw it out here while you're on this. The question is about deep research. So the question is, is deep research just asking ChatGPT questions about a role or competing companies? Can you elaborate more for those that don't know about deep research?

Matt Strain [00:16:01]:
Yeah, yeah. So great question. Deep research is you can ask it to do research on anything, right? So the thing I see a lot of times is people get stuck with like, I don't know what to ask it or how to ask it. And they will spend three minutes telling me offline, here's what I want to do. I want to know this, I want to know the other thing. And I'm like, don't tell me, put that into deep research. Deep research can really do anything in this context. So anything you want to do research on, you can use it for.

Matt Strain [00:16:26]:
And research can be a little confusing because people think it's academic research. But I would think of it anytime you want to deep dive on something and you want to go deeper than sort of the superficial answer you get from the regular GPT, ask deep research. And the other thing you can do is get a little meta with it and ask deep research. This is my area of interest. How can I best use you? And it will give you ideas to do that. Hang on for just a second, I'll show you the first one around the tools. Let me just set a framing for kind of how to think about this. So We've got our LLMs, the large language models and foundation models.

Matt Strain [00:16:57]:
The ChatGPT is Claude Perplexities, those types of companies. ChatGPT is the one I suggest as a go to, to start with. Most of us at our companies or our clients also have the major platforms, things like Microsoft Workday, ServiceNow and these platforms are going to be, I think, sort of adopting AI and they're all integrating AI to a certain extent. But really the companies that get really interesting for me are those specialized companies that build on top of the large language models and sometimes they'll have custom data, custom interface, custom training that they've done on the large language models. And I put crelate in there and I've outlined some of the things that these companies do around the search and talent and staffing space. There's a list of those things here. And these are really those companies that take the challenge of working directly with a large language model, wrap it in an interface, wrap it with some expertise. So for people who aren't AI experts, it allows you to access that power really quickly and easily.

Matt Strain [00:17:56]:
It's a really busy space. A lot of companies are popping up a lot of companies are getting a lot of valuation, so we won't get into what all those companies are. But part of the reason I'm doing this is I think Creelate is one of those companies that has taken a large part of that workflow and has made it really easy and really accessible. So when people are talking about AI, this is kind of how I At a very high level, I suggest you've got the large language models, which I'm going to make the argument are important to understand. You've got the major platforms that your companies are probably on and those are going to get better over time. But I think for the next X number of years, while those major platforms are catching up, these specialized companies are running really fast to do some very, very cool things. If you remember my initial comment on data, that's where some of the data comes back in, because they have access in some cases to proprietary or tailored data that feeds into their services that you wouldn't get or you might have a hard time finding with one of the large language models. Next part I want to spend some time on is the prompting question and both prompting and context.

Matt Strain [00:18:58]:
And this is the slide or the section that people find most interesting. In a lot of cases, it seems the most basic, but it's where people initially get tripped up. And it's really around the fact that search is not the same thing as AI. Right. And part of what's happened is it's part of what made ChatGPT so successful is if you look at these two interfaces, they look almost identical. So the vast majority of people take what they did in Google that worked well and they put it into AI and that really gives you an incredibly mediocre response. And the concern I have for a lot of people who don't quite understand AI is they have that experience. They take their Google search, they put it in the AI and it's okay, it might be a little bit better, but they really don't understand what the opportunity is.

Matt Strain [00:19:44]:
So in thinking about the two, I'd like to share a couple examples of how search and AI are different. Right. So we have the difference between a query and a conversation. And if you take nothing else away, the idea like the question of how do I interact with it. Well, you can't really ask Google how to work best with Google, but I do that all the time with AI. Right. So you can have a conversation, you can ask one thing, it comes back well, can you modify this? Can you explain this? Can you go deeper on this part? So we have query verse that we have links to insights. So, you know, search is much more like a card catalog.

Matt Strain [00:20:19]:
It will direct you to a source. AI will give you that insight, which can be frustrating at times because you don't quite know what the source is. But, you know, AI really is more about interactive iteration. It's more about being, you know, I like the Socratic mentor comparison because it's like that smart friend or your doctor, your advisor or your therapist who they may give you a slightly different answer each day, but it's going to be mostly the same, right? The words and the examples and the metaphors might be slightly different, which is different from search. The results may change, but it's more like an encyclopedia or a card catalog where it's going to direct you to something else. If you're musical, you know, the idea of a symphony is if you're playing a piece of music for the most part, right? Every top tier orchestra is going to try to make that symphony sound very, very similar. Whereas with jazz it might change depending on who the players are. One day, depending on what somebody's mood is, things are going to go in and out.

Matt Strain [00:21:11]:
And that's the way I would think about AI. If you have a typo, an obscenity, an emotional outburst, AI is going to be fine with that, right? It'll change and respond to you as if in a very similar context to how you you engaged with it. So I'm going to skip this part on kind of what the interface for AI looks like, particularly with GPT, because it sounds like the vast majority of people on the call have seen this before, and we'll touch on it a little bit later. But this is one. Even if you have been using it, there are two parts of sort of my quick prompting 101 that I think are going to be interesting. One is I also think of prompting as sort of the art of the conversation. And increasingly I've been thinking about it more like delegation. So if you're somebody who's clear on delegating and telling your employees what you need them to do, you're going to be in a much better spot when we start working with AI or even agents, because really the idea of delegating to a person and how you describe what you need done with an agent are surprisingly similar.

Matt Strain [00:22:05]:
So in doing this, and I would say that, you know, there are a number of different prompt frameworks that people use. So if there's another one that you use or you've been using, that's fine, but Particularly as you're starting off, you want to give it a clear role to play an audience. Who do you want it to write or create for? What's the task you want it to do? Describe the style and tone and then what's the output? Right. So some quick examples for this, depending on talking to. I also like to do it like in the Mad Libs format, right. If you're a little bit on the older side and you remember Mad Libs, you know you could do it in a fun way, right? You are a blank speaking to a audience, your task is blank. So some examples of that for this audience. Right.

Matt Strain [00:22:44]:
The role could be, you are a recruiting manager, staffing office manager, candidate experience coordinator. So you can tell it, do you want it to be a CEO? Do you want it to be any of these roles? And obviously this is not just for staffing, but thinking about what that role is that you want AI to play for you. The more specific you are with that, the more specific it will be in its response to you. The next one is the audience. So you would write and you would want output differently if you're talking to a candidate, a client or your internal team. As an example, the task, again, a random selection of things that you might do in any given day. DraFDEd job description, candidate outreach, email summarized candidate profiles and thinking about AI. Because it has seen and been trained on thousands and millions of examples of this.

Matt Strain [00:23:33]:
When you say you need a job description in the US with certain considerations, it pretty much knows what the framework of that is going to look like. And because most jobs are a derivative of an existing job, it could do a pretty good job, for example, of creating an initial job description. Same thing for tone, right? There are times that you want a professional, there are times you want it warm and welcoming. If it's a disciplinary email, you might want it to be legalistic and straightforward. You could make it funny, flirtatious. So you can go very, very broadly with this. The next one is the format and again telling AI what you want it to do. Do you want a brief email? Do you want a one page briefing document, a slide deck, a script, a three minute script for a video.

Matt Strain [00:24:16]:
And then there are also like other ways that you could, like I call this a twist, you know, keep it under 150 words, right? Use a sports or competitive analogy. So these are things where you can start to apply some creativity to it. So one of the things that I created to kind of help with this is I used Lovable, which is one of the vibe coding sites we Won't get into that now, but basically what I'm going to show you, I have no coding expertise. I did all of this through text. And the idea here is to help you construct a prompt. So what I'm showing you now is this was all built by just my saying. I want to take the prompting framework and to create a way for you to learn better. Prompting.

Matt Strain [00:24:55]:
What I just went through, right. The roles, if you come down here, you can select any of these roles, including create your own custom role. So if we said you are a staffing office manager speaking to a candidate, your task is to. I'm just going to pick these to go quickly. You want to have an instructional tone. The format is a one page brief and be sure to say like you're coaching a friend. So what this may look like is the prompt that comes out of it, right? Your staffing office manager speaking to candidates. Your task is to do such and such.

Matt Strain [00:25:33]:
The point here is when you're doing a prompt, you don't necessarily need to follow this framework, but the more of these things you put in, the better off the output is going to be. So in this case, this is the response that comes back from ChatGPT and what you'll see is it may be too long, but it's a pretty good first start. Then what you would do with this is you would then take this and ask it to be shorter, ask it to change. You would then have a conversation with this, much like you would an assistant to actually have it go through and help you out with it.

Kortney Harmon [00:26:03]:
Matt, while this is happening, can I throw a question in?

Matt Strain [00:26:07]:
Yeah, perfect.

Kortney Harmon [00:26:08]:
So from Mary Lou, I've been using AI for quite some time and I found that any imaging with words is usually wrong, even when it's instructed not to have spelling mistakes, use English, etc. I've asked ChatGPT how to fix it. Answers don't really help. Any suggestions for improving that?

Matt Strain [00:26:26]:
It sounds like you're doing all the right things. It's dramatically better and depends on what you're doing. Right. I've been using Gamma and Beautiful AI for presentations. So if you're doing a presentation that works pretty well. If you're trying to do infographics, there's a company called Napkin AI which allows you to put in and I'll show you example of that in a little bit too. So that's actually pretty good for infographics. And if you're trying to put text onto an image, like with that car that I did with the Ferrari where I had GPT on the side.

Matt Strain [00:26:57]:
Yeah. I went into Adobe Firefly and there's Generative Fill on that and you can use that for free. But when I asked ChatGPT to do a nervous high school kid in a Ferrari and put GPT on the side, it just, it got it so wrong. You know, there's a long technical reason for why it does that, but it's not you, Mary Lou. It's a technology right now.

Kortney Harmon [00:27:15]:
Thank you. I will leave you be. Obviously, people in the chat are saying they love this tool. I was playing with it as you were talking anyway, so thank you for sharing that.

Matt Strain [00:27:24]:
Check out lovable.dev. and if you have any interest in like creating prototypes and that type of thing, it's just joyful. You can take that and then find somebody on the team who's more technical to kind of fine tune it. But for prototyping this type of thing. And actually, so I went in and I said, make. I basically gave it my framework and I said, create an interactive tool for people to understand AI. With this framework, I spent about 15 minutes. I then spent a little bit more time, honestly.

Matt Strain [00:27:50]:
But within 15 minutes I had something functional. So again, I would just encourage you to go and use that. Yeah. And I'm happy to share that link so you can play with it. My request is I'd love feedback on it, what worked, what didn't work, ways to use it. So look at that as a prototype that I'm sharing with you, having gone through the prompt framework. Where AI really gets interesting is when you start pushing it with curiosity. So the things that I love on this, if you ask it, what have I missed? What could I have done better? Right.

Matt Strain [00:28:19]:
And these two, like, I even use this when I do meeting recordings. Like, I love the meeting recorders like Fathom and Otter and so forth. So the things like summarize this and track action items are kind of table stakes. What gets really interesting is when you take a meeting, particularly like a board meeting or a product meeting or an update, and you say, analyze this meeting and tell me what I could have done better in the meeting. It does a really, really great job of telling you how to be more effective. Or, you know, there are companies that do this just for sales as well. But it's really good at giving you feedback on yourself and on what other people can do. So that one, like literally makes the hair in the back of my neck stand up when I saw that.

Matt Strain [00:28:55]:
It also helps if you're doing something. Not everything, but if you're doing Something important, Ask it to tell me the steps you're going to take to do this. So if you're asking us to do something more complicated, this both helps it to load memory, but it also gives you a chance to go in and say, no, no, no, I don't want you to do it that way. I want you to do it this way. And if you have something documented that explains those steps, you can also upload that and ask it to follow the steps. This one I love. Because we all know that these things are somewhat sycophantic and they will tell you what they think you want to learn so that you'll continue to be engaged. I love asking it.

Matt Strain [00:29:25]:
If you say, be direct, I can handle the truth, it comes back with a much more honest response. Or even further, you say, play devil's advocate. Challenge my thinking that also allows you to take and this could be you're presenting to a client, you're presenting to anything that you have. If you really want to refine it, I'll either ask it to be devil's advocate, or if it's writing, I find I get a much better response. If I say, evaluate this as an English teacher, pick your favorite college, review it, grade it, and tell me what I can do better. So you're not asking it to write something for you, but you're asking it to challenge and improve your own writing. Again, sometimes I ask, are you sure this is right? If I'm doing something complex, I'll even say, describe this thing I've uploaded like a spreadsheet to me so you can make sure it's analyzing everything that you think it is. And sometimes when you say, are you sure this is right? It'll come back and it'll be like, oh, sorry, you're right, I made a mistake.

Matt Strain [00:30:20]:
Let me try again. Which is infuriating and like 5% cute. Tell it to check it again. And this also, like, kind of gets in. There's research behind this that says if you tell us to act as if you've just had your first cup of coffee, it will actually do a better job. And the reason for this, as it's been explained, is because this is trained on how people act on things, that it's read on people's behavior. So if you tell it to act as if you've had your first cup of coffee or you're well rested or takes your time to think about it, the models actually do a better job. So I add those in because for me, the prompting framework is table stakes.

Matt Strain [00:30:56]:
And that gets you to 80%. But these types of questions are where you really get to see the magic of what it can do. And you can apply that to just about anything that you're doing. So staffing and recruiting workflows. This next part again I mentioned Napkin AI. So the graphics here, and this is just a beginning part, this is just a tip of the iceberg of the graphics. But rather than trying to come up with some interesting graphic, you can put any of this stuff into Napkin AI. Say, create a graphic gives you like a whole range of different choices.

Matt Strain [00:31:27]:
You know, it's not the same thing as going to a graphic studio, but it's better than just a bunch of bullet points on a slide. So check it out. They just went from being completely free to now it's like a seven day trial. But I think you'll enjoy it. So what I wanted to do here was to kind of talk through the hiring, staffing workflow at a high level, starting with like sourcing, screening, assessing, selection, and kind of talk a little bit about where AI can be applied in that funnel. So if we start with sourcing right job specs. And again, I'll try to present this at a medium level and I apologize for those of you who are using this all day, but the things with job specs that I love is because it's such a formulaic output, it can take a very loose description of your job spec. I'll put a URL to a company in there, examples of other job descriptions that are like it, and it can draft the language.

Matt Strain [00:32:17]:
You can make sure that it has as much sort of DEI related language and considerations depending on who you're going after. You know, you can have a distinguish here, what are those must have things, what are the nice to have requirements. So it'll do a very, very good job of coming up with a draft job spec talent pool. You can go through and take this job spec. Here's where we are, come up with a list of potential candidates. What do they look like? Give me sample like specific individual candidates and we'll come up with that. As you saw in the first example that I gave. And also for outreach, right.

Matt Strain [00:32:49]:
So tailored outreach at scale. So even if you're working, if you're working with a prelate or something like this, these things are built in. But even if you're working with an LLM, I work with a very small nonprofit on their board and to do custom analysis of a donor base, for example, coming up with custom messaging for each of those groups, sort of getting down to hyper personalized messaging that as a small company you might not be able to to do or you might not be able to do it without accessing some kind of an internal system. So really, really helpful for that. Right. On the, on the sourcing front. And I'm just, this is just the tip of the iceberg on the screening side, right. I think it's very interesting to be able to, to look at profiles.

Matt Strain [00:33:28]:
You know, there aren't that many companies right now that can access LinkedIn, but if you start tracking, you know, people across all the sources of, of news and social, what's happening with company, you know, moves, where are they in the news? Are the titles changing? So there's some very interesting ways to kind of kick off screening with that. One of the key things is like what does it look like to manage in the AI era is I think there's going to be a lot of cases where there's a job opening or a need, but there aren't any candidates who are perfect right now because the need for what is happening with AI and the speed at which things are happening is so new. So I think we're going to have to find candidates that have skills that may not look like a traditional match, but maybe that there may be competencies and things that they've done in the past that match the need that we have. So somewhat convoluted way of saying that, but I think looking for a match between non traditional skills and the thing that you're looking to hire for AI can be a really nice job of bridging those gaps. I've done some work with high school kids who say they don't have any experience if they've cut lawns or watch people's kids. How can you take those experiences and translate that to a target job? And I think that's going to happen all throughout the organization. The evaluation is one that we've seen a lot of movement on. So if you're in an organization that is perhaps smaller, doesn't have a traditional rubric, has one person who's developed and hasn't shared it for evaluation rubrics, evaluation criteria, those types of things, AI is just excellent.

Matt Strain [00:34:53]:
And again it is because there's a pattern that it's used over and over to see this, so it can draw on that data to come up with something customized for what you have. The assessment side. Right. This is a really interesting space too. So there's a lot of work being done around video and audio analysis of interviews, using that for screening, using natural Language processing to evaluate sort of a discussion or an interview and finding some sort of counterintuitive nuggets that are in there with that. The thing that comes up, and there may be people speaking on that today, are some of the legal and regulatory landscapes on this. So there are definitely some issues around bias and also social appropriateness of the video and the audio interviews and analysis, but that's for a different conversation. Some very, very interesting things on the psychometric side.

Matt Strain [00:35:40]:
So, you know, using gamification and test to predict a fit or specific management styles. And you know, there are companies, I think it's Moderna and Unilever that have been using this for quite a while. So some really interesting things that have been going on for a while but are starting to get a little more traction. And then the selection side, I think the idea of consolidating information across the different candidates, one of the things, at least in the executive side is I think there's been a lot of bias in the ways these decisions and offers are made. So AI provides us for a way to look to reduce the different types of bias, to benchmark across all the different candidates and to quickly do that. So I think this is a place where clearly that human instinct and the one on one relationship is going to be absolutely key. But there are things that I think would augment that. So this falls into the category of using AI to support what people do and what humans do and humans do best.

Matt Strain [00:36:34]:
But I think anything that is like looking across multiple candidates, doing a comparison of one candidate to another to help a client determine who the best person is. Lots of interesting ways to use AI there. We've talked about bias and I think this is another one where as things change in the industry, we're going to see a lot of perhaps non traditional candidates coming in. So I think that's an area that will be good to think about with AI and even background checks. So I've talked to a number of companies that are using AI to look at a whole number of data sources to try to just get one or two signals, not necessarily to do the whole thing, but to try to find a signal that might come out of a YouTube video or from a post that will give them some indication of something that they should poke on a little bit more with the background check. And so that's sort of the work stream of staffing. And then I think there's a whole operational side. And this is where Crelate has some things that I think may be different from some of the other companies.

Matt Strain [00:37:28]:
And you know, this isn't necessarily just specific to staffing, but it's taking that those back office functions through a staffing lens, which I think is a huge opportunity. And I would say from the companies I work with, by deploying something like this, there's probably a 25 to 40% increase in just productivity that companies are seeing. So anyways, I want to spend a little time just talking about the competencies that I think are going to be needed for AI in this new era. And in a lot of these cases, when I've talked with people in the particularly the executive search world, you know, there's a discussion as are they new competencies or are they extensions of existing competencies? I think in many cases they're existing competencies that are going to need to be sort of extended in some new ways. So one of the first things I see for managers and leaders to be successful in AI is really around curiosity. Not just about AI, but are they curious about new business models, new ways of working? Are they willing to dive into AI and experiment with AI in the various parts of the recruiting chain that we just talked about? And that's probably the number one characteristic that I think is going to be important moving forward for the vast majority of employees. So anyways, any ways to identify and cultivate that are going to be important. The other thing we're seeing which is critical is navigating through uncertainty.

Matt Strain [00:38:49]:
And when I say navigating and communicating, I mentioned we do an AI readiness and activation survey. One of the key things there is that senior leaders and managers are concerned and worried. They don't know what the future is going to look like. And they're very uncomfortable talking about that. So the ability to communicate clearly and frequently about the uncertainty and what they do know and what they don't know is going to be a skill that's more important than it's ever been. Experimentation and resilience. And this is responsible experimentation. But we're still seeing a lot of companies that are frozen.

Matt Strain [00:39:24]:
They're afraid to tell employees that they can use AI. They don't know which version to use. They're worried about risk and private information and privacy, which are all very real and important concerns. But there's a balance between communicating what you can do, picking the right tools and allowing people to start experimenting and sitting back and doing nothing. And I'm not saying that's an easy decision to make, but it's a decision that leadership teams are going to need to get comfortable with to move forward. And this is the other thing that I, that I'm seeing is, you know, the ability to delegate. And we talked about this in the prompting, but I think any manager who can delegate well to a human is also going to be able to delegate well to an AI. So the comparisons that people have about AI being like a new employee from a different culture or an intern that's coming on, really what they're talking about is the ability to onboard a human or AI to provide them with training materials, to provide them with regular feedback, to give them small nuggets of responsibility, see how they do and expand it.

Matt Strain [00:40:27]:
So both of those things are what you would do for a new employee. And it's also the same thing you would do for AI. And when people talk about agents, it's very much the same discussion. So if you have somebody on the team who is great at delegation, that's the same person, even if they're not technical, who would be a very good candidate for working on some of the AI initiatives. And probably the most important thing I see as we move forward with AI is the ability to connect to other humans. I say at scale, but I mean, that's the one thing that only people can do, right? Increasingly, AI is going to take over more of the things that we talked about earlier. But really identifying those things that make each person and each of us uniquely human and uniquely competent in our role is where we need to focus. So the advice that I would give to everyone is think about those things that only you or only a human can do, and those things that you and your candidates are hired for, focus on those and then start thinking about some of those other things that you are, that you do, who are done on a day to day basis.

Matt Strain [00:41:31]:
And where can AI start assisting that? Because the recognition that there is a way for it to assist is going to be absolutely critical. So that's basically what I wanted to share with you today. I know I ran through a lot of it quickly, but I think all of those parts are important. So we talked about the mindset, tool set and skill set. You know, I think I've mentioned curiosity about 20 times, but I think that balance of curiosity, experimentation and resilience is absolutely critical. And just being aware that there are some new competencies that are going to come in and many of those are probably extension of existing competencies. But looking for ourselves and for our candidates at how we can stress those things that are human and supporting them with AI and some of those things that may be uncomfortable to transition to AI assistance. But again, going back to sort of Crelate.

Matt Strain [00:42:20]:
That's where a company like Crelate can come in and help ease that transition and provide a platform and scaffolding to make that happen. So thank you for racing through that with me, but that's what I wanted to share. So I think we have a couple minutes for questions and I would say, Kortney, also, if there are questions that people put in the chat, if we don't get to them, I'm happy to follow up with them with you on email.

Kortney Harmon [00:42:40]:
Amazing. I appreciate that. We'll give everybody a minute. If you want to add any chats, feel free. Feel free to email Matt or in the QR code. This is how I followed up with him post AESC and how I got him here. So thank you very much, Matt. I love it.

Kortney Harmon [00:42:56]:
Curiosity is the word for how we need to handle and how we're going to be changing and when our processes and our tools and everything that's happening in this world. So thank you very much. I jotted a few things down. I've never heard a napkin, so thank you. I personally took a few pieces away from this as well and I feel like I am in this day in and day out and it is ever changing. So thank you so much for your time today. Someone had asked about an AI101 course so I would encourage that person to reach out to Matt to get more details on. If you're interested in more details around, learning more around like a one on one or a prompting course, I believe Matt does more in that area as well.

Matt Strain [00:43:40]:
Thank you. Yeah, and there, there are a number of those out there. However you do it, if you're doing it for yourself and for your company. I really love an in person session. Right. So you can do it online but having like the people in the room working together because ironically learning about AI is it's a team exercise and if you have an AI buddy to do it with and somebody who can help you and help you, encourage you, that is one of the most important things to do. And I would also just say start with something that is low risk and low stakes. So if you go after like optimizing compensation for senior execs is your first bet out of the gate.

Matt Strain [00:44:14]:
There are all kinds of pitfalls to that. So start small and build up from there and I think you'll have a really fun kind of joyful ride.

Kortney Harmon [00:44:21]:
I love it. Matt, thank you so much for joining us. I sent you those screenshots. This is also available via replay for anybody who wants to share show their team or share a session. So thank you very much, Matt and I hope you have a wonderful day. Thank you for your time.

Matt Strain [00:44:34]:
Thanks Kortney. Really appreciate it.

Kortney Harmon [00:44:38]:
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 miss 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. Together, we're building a community of growth and learning. Until next time.