Hey there, it’s Felicia. Let’s record a video for today’s data digestible. I want to give a shout out to Blair Fanbro, who presented this morning at TA week and talked about GPT and did a live screen demo. That inspired today’s idea, so thank you, Blair. Today, I’m going to put a prompt called a One-Shot prompt. I’m going to put this prompt into both GPT4 and one of the newest competitors on the block, the AI block called Claude. I’m going to put the prompt in each of the large language models and see what results we get to determine which one is best.
This is something that I have not done before, so hopefully all systems are go. I’m going to go off camera and walk you through my prompt here, and then we’ll hop right into it. Great! So, I chose a position on LinkedIn that had zero applicants. It’s the role of dermatologists out in Roseville, California, which is an area I’m not familiar with. I am doing this just to demonstrate that you can do this with any role, but the point of this demonstration is to see how AI can help recruiters analyze and gather data more swiftly and comprehensively than we could on our own.
Let’s test it out on this role that definitely needs a little love. There have been no applicants over the past week. I’ve copied and pasted the job description and built out this One-Shot prompt. I’m asking GPT and Claude to act as an internet researcher with deep roots in Talent sourcing, Talent attraction, and talent analytics. Using the job description, I’m asking them to create a comprehensive market data analysis. This analysis will be used by the recruiting team working on this region and this role for the first time.
I said to use the internet or any other resources at your disposal because in chatGPT4, I have three plugins already affixed to that chat. So, that’s what I’m hinting at there. I’m expecting to see items like the ones below, in addition to anything else relevant. I’m expecting to see competitors in the space, similar job titles, possibly Boolean strings, BLS market data, online forums where I would find a candidate, and even a potential shortlist of candidates that can be found on the open web or niche websites that can be used in a future sourcing strategy.
I loaded this One-Shot prompt up to see what it can handle, what the AI can handle, and what it’ll bring back. Finally, I’ve asked the GPT and Claude engines to bring back the data in a digestible format that can immediately be shared with a stakeholder, maybe with a hiring team. The tone should be professional, friendly, and competent. I’m going to ask a question in the final analysis and let GPT or Claude interpret that as they will.
Let’s see. We’ll go in alphabetical order. We’re going to go with Claude first. I’m going to select this prompt, hit copy, and there’s Claude. If you’re just now meeting Claude, say hi Claude. I’m going to rename this one ‘Data One-Shot Prompt’. So, I’m seeing it go, and as you can see, it’s already given us some data on Roseville competitors and job titles. That’s interesting. We’ll look at it more in detail by giving you a real-time look at it.
BLS data looked interesting. Alright, I’m actually going to copy this and let you review it directly. I’m going to put it in a Google Docs so you can let me know and maybe vote on which one did a better job, Claude or GPT. Maybe I’ll include a little poll or something on Minty so you can vote on which one or you can comment below on which one you think did a better job. Once you’ve reviewed it, let me know. Alright, let me go to GPT4 with the requisite plugins, copying and pasting the same prompts, and let’s see what we get here.
Alright, GPT is really typing. Alright, it gave us some Boolean strings in an almost like a terminal little window. It was online forums. Already, it’s formatted pretty well. I like the bold, the bullets, underline, hyperlinks. Okay, what I’ll read here will be the question that the AI gave us, and then I’ll leave the link to both responses below, and you can comment on what you think.
GPT asked, ‘Would the hiring team be open to hosting a webinar or informational session on advancements in dermatology to attract potential candidates and showcase the organization’s commitment to the field?’ That’s a really great question. Let’s see what Claude asked. Claude asked, ‘Does the sign-on bonus and benefits package seem competitive enough to attract top talent?’ Well, I don’t know. It probably depends on what data you found. Like the median salary here is $381,000, and the job pays $477,000. So yeah, absolutely, it’s well within range, I would think. But that’s just looking on the surface. Take a look at these two, let me know what you think about the One-Shot prompt, and then comment below with your thoughts. Bye for now.
