AI Agents for Recruiting Hands On using Claude and OpenAI

Thanks for joining in to our weekly Workforce Monday seminar. And in this webinar or seminar, whatever you can say, that we are today talking about talent acquisition and workforce entry AI agent scenario catalog. So a lot in the previous sessions. So one of the previous sessions we talked about the different types of AI agents we can think of. That is something we have discussed. Okay. We will continue with that today and we will discuss what are the other possible scenarios and how do you create these agents also. Not only specific for HR point of view, but in general I'm talking about which we'll be talking about today, right? So let's talk about those details, and in our detail I will cover of those things. And in general, these are all general topics, okay? From whether it is a SuccessFactors or it is the Oracle or Workday, or in general if you're talking about, this will cover everything, right? So overall picture we will looking at, and that is what exactly we will talk about different hiring AI agent point of view, right? And many of the AI agents already SAP is offering that. And just, we like to understand what exactly, how we can design this. We will talk about that today. So we pick up one scenario from a talent acquisition and workforce entry. That's a kind of an onboarding point of view. We will see what are the possible AI agents we can create and what are the possible scenarios. So there are multiple domains. You can see talent demand, employer branding. Sourcing, we have application intake, then we have a screening and interviewing, and then selection and offer letter and workforce entry, recruiting operations, recruiting analytics. Okay, these are the possible scenarios we can have, and based on that we will start creating this AI engines, right? So out of this, what exactly we will try to do today is a few of those things we will do prioritizing, and based on that we will be creating these engines. Okay. So with that note, let me then get started. So let me get started and we will move further. So in this, we will be leveraging today the Claude, okay? And to create some of these agents and we will see how does it actually work. And based on that, we will take it forward. Okay? So now let's then get started and I will be sharing those details here and we will go further accordingly. Okay? So let's then get started with the hiring demand insight. So what does the agent will do? So it will identify emerging hiring demand from the workforce, plans, attrition, capacity gaps after projects, and vacancy trends. And based on that, we can use Claude or we can use ChatGPT also. Okay. So let's say I will try to use both. Okay. So let's see how does it work. So let me share this link with ChatGPT and, and as well as we will be going to Claude. So I will go to Claude first. So let's say I will go to Claude now and go to the ChatGPT as well. So I will say that I will go to the, what, my project. This is my HR transformation project. I will say that I'm having a series of AI agents to be created. And so let's start with the recruiting process, the talent acquisition workforce entry and AI agent scenario catalog. And here, the first agent to be created using Claude. Then we will go for the others. So this is a scenario I'm having. Let me try the same thing with ChatGPT as well. So I will, for this, I will go for HR transformation. HR transformation do not have directly. I will go for in general. Okay, I will be creating here GPTs. Okay, they have an option, I can create my own also, right? So yeah, so I'll be creating, defining your GPT creator, right? So follow this, then I will, let's see how it is doing. So the list of AI agents, I will be giving this. This is the list of AI agents. This is a big list, but we will start with the first one. First, Hiring Demand Insight Agents. We will start one by one and we will go here as well. So let's see, based on the thing, both the things, I will try to see that. So ChatGPT as well as— so it is— I'll use that Satya Nadella, the Codex Index Coding Agent that helps you build. It's okay. Does the name— a good study name would be Hiring Demand Insight Agent. What is the name of the agent? Hiring Demand Insights. Okay. Does that name work? Only favorite song, Headcount Planning. No, I do not want it. So I want this one only. Go with— this is the name of the agent. So what does it say? It will generate— I would generate the initial profile picture with simple profession like Commerce, Office Planning, and Demand Forecasting. After generating so and so, I will start with your function description. So and so, identify emerging trend. Please try the GPT in the preview panel in the right-hand. I can try to make a few changes for the next requirement. What planning methodology should it follow when forecasting? For example, with emphasizing best practices, AI-driven own planning. So it is saying to preview for that matter. Okay, so this is okay. It is actually asking question to me, right? So I will follow. Do not go by any product specific, although always follow the principles, the best practices, leading practices of the industry. And based on that, you create it and so that it can applicable for SAP SuccessFactors, Workday, or Oracle as for the demand or as for the need of the customer, depending upon what they're using. It should be AI-driven predictive analytics, skill-based workforce planning, as well as it should be given option that I can also add my own planning framework, right? So I have given my details. You can see that reading practices. Yes. Yeah. This is what exactly I told. To the thing. Meanwhile, I'm also checking with Claude as well what it has done. Hiring demand insight as well. It has brought in everything. It will create— okay, it has given a different approach, you can say that. I'll create a 4-part delivery: architecture design document, technical specification, then implementation guide. I'm really happy about it. Claude agent prototype, React artifact, interactive agent simulator, plus sample scenarios. Let me build this now. Okay, so it is building. Let it build. Because the approach is different here, so it is focusing on documentation first, which I really like it in Claude. But in ChatGPT, so we will see that what exactly they're saying. That's an excellent direction for GPT. I will incorporate those principles in its behavior. The hiring demand insight agent will now remain vendor neutral. AI-driven predictive analytics, applies skill-based workforce planning, supports scenario planning, hiring phrase, more conservative, expected, B framework, extensible, clearly distills. Please try it in the preview panel as we continue refining it. I can continue running it until we have my questions. Who is the primary audience of this GPT? Okay, so the GPT, the preview panel. Okay, so the preview panel I'm already looking at. Okay, so preview panel, start by defining your GPT. So what I will tell, we are hiring for a retail sector in US and Canada, and it is very much unpredictable because of high attrition rate, especially in the seasonal, seasonal factor also there. So how can I predict any of the hiring signal in my current organizations? We are having SAP SuccessFactors Smart Recruiters. Okay. Okay. So this is the preview anyway. It is not being having it like this rather. Okay. It has to be dictated. So I already done that anyway. It's okay. So anyways, that is something we will do that point of time. This is not the way actually I was thinking it will be submitting, but it is not submitting the preview way. But let's see. This is not the preview. Okay. It's okay. So now what I will do, I will be, this is not in this way it is working. Rather, I will be putting at a point of time who is primarily the primary, the primary audience for this is the workforce planning leaders, the talent acquisition leaders as usual, and your finance planner, recruiters, hiring managers, HRVPs and CHROs. All of these possibilities are there. So it is creating all of these things. So yes, whenever it is taking any decision, it will prioritize regulatory compliance, fairness in workforce planning, data privacy, cost optimization, speed to hire, etc. Also explainability. Okay, so these are the things I have given. It should behave like a strategic workforce planning consultant, ask probing questions like which industry, which sector, what is the attrition rate, what is your pain point, just like a Big Four or MBB consultant. Okay, and before making any recommendations, provide executive-ready reports by default and present concise answers with fast, uh, with detailed analysis available on request. Okay. So this is a very simple thing. Yes. So the report will be having all of these things. What artificial, what analytical frameworks would you be looking for? For example, approaches such as best scenario planning. Yes. Go for build versus buy talent strategy. Then make, buy, borrow, bought. McKinsey 7S, CCZ, workforce segmentation, skill gap analysis, SWOT planning. All of this possible, whatever the best. So in this way, it'll take care of all this framework. And meanwhile, I'm also looking at Claude, what actually it has done, right? So we can see that it has done a good job also. Like what are the roles and all these things, scenarios compared. It did a lot of other things, right? So these are the things I can see, but