The future of HR is AI-powered.
But, for many HR leaders, the technical jargon can be overwhelming and scary. What’s a “prompt,” how do you “train” AI, and why are people talking about “hallucinations”?
This guide breaks down these concepts into simple terms so you can confidently navigate the world of AI in HR.
“Traditional AI” typically focuses on analyzing data to make predictions, classifications, or decisions. For example, it can predict employee turnover. “Generative AI” (Gen AI) refers to a category of artificial intelligence that can create new content, such as text or images. It doesn't just analyze or classify data—it generates new data that has similar characteristics to the training data.
Gen AI can generate human-like responses and has given rise to a new wave of Virtual Assistants. In HR, the most common use case for generative AI is to introduce conversational chatbots that can answer questions from employees, like Applaud’s AI Assistant.
Other use cases include creating or updating knowledge articles and automating responses to HR service requests.
Above: Applaud's AI Assistant serves the right information to your employees at the right time via whichever device or channel they’re on.
“Prompts” are the instructions you give an AI to get the answer you want. Think of them like questions or commands. For example, if you ask an AI, “What are the top 5 qualities of a good team leader?”—that’s your prompt.
“Prompt Engineering” is the art of crafting these prompts to get the most accurate and useful responses from the AI. It’s like learning how to ask the right question to get the best answer. This is important because a poorly worded prompt can lead to confusing or irrelevant responses.
A basic prompt like the example below will generate a generic job description with standard responsibilities like “supporting HR initiatives” and “acting as a liaison between employees and management.”
1) Basic prompt example: “Write a job description for an HR Business Partner.”
An enhanced prompt like the second example below will result in a more detailed and relevant job description that emphasizes the specific qualifications, focus areas, and strategic impact expected from a Senior HR Business Partner in a multinational context.
2) Enhanced prompt example: “Write a job description for a Senior HR Business Partner at a multinational company. The role requires 7+ years of experience in HR, expertise in employee relations and talent management, and a proven track record of supporting organizational change. Include a focus on diversity and inclusion initiatives and the ability to influence leadership decisions.”
Prompt Engineering is an emerging skill that will be in demand from many organizational departments, including HR, as AI use in the business becomes more ubiquitous.
In the context of AI, hallucinations happen when the AI makes up information or gives you an answer that sounds believable but is incorrect or nonsensical. This can be problematic in HR, where decisions often rely on accurate information.
Example:
Imagine you ask an AI, “What leave allowance do employees in Singapore receive?” and it responds with a random number like “25 days” which may not be accurate. That’s an AI hallucination. It’s important to always verify the information AI provides, especially when it’s critical to your work.
If you are deploying an AI Assistant, it’s important to test for hallucination and improve the content being used by the AI to answer questions; the most common cause for hallucinations in HR is the content missing or being out of date.
An AI model is like a recipe that tells the AI how to process information and generate responses. It’s based on mathematical algorithms and trained on large datasets to understand patterns and make predictions.
Large Language Models (LLMs) are a type of AI model designed to understand and generate human language. They’re trained on vast amounts of text, like books, articles, and websites, so they can answer questions, generate content, and even have conversations that sound human-like.
Example:
Applaud uses Large Language Models to help HR by automating responses to common employee questions, drafting process and policy document articles, or suggesting content for processes like onboarding or offboarding.
Above: Applaud's AI-powered Knowledge Management uses Large Language Models to identify knowledge gaps and generate new policy documents all in your tone of voice
Training an AI is like teaching a new employee how to do their job. During training, the AI learns from large amounts of data (like documents, emails, and reports) to understand patterns and improve performance over time.
HR Leaders need to be careful about how the organization's data is being used by AI vendors. For example, imagine your AI vendor had access to all the compensation data in your company. A ‘rogue’ AI solution might use that data to train its public model so that people outside your organization could ask things like “How much does the average CHRO earn”. You do not want the AI to be ‘trained’ to answer this question using your data.
Applaud never uses your organization data to train our AI, we only ever use your content and data to answer questions ‘on the spot’.
Now that you know the basics, here are some practical ways Applaud is using AI to help HR leaders.
- HR AI assistants: AI chatbots can answer common employee questions, provide training recommendations, and gather feedback.
- Knowledge Management: AI can create new policy documents, suggest improvements to existing ones, and rate the content for tone of voice and readability.
- Journeys: AI can create suitable tasks and content during critical employee journeys such as new starters joining the organization, new Managers after promotion, or employees enrolling in benefits
AI can be a game-changer for HR, streamlining processes and providing valuable insights. Understanding the basics will help you make informed decisions about how to integrate AI into your HR strategy.