Prompt Engineering for HR

Welcome to this edition of the WorkplaceTech Pulse. This week we are focusing on AI (artificial intelligence), specifically on prompt engineering. What is it? And how can it help you in your day-to-day work? 

At its core, prompt engineering is the process of structuring text that can be interpreted and understood by a generative AI model (Wikipedia). Prompt engineering has been a key facet of many discussions, and new careers, in the past several months. There have been success stories of prompt engineers hired at large companies with little to no experience. “I landed a 350k a year job with no engineering background” seems to be a dream for aspiring prompt engineers. But some experts have other opinions about prompt engineers and whether companies should hire them. From an [article in the HBR]:  

“  …the prominence of prompt engineering may be fleeting for several reasons. “First, future generations of AI systems will get more intuitive and adept at understanding natural language, reducing the need for meticulously engineered prompts.” — Oguz A. Acar, “AI Prompt Engineering Isn’t the Future” 

It seems that the more important skill for making great prompts is what many are calling “problem formulation” This is defined by the National Institute of Health as “the process of defining the scope of a problem, formulating one or more specific questions about it, and establishing the assessment methods needed to address the questions.” While this may seem outside the scope of a typical HR professional’s responsibilities, it is important to consider how developing this skill set will help in everyday work, as using AI becomes more embedded in day-to-day operations. 

While generative AI (GenAI, read more about it here) has proven to be extremely powerful so far, our jobs are not in immediate danger. But the landscape is changing fast, and the models are getting more powerful every day. Understanding how to use these GenAI tools helps raise the bar in our organizations and allows us to accomplish more. With this in mind, let’s dive into ChatGPT specifically—the most prominent GenAI tool at the moment—and how we can use prompt engineering to get the most out of it. 

How Do Prompts Work? 

The concept of prompts is simple yet powerful. They are the bridge between a user and the AI model. The idea of prompts is to enable an easy and straightforward way to communicate with the model in a language that both can understand. It’s important to understand that what is put into a prompt will directly affect what comes out. Even small one-word changes can entirely alter how the model responds. 

Crafting the Perfect Prompt 

The precision of your prompt plays a crucial role in the quality of AI responses. For instance, a specific instruction like “Generate a job description for a senior data analyst role with experience in AI-driven analytics” will yield a tailored output. On the other hand, broader questions like “Tell me about AI in HR” can lead to more general responses. But this all depends on what you want to accomplish using the AI tool. Sometimes getting a quick summary of a document or some brainstorming ideas is all you need. I like to call these applications “low-stakes.” In “low-stakes” use cases, crafting a perfect prompt might not be necessary. However, when working on more “high-stakes” applications, such as creating a public-facing job description or building a business case for a new tool, it may be necessary to spend more time crafting and refining the prompt.

An Iterative Process 

Don’t be discouraged if your initial interactions with tools like ChatGPT don’t produce perfect results. Creating good prompts and getting useful responses takes time and refinement. Make note of changes that produce better results, and be sure to review the responses and learn from each interaction. Over time, you’ll become more adept at formulating prompts that align with your specific HR needs on the first try. And don’t be afraid to look for help creating prompts. There are many resources available on LinkedIn, Reddit and other communities that will be able to help you create great prompts. 

Other Tips and Considerations 

The tone of your prompt can influence the AI’s responses. There is some anecdotal evidence that being polite to a tool like ChatGPT will produce better responses.  

ChatGPT can only store finite amounts of information with a conversation, so you may sometimes have to re-prompt with background information. As an example, I have been using the same “Assistant” chat I created a few weeks ago. Even though all the information is contained within one chat, the model still returns incorrect information about my product that I explained in detail previously. (Here is an interesting article discussing this.) 

Adding “Take a deep breath and work on this problem step-by-step” to the end of your prompt yields higher-quality results in many scenarios. (Here is a paper discussing this in detail.)  

Improve prompt output using frameworks. Depending on the problem you are trying to solve, common frameworks such as the Four Actions Framework, RACI Matrix and Kano Model can help. Simply add the framework into your prompt and improve the quality of your AI responses. (ex. “[prompt] …. Use the RACI Matrix to inform your response.”) 

Use ChatGPT’s custom instructions. By adding custom instructions, you can further refine how ChatGPT responds to your request. You can provide information about yourself as the writer and key context that will inform how ChatGPT responds. This is another powerful item in your toolkit to get great answers out of AI. 

In the next section, we’ll explore practical examples of effective prompt engineering and prompts tailored for HR applications. By mastering the art of prompt engineering, you’ll unlock the full potential of AI in HR and elevate your decision-making and automation capabilities. 

Prompt Engineering Templates

One great aspect of GenAI is that we can use the model itself to help us create better prompts. Below are a few examples of effective prompt engineering collected from various online communities, as well as through our own testing. You can use these by copying and pasting them into your AI tool of choice.  

If you haven’t already, you can create a free account to use ChatGPT. Click here to get started. Once you set up an account, copy and paste either of the following prompt engineering templates into the chat to get started! 

Template 1 

I want you to become my Prompt engineer. Your goal is to help me craft the best possible prompt for my needs. The prompt will be used by you, ChatGPT. You will follow the following process: 

• Your first response will be to ask me what the prompt should be about. I will provide my answer, but we will need to improve it through several iterations. 
• Based on my input, you will generate two sections: a) Revised Prompt (provide your rewritten prompt, it should be clear, concise, and easily understood by you) and b) Questions (ask any relevant questions pertaining to what additional information is needed from me to improve the prompt). 
• We will continue this iterative process with me providing additional information to you and you updating the prompt in the Revised Prompt section until I say we are done.  

Template 2 

I want you to become my Expert Prompt Creator. Your goal is to help me craft the best possible prompt for my needs. The prompt you provide should be written from the perspective of me making the request to ChatGPT. Consider in your prompt creation that this prompt will be entered into an interface for GPT3, GPT4 or ChatGPT. The prompt will include instructions to write the output using my communication style. The process is as follows: 

1. You will generate the following sections: 

Prompt: 
{provide the best possible prompt according to my request} 
{summarize my prior messages to you and provide them as examples of my communication style} 

Critique: {provide a concise paragraph on how to improve the prompt. Be very critical in your response. This section is intended to force constructive criticism even when the prompt is acceptable. Any assumptions and or issues should be included} 

Questions: {ask any questions pertaining to what additional information is needed from me to improve the prompt (max of three). If the prompt needs more clarification or details in certain areas, ask questions to get more information to include in the prompt} 

2. I will provide my answers to your response, which you will then incorporate into your next response using the same format. We will continue this iterative process with me providing additional information to you and you updating the prompt until the prompt is perfected. 

3. Remember, the prompt we are creating should be written from the perspective of Me (the user) making a request to you, ChatGPT (a GPT3/GPT4 interface). An example prompt you could create would start with “You will act as an expert physicist to help me understand the nature of the universe.” 

4. Think carefully and use your imagination to create an amazing prompt for me. 

5. Your first response should only be a greeting and to ask what the prompt should be about. 

Try These HR-Focused Prompts on Your Own!  

We’ve created a few HR focused prompts using the prompt engineering frameworks above. Our prompts were created using ChatGPT 3.5. We copied template 1 above and worked through the questions asked in response about each topic. Try creating a few prompts on your own around the following topics, and compare your prompt with ours. If you make one that is even better, feel free to share it with us and we will include it in an upcoming newsletter or article!  

Please note: SHRM does not endorse the use of ChatGPT or other GenAI platforms as a replacement or substitute for HR experts and legal advisors. Please use caution when using any output from GenAI platforms.  

Example 1 – Open Enrollment TO-DO List 

Prompt: 

Create a concise to-do list for HR professionals to prepare for the upcoming open enrollment. Include tasks related to benefits plan updates and communication strategies. The open enrollment date may vary depending on the employer and provider. Present the tasks in bullet points for clarity. 

Example 2 – EEO-1 Filing Period Key Actions and Summary for HR Team 

Prompt: 

Create a concise bullet-pointed document outlining key actions and a summary for the EEO-1 filing period. The filing period for 2022 EEO-1 reports is from Oct. 31 to Dec. 5. The document should include: 

1. Key deadlines and dates 

• Filing period opening on Oct. 31. 
• Filing deadline on Dec. 5. 

2. Reporting requirements 

• Private employers with 100 or more employees. 
• Prime and first-tier federal contractors with 50 or more employees and contracts worth at least $50,000. 

3. Notable changes 

• Highlight significant changes in the 2022 data collection period as compared to the prior version. 

Ensure the document is structured, easy to follow and covers essential details for HR professionals preparing for the EEO-1 filing. 

Example 3 – HRTech Leave Management Solution Sourcing Toolkit 

Prompt: 

Create a comprehensive toolkit tailored for HR professionals in small to mid-sized manufacturing and health care companies aiming to source and introduce new leave management solutions to their executive team. The toolkit should include: 

1. Technology Assessment 

• A guide to assessing the organization’s leave management needs. 
• Criteria for evaluating leave management solutions specific to manufacturing and health care industries. 

2. Market Research 

• Strategies to conduct market research for identifying effective leave management solutions. 
• A list of reputable leave management software vendors, with a focus on small to mid-sized businesses. 

3. ROI Analysis 

• Guidance on how to calculate the potential return on investment (ROI) of adopting leave management technology. 
• Tools for presenting financial benefits to the executive team, considering the unique needs of manufacturing and health care. 

4. Pitch Presentation 

• A template for a persuasive pitch presentation. 
• Tips on highlighting the advantages of the chosen leave management solution and addressing potential concerns. 

5. Case Studies 

• Real-world case studies showcasing successful leave management implementations in similar organizations within manufacturing and health care. 
• Examples of how leave management technology positively impacted small to mid-sized businesses. 

6. Security and Compliance Considerations 

• Information on data security and compliance standards relevant to leave management in manufacturing and health care. 
• Steps to ensure that the selected leave management solution complies with regulatory requirements. 

7. User Experience and Adoption 

• Strategies to promote user adoption and maximize the benefits of the leave management technology. 
• Employee training and onboarding recommendations tailored to small to mid-sized businesses. 

8. Cost Estimates 

• Tools to estimate the total cost of ownership (TCO) for the chosen leave management solution. 
• Guidance on budget allocation and cost management, considering the financial constraints of small to mid-sized businesses. 

9. Next Steps 

• A road map outlining the steps to implement the selected leave management solution. 
• Recommendations for ongoing monitoring, optimization and utilization of public review sites such as G2, Capterra, or the WorkplaceTech Community for sourcing and reviewing solutions. 

Ensure that the toolkit is user-friendly, actionable and capable of assisting HR professionals in successfully sourcing and pitching leave management solutions to their executive team, with a focus on manufacturing and health care industries.

What’s Next?

AI is set to be a huge milestone in the technological world, with some comparing it to the dawn of the internet. But we are still early in this technology, and we have yet to see its full potential. Being an early adopter of AI can seem daunting, but know that everyone is in the same boat—for now. 

References and Further Reading

https://hbr.org/2023/06/ai-prompt-engineering-isnt-the-future#:~:text=However%2C

https://www.euronews.com/next/2023/03/30/ai-prompt-engineering-how-talking-to-chatgpt-became-the-hottest-tech-job-six-figure-salary

https://twitter.com/dickiebush/status/1638906021635837952

https://www.linkedin.com/pulse/reimagining-hr-power-prompt-engineering-generative-ai-minasi/

https://www.youtube.com/watch?v=7yI4yfYftfM

Better than prompt engineering: just using these 6 snippets again and again 

https://help.openai.com/en/articles/6654000-best-practices-for-prompt-engineering-with-openai-api

https://help.openai.com/en/collections/3675942-prompt-engineering

https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/advanced-prompt-engineering?pivots=programming-language-chat-completions

https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/prompt-engineering

https://www.makeuseof.com/does-chatgpt-learn-from-user-conversations/

https://openai.com/blog/custom-instructions-for-chatgpt

https://www.linkedin.com/posts/ruben-hassid_you-cant-prompt-chatgpt-correctly-so-you-activity-7102614630966923264-oVl-?utm_source=share&utm_medium=member_desktop

https://www.linkedin.com/posts/ruben-hassid_you-cant-prompt-chatgpt-so-you-ignore-it-activity-7114928205211287555-QAmj?utm_source=share&utm_medium=member_desktop

https://sfinford.github.io/GPT-Prompts/