Eight Tips for More Effective Generative AI Prompts
Academy of Management Scholar Herman Aguinis of the George Washington University School of Business, one of the most influential management professors and researchers, said creating prompts is the key to using OpenAI's ChatGPT, Anthropic's Claude, Google Gemini, Microsoft Copilot, Perplexity, DALL-E, Adobe Firefly, Stable Diffusion, Midjourney, and other generative-AI software effectively. When users input prompts that lack specificity and crucial contextual information, generative-AI platforms generate too many, too few, or vague recommendations and results that aren't useful.
The following guidelines were designed by Aguinis and coauthors Jose Beltran of Rutgers University and Amando Cope of the George Washington University to give leaders tools to improve their ability to write AI prompts and generate more precise, relevant responses:
Source: Shutterstock
1. Clarify the topic or issue:
Specify the particular area that you're inquiring about (e.g., recruitment, diversity, or performance management). State the question or problem clearly. For instance, "What are best practices in recruitment for tech startups?" is more specific than "Tell me about recruitment."
Source: Shutterstock
2. Provide relevant organizational context:
Include details about the industry, company size, region, or any specific policies that may apply. Mention any unique organizational culture or constraints. "Given that we are a mid-sized software company with a diverse workforce, how should we structure our diversity and inclusion initiatives?" or "How can we tailor our training programs for our Gen Z employees?"
Source: Shutterstock
3. Use precise and standard terminology:
Employ clear and specific terms to guide the model and avoid ambiguity. Use standard language that aligns with industry best practices. For example, "How can we enhance our employee value proposition (EVP) to attract top talent?"
Source: Shutterstock
4. Indicate the desired outcome or action:
Explain what you hope to achieve or the recommendation you seek. Specify if you need a strategic overview or tactical steps. "What strategies can we implement to reduce employee turnover by 10% in the next year?"
Source: Shutterstock
5. Request evidence-based recommendations if needed:
Ask for supporting evidence, research, or citations if you need substantiated information. Ask for case studies or examples to understand the practical applications of certain strategies. You can ask, "Can you provide a case study, research study or paper, expert testimonial, or a citation from an authoritative source, that describes a company that successfully revamped its performance management system?"
Source: Shutterstock
6. Adhere to established best practices:
Adhering to best practices mitigates the need for costly and time-consuming corrective measures, optimizing operational efficiency and resource allocation. Ask ChatGPT to give detailed recommendations based on established best practices in a specific field and emphasize critical strategies and principles for achieving an optimal outcome.
Source: Shutterstock
7. Maintain confidentiality:
Don't sharesensitive or personal information about employees or the organization. Use general or anonymized terms if describing confidential scenarios.
Source: Shutterstock
8. Engage in follow-up communication:
Ask follow-up questions for clarification or more information. Provide feedback if the initial response needs refinement. Take an interactive prompting approach by asking multi-step questions (10 to 20 iterations). Approach ChatGPT as though you are talking with a coworker.
Source: "How to use generative AI as a human resource management assistant," Organizational Dynamics , Vol. 53, Issue 1, January–March 2024, https://doi.org/10.1016/j.orgdyn.2024.101029
The post Eight Tips for More Effective Generative AI Prompts appeared first on Academy of Management Today .
