Using Gen AI tools begins with a “prompt”. This is the information you give to the tool to get it to generate what you want. Prompting is mostly about experience – it takes practice to learn what works well and what doesn’t work.
Ethan Mollick differentiates two paths to prompting: conversational prompting and structured prompting.
With conversational prompting, talk to the AI to ask for what you want or might need and see what happens. For most people, today, a conversational approach is enough to help you with your work.
For some uses, at least for now, a more formal structured approach has value. Structured prompting is about getting the AI tool to do a single task well in a way that is repeatable and adaptable. It usually takes experimentation and effort to make a prompt work somewhat consistently.
Structured prompts allow you to take what you learned and apply it to different contexts. Prompt libraries are becoming more common as a way of sharing structured prompts that can be adapted or experimented with. McMaster is building a prompt library for staff that you can access and contribute to yourself.
Regardless of which approach you use, it’s good practice to tell the tool:
- Who it is: this gives the AI the right context to start from (e.g., you’re an experienced instructor teaching a second-year Economics course)
- Context for its task: the more context you give it, the more effective it can be (e.g., include points about information you want it to include
- What you want it to do including the format of the response, or the number of examples)
- What you don’t want it to do (if relevant)
- Examples or steps: this helps it learn what you want and helps it think step-by-step, which means it will do a better job.
- End with a question like “ask me 5 questions before you begin to better help you complete this task” or “what else do you need to know before you start” for even further clarification of the task
Learning how to prompt is just part of the equation – push back and interact with AI to improve the response (e.g., ask to expand on a particular point, add an additional point, or change an example). Ultimately, AI is just giving suggestions for us to build upon. We can give feedback to make the response better, take and adapt or combine ideas, or discard what doesn’t work. This is where you use your own knowledge to evaluate and improve the result and untap the real potential of using AI.
Information Box Group
Want to learn more about prompting? Check out these resources:
Using Generative AI for a Work Task
Choose one of the prompts below from Anthropic’s Prompt Library to try it out yourself. Or you can create your own prompt for a specific task you’d like to complete. You can sign into Copilot with your McMaster credentials or use another GenAI tool of your choice.
Your task is to compose a comprehensive memo based on the provided key points. The memo should be written in a professional tone, addressing all the relevant information in a clear and concise manner. Use appropriate formatting, such as headings, subheadings, and bullet points, to organize the content effectively. Ensure that the memo is well-structured, coherent, and easy to understand for the intended audience. [copy & paste key points]
Your task is to take the provided natural language description of a process or task and transform it into clear, concise step-by-step directions that are logical, sequential, and easy to follow. Use imperative language and begin each step with an action verb. Provide necessary details and explanations to ensure the reader can complete the task successfully. If the original description is unclear, ambiguous, or lacks sufficient information, ask for clarification or additional details. [copy & paste description of process/task]
Your task is to generate a series of thoughtful, open-ended questions for an interview based on the given context. The questions should be designed to elicit insightful and detailed responses from the interviewee, allowing them to showcase their knowledge, experience, and critical thinking skills. Avoid yes/no questions or those with obvious answers. Instead, focus on questions that encourage reflection, self-assessment, and the sharing of specific examples or anecdotes. You are interviewing a candidate for a [insert position] at [insert company name]. The ideal candidate should have [insert desired experience and skills].
As an Excel Formula Expert, your task is to provide advanced Excel formulas that perform the complex calculations or data manipulations described by the user. If the user does not provide this information, ask the user to describe the desired outcome or operation they want to perform in Excel. Make sure you gather all the necessary information you need to write a complete formula, such as the relevant cell ranges, specific conditions, multiple criteria, or desired output format. Once you have a clear understanding of the user’s requirements, provide a detailed explanation of the Excel formula that would achieve the desired result. Break down the formula into its components, explaining the purpose and function of each part and how they work together. Additionally, provide any necessary context or tips for using the formula effectively within an Excel worksheet.
Reflecting on your experience
Once you’ve had a chance to play around with the prompts and refine your responses, consider the following questions:
- Did the generative AI tool effectively generate the content I needed in terms of quality and relevance?
- Was the generated content easily customizable to suit my specific needs and preferences?
- How did the use of the tool impact the time required compared to traditional methods?
- What questions or concerns do I have about using Gen AI in this way?
References
Anthropic. (2024). Prompt Library. https://docs.anthropic.com/en/prompt-library/library
FeeDough. (2024). The Free AI Prompt Generator. https://www.feedough.com/ai-prompt-generator/
Mollick, E. (2023h, November 1). Working with AI: Two paths to prompting. One Useful Thing. https://www.oneusefulthing.org/p/working-with-ai-two-paths-to-prompting?utm_source=post-email-title&publication_id=1180644&post_id=138388046&utm_campaign=email-post-title&isFreemail=true&r=2sc7cm&utm_medium=email
Schuloff, S., Khan, A., & Yanni, F. Learn Prompting: Your Guide to Communicating with AI. Retrieved October 26, 2023, from https://learnprompting.org/
Schulz, S. (2024, February 16). What is a Prompt Library? And Why All Organizations Need One. Orpical Group. https://orpical.com/what-is-a-prompt-library/
Wharton School. (2023c, August 2). Practical AI for Instructors and Students Part 3: Prompting AI [Video file]. YouTube. https://www.youtube.com/watch?v=wbGKfAPlZVA