Artificial Intelligence (AI) is no doubt here to stay and seems to be growing almost exponentially. Business professionals undoubtedly continue to find new uses for AI, yet so many opportunities remain. Unfortunately, like many other technologies, most business professionals have not received formal training on issuing better AI prompts. Thus, although they may receive good results, it’s likely that they are having to modify their prompts multiple times and they may not be getting the best results possible. In this article, you will learn some of the best practices to issue better AI prompts to yield better AI results.
Four Key Components Of AI Prompts
According to Microsoft, four of the key components of your prompts include:
- Tasks
- Context
- Expectations
- Output
A task is the most fundamental of these four components. You use tasks to explain to your AI engine what you want it to do for you. Often, you will identify your task in some context; the context describes the data that the prompt will act on. It can also identify any input variables. Two other fundamental – and optional — elements are expectations and output. Expectations help to identify the goals associated with the prompt. Likewise, output is used to describe how you want the results of your prompt to appear. For example, you might choose to have the results appear as text, graphs, or data tables, among other choices. Note that of the four items listed above, the only required item is a task.
An Example
Consider the following example of a prompt.
Create a summary of employee travel expenses. The summary should identify key trends and items and any noted errors, irregularities, or unusual transactions. Summarize these expenses by amount, month, and employee. Create a report that is suitable for a senior management meeting.
In the example above, the first sentence represents the task, the second represents the context, the third provides expectations, and the fourth defines how the output should appear.
Another Approach
Continuing with the above example, issuing the prompt in full is unnecessary. Instead, many find it easier – particularly if they are new to AI – to create a series of prompts instead of a singular “uber prompt.” For example, you could issue four separate prompts shown below, in succession to obtain the same results.
- Create a summary of employee travel expenses.
- Identify key trends and items and any noted errors, irregularities, or unusual transactions.
- Summarize these expenses by amount, month, and employee.
- Create a report that is suitable for senior management.
Using this approach, you can inspect the results incrementally to help ensure you’re on the right track to get a correct response. This methodology can be particularly useful to those just beginning their AI journey.
Additional Guidance
Of course, the more explicit you are with your prompts, the more likely you will receive responses relevant to your needs. For instance, if you use a prompt that refers to all team members in the organization, your results will reflect all team members. But, if you provide context (i.e., a filter) that you are interested only in team members in the Sales department, then your results will reflect the implicit filter provided in the prompt.
Another critical issue to consider relative to issuing prompts involves understanding the near-universal imitations of generative AI platforms. While these tools are great and getting better every day, they are incapable of doing everything we might want them to do. Thus, recognize where AI excels and use it where you can maximize your chances of success.
For example, generative AI performs well in the following six tasks.
- Classifying text,
- Performing sentiment analyses,
- Rewriting content,
- Summarizing information,
- Extracting data,
- Creating a response to other text, such as an email.
Using AI in situations where it is best suited increases your chances of success substantially. Therefore, focus on creating prompts to perform the six tasks listed above.
Summary
To close, generative AI tools such as ChatGPT, Gemini, and Copilot are here to stay and can help us improve productivity almost immediately upon adoption. The key to this success, though, rests with issuing effective prompts so that the tools complete the tasks at hand accurately and efficiently. Follow the guidance in this article, and you’ll be well on your way.
At K2 Enterprises, our commitment lies in providing unwavering support and expert instruction to CPAs. Explore the wealth of resources on our website, where you’ll find valuable insights on selecting the most suitable accounting software, ensuring your firm is equipped with the right tools for the journey ahead. If you work in accounting or finance, K2 Enterprises provides continuing education programs to enhance your skills and credentials. Need help learning how to solve your business’s accounting technology needs and selecting the right software for accounting or CPA Firms? Visit us at k2e.com, where we make sophisticated technology understandable to anyone through our conferences, seminars, or on-demand courses.
