4 ChatGPT Use Cases: How to Strategically Use Generative Tools

As our clients explore the capabilities of generative tools such as ChatGPT, we asked the lead of our Data Analytics and AI practice, Jeffrey Hawkins , to weigh in on when and how to best employ these tools.

His response: Two key factors to consider when interacting with generative tools are the “objectivity” and “verifiability” of the generated output. Objectivity refers to the extent to which the outputs are facts versus subjective opinions. Verifiability refers to the ease with which the correctness of the outputs can be validated.

By considering these factors in tandem, we can identify the most effective use cases for ChatGPT and other generative tools in the image below.

These use cases are:

Clear-Cut Facts: In the top right, we have examples of situations where the output generated has high objectivity and is easy to verify, such as writing code, step-by-step instructions, or scientific facts. ChatGPT may not always give the correct answer, but we can easily run, test, or check it. Given the high verifiability, these tools can generate significant productivity gains if correctly applied in this quadrant.

Expert Judgement: In the bottom right, we have situations where the output has high objectivity but is difficult to verify, such as document summaries, legal interpretations, and data analysis. To validate correctness and relevancy in this instance, you’d have to be a subject matter expert and familiar with the anticipated outcome. Caution should be exercised when using generative tools in this manner as the outputs may be skewed or contain errors that are harder to spot.

Valid Option Range: In the top left, we have situations where the output generated has low objectivity but is easy to verify, such as a recipes, recommendations, or travel itineraries. The outputs will have multiple correct answers, but you can confirm that they fall within the category requested. The outputs in this quadrant are useful, but often of lower value.

Personal Preference: In the bottom left, we have situations where the output has low objectivity and is difficult to verify as opinion, and tone and style play a role. This includes examples such as a product descriptions, thought pieces, and creative writing. Generative tools are useful in this quadrant as they assist with brainstorming and there is no true “correct” or “incorrect” response.

Overall, ChatGPT and other generative tools can be incredibly valuable resources, but it's important to use them in the right contexts. Each category results in a series of trade-offs that must be carefully considered and understood prior to using the tools.

As consultants, it's our responsibility to guide our clients towards the most effective and efficient use of technology. By carefully considering the objectivity and verifiability of the output, we can make the most of these tools, avoid pitfalls, help our clients achieve their goals, and stay ahead of the curve.