GPT Playground and GPT-4o

I use Playground to engineer my prompts. I use different prompt engineering techniques, such as prompt chaining or tree-of-thought (ToT), and let them run on the different models. This way I can refine which combination of a prompt technique and model give the best result for a specific use case.

For instance, you want to create a Project Charter with certain characteristics known to you and use one of PMI's templates. Using different techniques and models give you different outputs. You pick the combination that meets your requirement the best, and refine your prompt further until you have the result you needed.

Now you have a Project Charter tailored towards your requirements. But it doesn't stop there!

Because now you know what you did gave you a great output, you can easily apply the same combination for similar document types, like a PMI-templated communication plan or a resource management plan.

You see, this can save you a lot of time over 'manually prompting' in ChatGPT and then also only be limited to a certain GPT model. As Playground is limited to OpenAI models, you can also use IBM watsonx, Dust, or PromptPerfect which will allow you to test various different models.

Happy prompting!