Before you request your new GPT, clearly define why the GPT exists, who it is for, and what problem it is meant to solve. This step prevents building a tool without strategic intent or practical value. Also think clearly about what final result you expect to receive when using the GPT. Specify the exact outcome the GPT should produce, including what “success” looks like for the user. This anchors the GPT around results, not features or prompts.

Use the Custom GPT Request Form to submit your GPT. This will be sent to the AI Department and your manager for approval. The entire process will take 3-5 business days, unless you have an urgent request. If that is the case, talk to your manager and they will communicate that to the AI Department for a rush order.

After your manager reviews and approves your request, the AI Department will draft your GPT by writing clear, structured instructions that guide how the GPT should think, respond, and behave. These instructions translate the purpose and desired result into repeatable system logic. You will be provided access to the GPT and the project notes with a link on the PAE.ai website here.

You and your manager will be provided with the new GPT for testing. Test the GPT internally using realistic scenarios to identify gaps, inconsistencies, or confusion. Refine instructions and logic until outputs are reliable and aligned with the intended outcome. Submit your changes to the AI Department as comments on the GPT Production page for your new GPT.

After the GPT has been tested and refined, the AI Department will release the GPT to its intended users on the PAE GPT Database found here. Ensure positioning, description, and access are clear so users understand how and when to use it.

Collect real-world feedback from users to evaluate performance, usefulness, and clarity. Use this input to continuously refine the GPT so it improves over time and stays aligned with evolving needs.