Hi everyone,
I’m exploring how to use OpenAI’s tools to create a conversational flow for a specific use case: allowing prospective students to register for a university’s entrance exam.
The key detail is that this registration is completely free. There is no payment involved. The goal is simply to collect the user’s information (name, email, etc.) through a chat interface and submit it to our university’s registration system. It would be really nice (in the future) for it to be present in the ChatGPT UI. I’ve been reading about the Agentic Commerce Protocol and I’m studying the specs so we apply to this new feature, but as I saw it was designed to work with monetary cases only.
I’ve been looking at two potential approaches:
-
The Agentic Commerce Protocol (ACP): My initial reading of the ACP documentation suggests it’s fundamentally designed for monetary transactions. The specifications for Product Feeds, Agentic Checkout, and Delegated Payment all seem to be built around a product with a price and a payment flow. My question is: Can the ACP be adapted for a zero-value, non-commercial registration flow? Is this a supported use case, or would it be forcing a square peg into a round hole by, for example, setting a price of “0.00” and trying to bypass the payment steps?
-
Custom GPT with a GPT Action: The alternative seems to be creating a Custom GPT (e.g., a “University Admissions Advisor”) and defining a GPT Action. This action would call a custom API endpoint on our end (e.g.,
POST /api/exam/register) to submit the student’s data after collecting it in the conversation. This feels like a more direct and appropriate solution for a data-collection task that doesn’t involve commerce, but it will not be integrated into the defaults ChatGPT interface.
Can this be achieved using the ACP? I see that’s a new fast-growing feature and I’d like to be part of that as well, but I don’t know if my use case will fit to it. It would be really nice.
Thanks in advance for your insights!