4 posts were split to a new topic: Assistant API Retrieval Falls Short of Identical Custom GPT Retrieval
So, I’m missing a ‘guest’ role (like on Asana, Monday or even Slack) exactly for clients. There are so many use cases here, from customer service to briefing GPTs. And yes, I realise we can go API, but this would mean more budget spent on dev and design, which not all SMPs can afford (which I’m assuming is part of the reason for creating the GPTs for teams in the first place)
@EricGT I’m also missing a ‘guest’ role like on Slack or project management tools (Asana, clickup, Monday) where we can share our GPTs securely with our clients, as we’re looking to create customer service and briefing GPTs we can share with our clients directly, privately and securely.
As a moderator I will do what I can to get this noticed by the OpenAI staff.
I have a use case where I want to share my CustomGPT with clients. The teams workspace seems to be ideal as I can create a separate member account and share it with one of my clients. However, I don’t want my client to do anything else than use my CustomGPT.
- A client is not part of your organization and doesn’t belong in your workspace.
- You can share your GPT with a client without them being part of your workspace.
how?
they’d need at least a chatgpt+ account first. and making one for your customer also isn’t a good idea.
in any case, i wanted to advise folks to take a look at assistants for this use case, but that’s also not a real solution for the no-code custom gpt developers.
I feel like the best solution would be for OpenAI to add a ‘guest’ role for clients (like they have on Slack or on any project management software) who you can then assign specific GPTs for
Yes. The customer should have their own that they make for themselves.
Look…
The entire problem here is no-coders want this to be something it isn’t and are trying to jam a square peg into a round hole.
ChatGPT Team is meant to be used by a group of people, working together within one existing organization, to be able to collaborate on development.
You wouldn’t buy a motorcycle then complain it doesn’t have LATCH attachment points for your toddler’s car seat, right?
It’s the wrong product for that use case.
The right path forward for this type of use is,
- The client gets their own ChatGPT Plus, Team, or Enterprise account and the business shares the GPT with them.
- The developer builds something using the API and shares that with their client.
Otherwise, it really seems like the business is, in effect, reselling a ChatGPT Team account—which is against the terms of service.
¯\_(ツ)_/¯
Not my pig, not my farm.
Everyone is free to do what they will.
That’s very true, but lacking a lathe it’s the best they can do ¯\_(ツ)_/¯
So what is OpenAI’s long-term strategy regarding CustomGPTs and the store? Do they think that everyone is willing to create a ChatGPT+ subscription just to try out some GPTs? I don’t think so.
How can I serve a customer who is looking for a very specific solution that I can provide with a CustomGPT, but who has no interest in creating an account or a CustomGPT? I’m not a developer, so it’s difficult for me to develop an API solution.
I didn’t think so either.
I think they think people will come for GPT-4. Being able to make your own custom GPTs is a nice bonus. Being able to use GPTs created by others is another nice bonus.
At some point someone will come up with the “killer app” for ChatGPT and that could very well end up being the canonical version of ChatGPT for a large number of users. But, given what I’ve seen in the GPT Store so far, we’re not there yet.
There seems to be a little bit of a disconnect here.
The GPT Store (at least to me) seems to be a platform where you can create products for OpenAI’s customers, not yours. Ideally, if you create a product that appeals to enough of OpenAI’s customers it will have the halo effect of helping to bring in and retain new OpenAI customers.
No one is saying you can’t develop something a a one-off for an existing customer of yours, but that’s not really what I see as the impetus behind the GPT Store.
I’m just pointing out that most of the people complaining about this aren’t exactly the types of people for whom this product is intended.
Actually I am the owner of a 32-member Team, I think I really need the finer granularity of permission.
Whether intentionally or unintentionally, the unexpected invitation operation would still incur additional losses for this team.
Hey! Thanks for flagging this, these are part of the tradeoff between Teams and Enterprise. Like other SaaS products, the ability for members to invite other team members is intentional. For Enterprise, there are additional toggles for things like this but the teams plan is intended to be more open.
But the payment is not individual, so the owner should have permission to forbid members from inviting unexpected new members.
Completely agree that only those those with access to billing should be able to invite new members.
150 People is much too large a group to be able to completely trust each member.
Edit: I’ve created a new topic on this subject here if people would like to continue this conversation in a thread that is not marked as resolved.
The problem we are having is that our organization does not have enough members to qualify for Enterprise, but more than enough for the lack of this protection to be a deal-breaker.