OpenAI GPT Store allows developers to share custom GPT apps
The AI vendor introduced the highly anticipated store that enables developers to share their applications. It also introduced a new collaborative workspace for employees.
OpenAI wants to make ChatGPT more accessible to employees and give users the opportunity to monetize their GPTs.
The AI vendor on Jan. 10 introduced its anticipated GPT Store and ChatGPT Team platform.
GPT Store is for developers who create custom versions of ChatGPT and want to share them with others, according to OpenAI.
The store features different GPTs from OpenAI partners and users. For example, custom apps on the store include a Khan Academy programming tutor and a Canva design tool. GPT builders can build and save their GPT in the store.
GPT builders in the U.S. will soon be paid based on user engagement with their GPTs through OpenAI's GPT builder revenue program, which is expected to launch soon, according to the AI vendor.
Meanwhile, ChatGPT Team offers users access to OpenAI's models such as GPT-4 and Dall-E 3.
ChatGPT Team includes a secure workspace for employees and the ability for users to create and share custom GPTs, according to OpenAI. ChatGPT Team costs $25 per user, per month, when billed annually, or $30 per user, per month, when billed monthly.
GPT Store
GPT Store is similar to the Apple and Android app store marketplaces, Gartner analyst Arun Chandrasekaran said.
Many of the initial apps available in the GPT Store will likely be consumer-facing applications, he said. For example, the store could have a GPT app that is a movie recommendation engine or a GPT app to help young students learn the basics of programming.
The whole idea behind that is the fact that ChatGPT is a general-purpose chatbot, and with GPT Store apps, you will get these specialized versions of ChatGPT.
Arun ChandrasekaranAnalyst, Gartner
"The whole idea behind that is the fact that ChatGPT is a general-purpose chatbot, and with GPT Store apps, you will get these specialized versions of ChatGPT," Chandrasekaran said.
But enterprises might not jump to adopt or use the GPT Store, though they might eventually migrate there, he said.
"In the enterprise space, we might see both application vendors building capabilities as well as some end-user organizations trying to become a quasi-vendor," Chandrasekaran added.
Application vendors might decide to create an application in which they build on the capabilities of ChatGPT with some domain-specific data and make that available in the GPT Store. The end-user organization might be a bank that creates a custom implementation of GPT and offers that to others in the financial services by making it available in the GPT Store.
"The simple way to think about GPT Store is it could be an effective distribution mechanism for these applications," Chandrasekaran said.
What remains unclear about the store is the issue of monetization.
OpenAI has not yet said how it will dictate the pricing model for the apps in the store and whether it will take Apple's and Google's approach of taking a percentage of the cost of each application, Chandrasekaran said.
Moreover, it will also be interesting to see how the GPT Store user interface will be integrated into the existing user interface of ChatGPT, he added.
"Just in terms of making sure that it's a seamless experience for the user, I think that's an important goal for OpenAI," he said. "Given thousands or possibly millions of apps that I expect to see in the GPT Store, I don't know how they could keep that layout very simple and the user experience pretty straightforward."
ChatGPT Team
While the GPT Store might appeal to developers looking for a way to create GPTs that promote or enhance a brand, ChatGPT Team could also interest small and medium-sized businesses, Futurum Group analyst Mark Beccue said.
"I don't see [ChatGPT] Team necessarily appealing to developers per se, but more likely SMBs that would like to collaborate on ChatGPT products with a dedicated channel or resources rather than typical company collab tools," Beccue said.
ChatGPT Team also looks like an indirect way OpenAI is competing with and offering similar capabilities as its partner Microsoft's Copilot system, Forrester Research analyst William McKeon-White said.
"It seems that there's a bunch of different ways of interacting that are combined and a bunch of different OpenAI models that are put together into something that would act as a copilot for organizations that aren't potentially all in Microsoft," McKeon-White said.
Moreover, ChatGPT Team seems to offer a degree of flexibility and customization like Microsoft AI Builder, he added.
However, ChatGPT Team lacks the Microsoft Graph underlying data layer, McKeon-White said. Microsoft Graph is a Microsoft API developer platform that connects various services and devices in Microsoft 365 and is linked to Copilot.
The lack of that data management layer is a drawback, but "it is a very interesting proposition to be able to combine all of these different powerful capabilities under one product," McKeon-White said.
Since ChatGPT Team is designed to be a digital assistant to employees, OpenAI will have to figure out how to manage some of the problems that come with integrating the collaborative systems with other tools that employees work with, he added. If an enterprise works with an older system that doesn't have a modern API connection, that could be problematic.
Another problem, especially relating to the lack of data management, is if an employee wants to perform a task or access a data resource, but is unable to because ChatGPT Team doesn't have that ability.
"There's all these sorts of weird caveats to this," McKeon-White said. "Now that they are explicitly building work accelerators for individuals and employees, they're going to need to find a way around those expectations."
ChatGPT Team users can access a private section of the GPT Store, which includes GPTs securely published to their workspace.
ChatGPT Enterprise customers will soon have access to the store and will have stronger admin controls, OpenAI said.
Esther Ajao is a TechTarget Editorial news writer covering artificial intelligence software and systems.
