On 10th January, OpenAI unveiled the ‘GPT Store’, a virtual storefront available to paid subscribers of ChatGPT that allows users to create, and obtain, custom iterations of ChatGPT.
These iterations have been termed ‘GPTs’, and were announced by OpenAI last November. The GPT Store was slated to follow shortly after, but its rollout was delayed by the events that saw CEO Sam Altman removed, and then later reinstated, as the head of the company.
Now, the GPT Store has finally arrived (at least, in the United States), and many are hailing it as a watershed moment equivalent to Apple’s App Store launch in 2008.
Here’s why I don’t think that’s quite warranted – but why it’s still significant for marketers and businesses who want to exploit generative AI’s full potential.
What are GPTs really?
The question of whether the GPT Store can be compared to Apple’s App Store hinges on whether it’s fair to compare GPTs to iOS apps.
The App Store famously opened up Apple’s ecosystem to developers (after an initial foray into web apps that was received poorly by developers) to create third-party programs and experiences for the iPhone, ushering in what Apple rather bombastically called a “worldwide … phenomenon” of “creativity, innovation and entrepreneurship” in its 10th anniversary retrospective. Finally, it was possible for anyone to make – and crucially, sell – an app for the iPhone.
The difference between this and ChatGPT’s situation is that OpenAI introduced a way to create native apps for ChatGPT some time ago, when it opened up ChatGPT’s API (for a fee, naturally) and introduced ChatGPT Plugins.
These are available through the ‘Plugin Store’, which allows ChatGPT Plus users to enable plugins from businesses like Expedia, Instacart, Klarna, and OpenTable, augmenting ChatGPT’s capabilities so that it can be directed to perform actions like ordering food (Instacart) or finding a restaurant (OpenTable).
GPTs, on the other hand, are less sophisticated than Plugins – by design, as they’re intended to be no-code creations that non-developers can put together using OpenAI’s GPT Builder (itself a chatbot within ChatGPT).
Daniel Li, cofounder and CEO of Plus, has a good explanation of the difference between GPTs and Plugins (and Custom Instructions, an even ‘lighter’ customisation option that was rolled out before GPTs), but the short version is that GPTs are a way of pre-loading ChatGPT with certain prompts and parameters to get a desired output. As OpenAI explained in its blog announcement:
“Many power users maintain a list of carefully crafted prompts and instruction sets, manually copying them into ChatGPT. GPTs now do all of that for you.”
Plugins are more likely to be of interest to brands wanting to create a version of their product capabilities that can be launched from within ChatGPT, while GPTs are likely to be more attractive to individuals (although they have a strong use case for businesses that I’ll explore shortly). Particularly because OpenAI has promised that revenue sharing for GPT creators will arrive at some point during Q1 (the details of which are as of yet unknown).
Why the GPT Store is still significant
While I don’t see the GPT Store as being on a par with the advent of Apple’s App Store, that doesn’t mean it isn’t still significant.
GPTs allowed power users of ChatGPT – and anyone else who was so inclined – to load in their own custom setup for ChatGPT rather than beginning from scratch every single time, and the GPT Store now enables them to share it.
This means that anyone can benefit from that customisation regardless of whether they know how or why it works; they don’t need that. They can simply look for a GPT that accomplishes what they need, and use it (although of course, there may be a variety to choose from, and in that case recommendations and good standing would probably determine which GPT they choose – in that respect, the GPT Store is like the App Store).
The addition of the GPT Store to ChatGPT is a lot more like what Benedict Evans mused about in his Unbundling AI essay of October 2023 when he compared LLM prompts to computer command lines and wondered what might serve as the equivalent of the graphic user interface (GUI) for LLMs.
Evans concluded: “…ChatGPT sometimes seems … like the original PCs … It’s a general purpose technology, there’s a command line, and some stuff that’s theoretically magic, and a few things that are extremely useful to a few people, but we don’t yet have the richness of all the software that came on top – all of the embodied use cases.
“…for a lot of … people it looks a bit like those PCs ads of the late 1970s that promised you could use it to organise recipes or balance your cheque book – it can do anything, but what?”
What does it mean for marketing?
As ChatGPT rapidly took off and users explored its capabilities, many began to speculate that specialists in ‘prompt engineering’ – who know the most effective phrasing to use to get a desired response from the AI – would be in high demand.
Then it started to become apparent that facility with LLMs like ChatGPT might need to become part of a worker’s standard skillset. As eSure Group’s Elisabeth Ling said at Econsultancy Live in November, “[Y]ou need to do your homework … Use [AI] in your daily life, and then when you know what type of algorithm does what – immediately you will make the connection. ‘I have this type of problem, I can use a recommender… [or] I can use generative AI’.”
Now, the GPT Store and its offered GPTs may elide some of the need to spend time working with the tool to gain a desired output (although familiarity with ChatGPT and its capabilities will still be important). Econsultancy’s Future of Marketing Report 2023 found that many marketers are already employing generative AI for marketing use cases such as written content creation and copywriting (58%), SEO keyword research (43%) and summarising emails, meetings and actions (38%).
GPTs could facilitate these use cases by allowing marketers to get started quickly with a purpose-built iteration – or by allowing companies to easily assemble and distribute their custom ChatGPT setups to employees. OpenAI is allowing ChatGPT Enterprise customers to design and share GPTs internally within a business – and its new Team plan also seems tailored for this use case, allowing smaller groups of users to access a private section of the Store and securely publish GPTs.
The GPT Store has its share of teething issues, such as complaints from users about copycats and the need to regulate and control for quality, but OpenAI will surely have a plan for how to tackle these. The revenue sharing terms are also still unknown and their favourability may determine how attractive GPTs are to would-be creators (who need to ‘pay to play’ by investing in a ChatGPT Plus subscription).
There are likely to be many less-than-worthwhile offerings, but at least for power-user businesses and teams, the ability to create and share bespoke GPT setups internally (thus making sure all uses of ChatGPT are on-brand) will be valuable. And for more sophisticated offerings, brands can still create a ChatGPT Plugin.
Now, brace for a wave of ‘How to optimise for the GPT Store’ guides…
Econsultancy offers a short course in AI for marketing.
Comments