This briefing is part of our Generative AI: Year in Review series, which distills some of the overriding trends in generative AI from the past year and considers what they have to offer brands and marketers looking to take advantage of this influential technology.

It’s a little hard to believe that ChatGPT had only just celebrated its first anniversary by the end of 2023 – turning one right before the year was out, on 30th November.

The year that followed was full of significant moments for generative AI, from the Microsoft versus Google Search Generative Experience (SGE) battles to the release of ChatGPT Plus and GPT-4 to an explosion of generative AI-powered and enabled tools.

But as with any much-hyped technology, generative AI is often being applied first and justified later, without all of its applications being backed up by a genuine use case (see our 2024 digital and marketing trends for a discussion on ‘finding business value‘). And while the technology has real transformative potential, its drawbacks are not to be ignored, particularly as it becomes ever more ubiquitous.

In this briefing, we’ll examine OpenAI’s recent moves to turn ChatGPT from a ‘simple’ chat tool into a platform that can support other tools and experiences via Plugins and GPTs: what the developments mean, and whether they are worth the investment for brands and marketers.

A new ‘app store’ for ChatGPT

March 2023 was a busy month for OpenAI. Following on from the release of ChatGPT’s API and GPT-4, the company announced ChatGPT plugins: a way to build tools and products to be operated from within ChatGPT itself.

This signalled a sea change in the nature of ChatGPT, turning it from a chat application (albeit one that could already accomplish a wide variety of tasks) into a platform that could support other tools and experiences via a chat-based interface.

Expedia, Instacart, OpenTable and Klarna Shopping were among the first brands to launch ChatGPT plugins. The Expedia plugin, for instance, can answer queries about an upcoming trip and surface flight details, accommodation suggestions, activities in the area, and vehicle rentals from Expedia, complete with links to finalise the booking.

This means that the full end-to-end user journey cannot be completed within ChatGPT, but ChatGPT can do significant work in bringing the user further along the sales funnel if the customer experience is right.

Speaking to PhocusWire about the potential of generative AI in travel, Rajesh Naidu, Senior Vice President and Chief Architect at Expedia Group, said, “[A] real opportunity for ChatGPT in travel is trip planning. … It can simplify an extremely manual planning process involving sifting through lots of options down to minutes.”

In effect, online travel agents (OTAs) powered by generative AI are moving back into the role that human travel agents occupied for decades. This removes some level of consumer choice, but may be preferable as long as the recommendations are considered high-quality and trustworthy – something that will fall to the brand to accomplish by ensuring that any generative AI experience lives up to its own standards.

ChatGPT’s Plugin Store and the more recently-released GPT Store have both been compared to the launch of Apple’s App Store in 2008: the thing that will open the floodgates to a new ecosystem of software built atop the ChatGPT platform, enabling products and solutions that we can’t yet imagine.

How worthwhile are Plugins and GPTs for brands?

But there’s a catch that hampers the potential appeal of Plugins and GPTs for brands: they are solely available to ChatGPT Plus subscribers, offering an attractive perk for any users considering paying for a Plus subscription, but less of an incentive for businesses, particularly when they already have the ability to embed ChatGPT with their own products via an API – as the likes of Instacart and Expedia have already done.

It’s difficult to find exact figures on the number of Plus subscribers, but a graph created by Andreeson Horowitz using data from Yipit put the number at more than two million users as of July 2023. A reported spike in subscribers following OpenAI’s first Dev Day may have added to that number, but it’s a far cry from the 100 million users reported to be using the free version of ChatGPT every week.

In marketing terms, a ChatGPT plugin has the potential to act like an interactive search ad: a well-placed signpost for a brand that can net a portion of ChatGPT’s ‘traffic’ and hopefully steer it towards the brand’s own space.

However, does the relatively limited audience make it worth a brand’s while to build a ChatGPT plugin, especially when this means giving up total control of the user experience to OpenAI? Opening up plugins to all ChatGPT users would be a game-changer for brand visibility, but would undermine much of the value-add that a Plus subscription has to offer users.

This goes some way towards explaining why, even though Plugins should have been an huge moment for ChatGPT, the buzz around them didn’t last very long. It has picked up again following the release of GPTs – custom iterations of ChatGPT that enable the user to pre-load certain prompts and parameters to avoid starting with a blank slate – and the GPT Store, a storefront for Plus, Team and Enterprise users to browse and obtain GPTs.

But the store is experiencing quality control issues that might put a dampener on the appeal of GPTs for users, and OpenAI has yet to disclose details of the promised revenue sharing that would incentivise developers.

In a recent piece for Econsultancy, I argued that GPTs will be valuable in doing away with the need for extensive ‘prompt engineering’, allowing anyone to get started quickly with a specialised ChatGPT use case regardless of their knowledge. A particular boon for businesses is that the Enterprise and Teams ChatGPT subscriptions permit GPTs to be shared only within an organisation or team, facilitating the use of internal GPT setups.

Beyond this, however, it seems likely that brands will err on the side of using ChatGPT’s API to build conversational products for consumers, while Plugins and GPTs will remain a benefit primarily for the user.

Further reading

Need training?

Econsultancy runs marketing academies and offers training in AI and machine learning for marketers and ecommerce professionals (including a short course and elearning).