Is ChatGPT going to be your next marketplace?

How marketplaces evolve

According to Brian Belfour of Reforge, digital distribution channels or marketplaces undergo four different phases in their lifespan. First comes the emergence of new technology, and then a new distribution channel opens up. The four steps are: 

(1) establishing a new technology category,
(2) building a moat to secure category leadership,
(3) platform opening for third-parties for maximum reach, and 
(4) platform closing for monetization by the platform owner. 

We’ve seen this cycle play out in several well-known services. Facebook, LinkedIn, Google, and even Spotify have had periods of openness. Eventually each has reverted to a closed environment in which all visibility for third-party distribution must be paid for. For example, Google Maps remains a good place to find Western points-of-interest. To be highlighted on the map, however, takes promotion – not just great user reviews. 

With each new transformational generation of technology, a new marketplace has usually formed. I argue we are now facing such a moment with LLM-driven AI systems such as GPTs. The reasons are economic. To spur the feeling of infinite growth, the platform owners attract new crowds (users and service providers) with new use cases offered by third-party vendors. They aim to push for commercial transactions or maximize engagement for ad-based revenue models.

With new users and use cases, arrive the new opportunities to monetize on shopping needs and consumer time. 

The ChatGPT marketplace moment

So far, the major LLM chat platforms have been quite closed systems. Within the past year, LLM providers have realized that opening up more context – that is, connecting with users’ other data sources – is key. In fact, it’s central to improving model performance. Initiatives such as the Model context protocol (MCP) enable users to feed the LLM new material. They haven’t transformed the platform more than an API does, however. LLM APIs have been around for some time.

Some goes for custom GPTs or Gems (Gemini equivalent). Personalized GPTs have neither made a major breakthrough nor opened new venues for monetization for their builders. In my opinion, they don’t really count as opening moves in the great game we’re awaiting. 

What we are waiting for is a marketplace for ideas, products, software, and who knows what, that is deeply integrated within an LLM. This is now inevitably happening in a movement known as agentic commerce. Agentic commerce has both technical components (OpenAI’s Agentic Commerce Protocol or Google’s AP2) and a customer-facing part, such as ChatGPT’s Instant Checkout. 

ChatGPT just announced that Etsy sellers are being rolled out as the first real-life test group of the Instant Checkout feature. More vendors, such as the whole Shopify platform, are soon to follow! Especially for subscribers whose credit card is already on the file, shopping could be even easier than with Amazon’s One-click ordering. Just say ‘yes’, and the order is in.

For now, OpenAI describes that the visibility of the products will be organic and depend on “relevance to the user”. This sounds awfully lot like Google’s Pagerank used to be. But what is expected to follow?

Reflecting on Belfour’s steps, this would look like:

1. LLM chatbots become a thing ✓
2. OpenAI gains an advantage by amazing acquisition and product availability ✓
3. OpenAI provides mechanisms for third-parties to sell products and services on ChatGPT ✓
4. OpenAI limits the distribution and increases its share of revenue on all third-party sales, possibly with charging money for the mere visibility?

How it will begin and how it will end 

In the interview on Lenny’s Podcast, Brian Belfour explains how the new marketplace momentum builds and how it also dies. Particularly, the moment of birth is exciting, because it offers a huge opportunity for the first movers on the platform to reach audiences they could not (afford) earlier. Just imagine putting your digital product in the hands of 800 million weekly active ChatGPT users!

But the pressure to move with top speed will also be remarkable. If the new marketplace follows the pattern of its predecessors, there will be a clear divide between winners and losers of the new attention share delivered by the LLM. In Balfour’s words, small players must place all their bets on the table immediately. It’s a high-risk, high-reward situation. You win big, or lose it all. 

The feeling of urgency is not artificial in emerging marketplaces. It reflects the reality that the platform owners tend to be more generous in their cuts early on in the game. When the novelty of the marketplace wears off and the benefits of user acquisition for the platform owner have been consumed, it is time to monetize. The focus shifts to the use cases that create real revenue. What follows are falling margins for third-parties – or even a totally collapsing business, if the platform decides to compete within the category they’ve found big enough. 

Is this relevant for your company?

The new AI-driven marketplace scenario is not yet fully here. Nobody knows exactly how it will be revealed. OpenAI’s first gambit seems limited to the US, so far. Their previous shopping experiments were also geographically limited, but they naturally have bigger plans than that.. 

Companies that can provide goods or services globally will be best positioned to enter the new marketplace when the gates open. For instance, here in Finland, most incumbents probably don’t feel a great urgency to follow suit. We’ve got plenty of businesses doing just fine in the local, relatively enclosed market. 

But this doesn’t mean that the new distribution channel could not affect domestic niche markets. For example, large global insurance companies might find instruments such as life insurance appealing. These can be sold with little difficulty (don’t take my word about the simplicity, though) across borders with new channels. We have seen this happen with Temu, we might see it again on your go-to LLM chat. Given the superior lead ChatGPT has over its competitors, it seems like a safe bet to argue that OpenAI is the one to watch out for if you are ready to race for the new platform era. 

This brings us to the new witchcraft that concerns the challenge “how do I make my products seem relevant to ChatGPT users algorithmically?”. The emerging art of AI engine optimization (AEO or similar) is there to address this part, but that’s a different story (See Ethan Smith’s interview, for example.)

 

Sources:

https://www.voguebusiness.com/story/technology/chatgpt-launches-new-shopping-features 

https://hbr.org/2025/05/could-the-gpt-store-turn-chatgpt-into-a-platform-powerhouse