Last week I watched someone ask ChatGPT to find a cheap flight. Expedia opened — inside the chat window. No browser tab. No copy-pasting. Full booking interface, right there.
That was the moment I stopped treating OpenAI's Apps SDK as a developer story.
This is a distribution story.
Since late 2025, ChatGPT has had a live app ecosystem. Expedia, Spotify, Canva, Booking.com, Adobe, Zillow, DoorDash — all integrated via the Apps SDK, all surfaceable within a conversation. Not as links. Not as search results. As first-class apps that open inside the chat when ChatGPT decides they're relevant.
And that decision — which app shows up, when, for whom — belongs to the model, not to you.
Here's what makes this different from every other platform distribution battle.
In the App Store, users look for your product. They search, browse, read reviews, and choose. In ChatGPT's app ecosystem, users don't search at all. They describe a problem. "Help me plan a trip." "I need to design a banner." "Find me a rental near downtown." ChatGPT maps their intent to an app and surfaces it without the user ever asking for a specific product by name.
That's a new discovery paradigm. SEO was about keywords. App stores were about intent matching in a search interface. This is conversational intent — and you have very limited leverage over how the model interprets it.
I work in vertical SaaS. Healthcare software: clinics, practitioners, scheduling.
Here's the thought experiment I've been running: a patient opens ChatGPT and types "I need help booking an appointment with a specialist near me." What happens?
At best, nothing specific comes up and ChatGPT gives generic guidance. At worst, a competitor who built a ChatGPT app owns that conversational touchpoint — and gets the lead that should have been yours.
Test it yourself. Open ChatGPT, describe the exact problem your product solves, and see what surfaces. If your product doesn't appear and a competitor does, that's not just a brand awareness gap. It's a distribution gap — and those compound.
Product teams have navigated two major platform wars in the past 15 years.
The App Store moment in 2008 was swift and unequal. First movers got organic discovery, early reviews, and compounding word-of-mouth before the market filled up. Late entrants paid significantly more for the same visibility — through ads, influencer deals, and PR pushes.
The SEO war was slower but just as decisive. Eventually, every web-based product had to reach an equilibrium with Google: content, backlinks, or paid search. There was no opting out.
Now there's a third one.
What makes it different: the rules aren't fully settled yet. The Apps SDK is still in preview. The algorithm that matches user intent to app recommendations isn't transparent. Standards are forming in real time.
That ambiguity can be read two ways: "It's too early, let's wait." Or: "It's early enough that entering is still relatively cheap."
The honest complexity here: the brands already in ChatGPT's app ecosystem — Expedia, Adobe, Spotify — have resources that most vertical SaaS companies don't. The real cost of building and maintaining a quality ChatGPT app for a smaller company is still unclear.
But platform wars don't wait for smaller players to get ready.
The companies that moved first in the App Store era didn't just accumulate downloads — they got more data, faster feedback loops, better retention signals. They built moats from that compounding advantage. The second and third cohorts of entrants spent much more to achieve less.
The same dynamic tends to replay.
Here's the question I keep coming back to: if someone typed your product's core use case into ChatGPT right now, would anything happen?
Not "would your website come up in a search" — that's a different distribution channel. Would your product appear as a ChatGPT app? Would it be suggested?
If the answer is no, you're not necessarily late yet. But you're watching the third platform war from the sidelines.
Three years from now, someone will ask whether being in ChatGPT's app ecosystem in 2026 made a difference. Most answers will include the phrase "we should have moved sooner."