The ChatGPT App Store Just Opened to 800 Million Users. Should Your Small Business Build One, or Just Get Found Inside One?

Every few months, a platform flips a switch and every small business owner gets the same anxious question in their inbox: "Do I need to be on this now?" This time it's ChatGPT. OpenAI quietly opened its App Directory to third-party developers, apps are now live inside the sidebar at chatgpt.com/apps, and the platform sits in front of more than 800 million weekly users. That's a bigger built-in audience than any app store launch in small business history got on day one.

Here's the part the hype cycle keeps skipping: as of this month, roughly 100 companies have a ChatGPT app, and exactly one of them — Instacart — can actually process a sale inside it. Everyone else is either not selling anything through the app yet, or routing people back out to their own website to pay. If you've been feeling pressure to "build a ChatGPT app" before your competitor does, the honest answer is: slow down, because there's almost nowhere for that app to make you money today. But there is something worth doing right now, and it's not what most of the advice floating around is telling you.

What Actually Launched, in Plain English

OpenAI's Apps SDK lets developers build mini-apps that live inside a ChatGPT conversation — triggered when a user types "@" and your app's name, or pulls it from the tools menu. Under the hood, it runs on the open Model Context Protocol (MCP), the same connective tissue a lot of AI tooling is standardizing around this year.

• Submissions opened December 17, 2025, with approved apps rolling out starting early 2026 — this isn't brand new, but October 2026 is when the gap between the hype and the reality has become undeniable.
• Reach: 800+ million ChatGPT users can discover and launch an app without leaving their conversation.
• Requirements are strict but not exotic: a privacy policy, a support contact, general-audience content, compliance with OpenAI's usage policy, and no advertising or prohibited commerce categories.
• About 300 integrations exist today, and OpenAI itself admits most users can't find them. Discoverability, not submission, is the actual bottleneck.

The Monetization Wall Nobody's Advertising

This is the detail that should change your decision, not the user count.

• Physical goods can be sold through outbound links. Digital goods, subscriptions, and in-app services are explicitly "not yet allowed," per OpenAI — the company says it's "exploring additional monetization options over time," which is corporate-speak for "no committed date."
• No revenue-share terms have been published, despite OpenAI floating the idea back in 2024. If you build something that works, you still don't know what OpenAI's cut would be if monetization ever opens up.
• Instacart is the lone checkout example, built directly with OpenAI as a launch partner with engineering resources no small business has. Everyone else building an app today is doing it on faith that the rules change in their favor later.
• Developers who already shipped apps are reporting the real friction points out loud: low traffic conversion from the mini-apps, opaque engagement metrics, slow review timelines, and real consumer hesitation about entering payment details inside a chat window at all. OpenAI's own spokesperson admitted in October that "we're still early in building this out, and we recognize there are areas where the developer experience needs to improve."

That's not a platform ready for a small business to bet its Q4 roadmap on. It's a platform worth watching closely and positioning for — a different thing entirely.

The Smarter Move: Get Found, Don't Build Yet

Here's where this connects directly to the AEO and GEO work we talk about constantly: the problem OpenAI has with its own App Directory — 300 integrations that almost nobody can find — is the exact same problem your business has inside every AI assistant right now. Being technically present isn't the same as being discoverable. We laid out the mechanics of winning that discoverability fight across platforms in our GEO strategy breakdown, and it applies here directly: ChatGPT, Gemini, and Perplexity are all becoming places customers ask questions and expect an answer to just appear, not places they browse app stores.

So instead of racing to build a ChatGPT app with no clear way to get paid, do the three things that actually move the needle this month:

• Check whether a tool you already pay for just got a free ChatGPT integration. OpenAI's separate "ChatGPT for Small Business" program, launched this past July, ships curated partner plugins from Dropbox, Shopify, Intuit, Slack, Atlassian, and Wix — with exclusive offers attached. If you already run any of those, you may have a working integration sitting unused.
• Audit whether your business is answerable, not just searchable. Structured service pages, clear pricing, a complete Google Business Profile, and schema markup an AI agent can parse are what gets you cited inside a ChatGPT answer today — no app submission required, no monetization wall, no 300-integration haystack to get lost in.
• Treat "build a ChatGPT app" as a 2027 roadmap item, not a 2026 emergency. Revisit it when OpenAI publishes real revenue-share terms and digital-goods support. Until then, every hour spent on a submission is an hour not spent on the discoverability work that pays off on every AI platform at once, not just one.

This is the same build-vs-buy-vs-wait discipline we walked through with no-code AI app builders a few weeks back — that piece was about owning custom internal software; this is about showing up inside someone else's platform. Different decision, same rule: don't spend real money chasing a feature that isn't finished yet.

Who Should Actually Move Now

Not every business should sit this one out. A few situations where building toward a ChatGPT app earlier makes sense:

• You sell physical products with a clean checkout flow already. Outbound-link commerce works today, so if ChatGPT shoppers finding your product and clicking through to buy is a plausible path for your category, there's upside in testing it early while competition for attention inside the Directory is still thin.
• You have engineering capacity sitting idle anyway. If a technical hire or agency relationship has slack this quarter, prototyping an app now — knowing it won't monetize yet — buys you position once the rules change, and lets you learn the MCP tooling before it's crowded.
• Your customers are already heavy ChatGPT users and ask it product questions. If you know that from support tickets or sales conversations, presence matters even without a transaction happening inside the chat.

For everyone else — the majority of small businesses reading this — the discoverability work is the higher-leverage move, and it's work you should already be doing regardless of what OpenAI decides next.

FAQ

Can my small business actually sell products inside a ChatGPT app right now?
Only physical goods, and only by linking customers out to your own checkout. Digital goods, subscriptions, and in-app payments are not yet allowed on the platform, and only one company — Instacart — currently has in-app checkout live.

Is it worth submitting an app to OpenAI's ChatGPT App Directory today?
For most small businesses, not yet. With no published revenue-share terms, 300 integrations most users can't find, and OpenAI itself calling the developer experience early-stage, the return on the time investment is unclear. It's worth more as a 2027 planning item than a Q4 2026 project, unless you already sell physical goods with a simple checkout.

What's the difference between OpenAI's "ChatGPT for Small Business" program and the App Directory?
They're separate initiatives. The small business program (launched July 2026) is training, in-person events, and curated partner plugins from companies like Shopify, Intuit, and Slack — no app-building required. The App Directory is where developers submit custom apps built on the Apps SDK. Most small businesses get more value, faster, from the first one.

If I'm not building a ChatGPT app, how do I make sure my business still shows up when customers ask AI assistants about it?
Focus on answer-shaped content, structured data, a complete and accurate Google Business Profile, and clear service pages — the same fundamentals that drive AEO and GEO performance across ChatGPT, Gemini, and Perplexity. That work pays off everywhere at once, unlike a single-platform app.

The Bottom Line

OpenAI just handed 800 million people a new way to discover businesses inside a chat window — and then quietly admitted the discovery problem isn't solved, the money isn't flowing yet, and the developer experience still needs work. That's not a reason to ignore ChatGPT apps. It's a reason to stop treating "build an app" as the urgent move and start treating "be the business that actually gets found and cited" as the one that compounds regardless of which platform wins the app-store war.

That's precisely the kind of platform-risk call cCoreVentures' AI Strategy & Implementation practice makes for clients every week — which emerging AI surface is worth building for now, which is worth watching, and which discoverability work pays off no matter how the platform race shakes out. If you're wondering whether a ChatGPT app belongs on your roadmap or your AEO fundamentals need attention first, let's map it out before you spend a sprint on the wrong one.