Google Connected Apps and the App Store You May Never See

The choice before the choice

Before a developer’s Connected App reaches anyone, Google says the developer chooses whether to submit it for Search, the Gemini app, or both. Then comes Google’s review. That sequence interests me: the developer chooses where to knock, while Google decides whether the door opens. On October 28, Google launched a U.S. developer preview, available through a waitlist, for bringing more outside app capabilities into those products. ^{[1] [2]}

The demonstrated change is plain: developers can now use one Google platform to register, test and submit app capabilities for Search and Gemini. Google says more than 50 Connected Apps have launched across areas including productivity, ticketing, health and fitness. Availability is uneven: some integrations are already in Gemini, while Google said several were due to reach Search in the following weeks. The development platform is new; connected integrations themselves are not entirely new. ^{[1] [2]}

The plumbing carries two names most people should not need before breakfast. Developers can connect through a cloud-hosted MCP server or through Android’s AppFunctions APIs, then use Google’s console to prepare the integration. In the examples reported by TechCrunch, that connection can let someone ask a service to make a flyer or build a playlist from within a Google product rather than opening the provider’s app first. ^{[1] [2]}

For an ordinary person, the appeal is wonderfully untechnical. Ask for the outcome and avoid the small obstacle course of remembering which app, finding its icon and moving the task across several screens. Friction is real, even when it arrives dressed as three harmless taps. I think many people will gladly trade those taps for a sentence—provided the sentence does what they meant, not merely what the system found convenient. ^{[1] [2]}

Direct becomes indirect

Here is the reversal: the route to the task becomes more direct, while the relationship with the company doing it may become less direct. A service can remain essential yet become less apparent. You may know that Gemini understood your request while barely noticing who fulfilled it. That does not mean apps are dead or providers disappear. It means capability and visibility can separate, and attention tends to gather where the request is understood. ^{[1] [2]}

Google says developers can retain verified logos and authenticated accounts while receiving reports on traffic, latency and error rates. Those are meaningful provisions for identity and performance. But the captured announcement does not specify discovery or ranking rules, economic terms, rejection criteria, or why one connected service might be presented rather than another. “App store” is therefore an analogy, not Google’s stated product category. We have the doorway, but not yet the map of the hallway. ^{[1] [2]}

It would be premature to claim that Gemini is already choosing every shop on our behalf. The evidence does not establish that. The consequential possibility is narrower: if people increasingly begin with an AI request rather than a named provider, the interface interpreting that request gains influence over which options they encounter. The hidden cost of saved time could be fewer visible moments to compare price, privacy, quality or simply another company. ^{[1] [2]}

Who gains agency? Someone who wants a routine task completed quickly may gain plenty. A small developer may reach people without first winning a permanent square on their phones. Google gains a central position between request and result. The provider may gain distribution while losing recognition. That is the peculiar bargain: a business can become easier to use and harder to remember. Distribution expands; the customer relationship may contract. ^{[1] [2]}

More doors, same question

There is a serious counterargument. Google says developers can build for both Search and Gemini, choose where to submit, preserve their brand identity and monitor performance. Separately, binding European Commission measures issued in July require Google to give competing AI assistants access to key Android functions for EU users, including activation by voice and actions inside apps, subject to privacy, security and device-integrity safeguards. ^{[1] [3]}

That could produce more competing gateways rather than one dominant assistant. If people can choose another capable assistant, some power moves back toward them and toward developers seeking alternative routes to customers. But assistant-level competition does not automatically create transparent choice inside each assistant. Several doors into the shopping centre are useful; they do not tell you why one shop appears at the entrance and another is tucked behind the lift. ^{[1] [3]}

I would use a simple rule: the more consequential the task, the more visible the provider should be. A playlist can tolerate some mystery. Health, credit, travel or an expensive purchase deserves a pause. Ask which service is acting. Notice which account is connected. When terms, price, privacy or judgment matter, visit the underlying provider and compare alternatives directly. Convenience should shorten the journey, not remove the signposts. ^{[1] [2]}

Google’s preview may become a useful bridge between what we ask and what software can do. It may also become another doorway businesses must pass through to reach us. The platform is young, broader availability is undated, and its discovery rules remain unspecified. So the durable question is not whether apps vanish. It is whether, when the app becomes less visible, our choice vanishes with it—or whether we remember to look through the doorway before walking on. ^{[1] [2] [3]}