Gemini is becoming more than a chatbot: Google wants your apps to work inside it

Smartphone screen displaying the Google Gemini app store page | Image Credit: MARCO, Unsplash

This week, Google announced that it is breaking down the barriers separating Gemini from the other programs that people use on a regular basis. The company introduced the Connected Apps Platform, a system that lets your favorite applications work directly inside the chat interface.

Instead of bouncing between Gemini and separate apps to complete tasks, users will soon be able to book restaurant reservations, order groceries, or check fitness data without ever leaving the conversation. This is Google’s bet that rather than being a separate destination that users would visit, the future of AI will be a hub where work is accomplished. This announcement came amid intense competition.

Google wants Gemini to reach beyond its own tools, offering a new publishing route for app developers

ChatGPT established the consumer AI chatbot category, but Google’s Gemini has been growing faster. With 1 billion monthly active users, up 100 million in a single quarter, Gemini is narrowing the gap against ChatGPT’s estimated 800 million weekly users. What matters most is what users can actually accomplish without friction.

Since these figures only provide a part of the picture, Google is relying on Gemini’s seamless integration feature to help tip the scales with this move. Historically, getting a service into Google’s AI meant striking a deal with Google directly. The company would collaborate on a custom integration with the appropriate engineering team.

This arrangement favored large corporations with dedicated resources and could leave out start-ups and mid-market companies. Google has now dismantled this bottleneck by creating a standardized onboarding pathway.

Developers can now use Google’s management interface, Google Connected Console. Here, they can register their application and monitor and test its performance. The console displays metrics that matter, like how many people are using the service through Gemini, how fast responses arrive, and whether errors occur.

This allows developers to maintain control over branding and user permissions via the self-service model. Fifty-plus applications are already live through the system. Adobe handles creative tasks. Experian manages financial inquiries. Peloton tracks workouts. Instacart fulfils grocery orders. OpenTable processes restaurant reservations.

When they’re connected through Gemini, these services behave as if they’re built into the assistant rather than bolted on top. Users can intentionally activate them by using the @ symbol, or they can be mentioned organically in chats. Nothing requires users to exit the conversation.

ChatGPT offers plugins, but using them feels like accessing external tools that live within the interface. The primary feature and supplementary features are distinguished by the user experience. Google’s architecture collapses this distinction.

The appeal depends on what Connected Apps can actually do

With this experience, Google is supporting Anthropic’s open standard, the Model Context Protocol. Replit, Sourcegraph, and Apollo already support MCP. For developers, this choice carries practical weight. By creating an MCP integration, services may be able to contact people on several AI platforms.

The choice is no longer between Gemini, Claude, or other MCP-compatible assistants. The developers build once and distribute broadly. Google is rolling out the platform gradually through a U.S. developer waitlist. This controlled approach lets the company manage quality, identify problems, and refine the system before global expansion. Developers can register at the official Connected Apps website.

Google owns search, where billions of queries happen daily. Google also owns Gmail, Calendar, Photos, and Android. This advantage reinforces itself.

More data guides development priorities. Each new connected service increases Gemini’s utility, which attracts more users. Meanwhile, OpenAI lacks the foundational assets, like Android, Gmail, and search, that would be required to match this approach. These services are the core of Google’s business and could not be licensed to competitors.

Sources: TechCrunch, The Motley Fool

Affiliate links on Cult of Mac may earn us and our partners a commission.

Comments (0)
Add Comment