Gemini Connected Apps are becoming part of a broader Google system that could eventually let people use third-party services from both Gemini and AI Search. On 28 October 2026, Google announced a developer platform intended to bring those connections together. It also named services already arriving in Gemini, while saying further availability in Search would follow in the coming weeks.
The company’s launch announcement mentions Adobe, Experian, Localiza and Peloton among partners with integrations live for users in Gemini. It says Calm is due to arrive across Search and Gemini. Availability differs by app, product and location. The announcement is therefore not a promise that every person can already ask Google Search to use every named service.
What Gemini Connected Apps are meant to do
Imagine wanting help with an everyday task that cannot be finished by generating text alone. Finding something is one part of the job. Opening the appropriate service, passing the right information and obtaining a useful result may be another. Connected Apps aim to reduce that gap by letting the assistant work with approved third-party tools rather than simply telling someone where to go next.
The appeal is easiest to see in activities people already understand: organising a workout, looking up a business or working with a creative application. A connection could help an assistant move from describing a possible action to interacting with a service designed to perform it. Which actions are actually available depends on the individual integration. A logo in a catalogue is not evidence that every possible task is supported.
Google says it has launched more than 50 Connected Apps across Search and Gemini in areas including productivity, ticketing and fitness. That total refers to the wider ecosystem, not 50 new integrations announced on the same day. Existing tools, new services and future planned connections need to be kept distinct when judging how much has changed for ordinary users.
One connection for developers, several places for users
The new platform is partly aimed at the people who build apps. Previously, developing an assistant integration could mean separate technical work for different destinations. Google wants a unified route through which a developer can build and manage a connection intended for Gemini, Search or both. That could make integrations more economical to maintain, but it remains a platform promise rather than a universal result.
According to Google’s developer documentation, integrations can use a cloud-hosted connection based on the Model Context Protocol, or Android mechanisms that let apps expose suitable functions. Developers register and test connections through Google’s console before submission. That infrastructure is the less visible part of the story, but it helps explain why Google expects the number and variety of available services to grow.
The platform also gives developers monitoring information, including traffic, latency and errors. Those details matter when a service is asked to do real work. A connection that works in a demonstration but repeatedly fails during an actual task is unlikely to build confidence. Companies supplying integrations will need to think about maintenance, account access and whether the correct result reached the user.
More convenient does not mean more automatic
A connected service often involves an account, permissions or personal information. A fitness application may hold activity details; a shopping account may hold addresses and payment preferences; a creative service may contain unpublished work. A connection makes those boundaries more important, not less. Users should see what they are authorising and have a way to stop using an integration they no longer want.
The important distinction is between asking Gemini about an app and giving Gemini permission to use that app. The first can be little more than general assistance. The second may involve a service receiving data or carrying out an action. Anyone evaluating the feature should look at the available controls for the particular product and account rather than assuming that a general description explains every permission.
LiveAIWire has examined the permissions problem for AI shopping agents. The same principle applies here: the more an assistant can do beyond providing information, the more carefully people need to distinguish observation, recommendation and execution. An easy route into another service should not remove an equally easy route to review or revoke access.
How this differs from Google's existing work assistants
Google already has AI tools designed to carry out work inside connected software, including the Gemini agent capabilities covered in LiveAIWire’s earlier enterprise report. The new Connected Apps story is different. It concerns a common developer route across consumer-facing Google surfaces and a fresh wave of third-party partners, not merely a renaming of an existing workplace assistant.
That difference may matter to the audience. People do not always think in terms of model providers or software architecture. They think in terms of whether an app can perform the job they have in mind without forcing them to repeat information across several screens. If the platform succeeds, the visible experience might be a shorter route to a familiar service, even though the underlying connections are technically complex.
There is a commercial question too. Developers want their services to remain recognisable and useful when a large assistant sits between them and the customer. Google says its system lets participating providers retain verified branding and manage their connection. How discovery, competition and customer relationships work in practice will become clearer as more integrations reach Search.
For now, the verified news is a developer preview and a specific set of current and planned app connections. The bigger ambition is a Google assistant that can do more with the services people already use. The test is whether the controls, reliability and regional availability become as clear as that ambition sounds.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity and the social impact of emerging technology. LiveAIWire is an independent, human-led technology publication using AI-assisted research, editorial production and original AI-assisted editorial illustrations under his direction.