Why some users find Gemini less capable than Google Assistant

Quick answer

Many users perceive Google Gemini as less capable than the original Google Assistant for certain everyday tasks. While Gemini offers advanced conversational and generative AI features, it currently lacks some of the deep system integrations and quick actions that Google Assistant had developed over many years. This can make simple requests, like adding to a grocery list, require more steps or phone unlocks.

What changed with Gemini's arrival

Google began replacing the long-standing Google Assistant with its new Gemini assistant on Android phones in early 2024. This shift introduced a new user interface for accessing the assistant. Pressing the power button or saying 'Hey Google' now often activates Gemini instead of the familiar Assistant.

The change is not just cosmetic. Gemini is built on Google's large language models, designed primarily for complex conversations, summarisation, and content generation. Google Assistant, by contrast, evolved over a decade as a more rule-based system, deeply integrated with Android and Google's services. This fundamental difference means that while Gemini excels in some areas, it struggles with others where Assistant was strong. Tasks such as adding items to a grocery list, controlling smart home devices with specific commands, or accessing 'at a glance' information have been reported as either broken, requiring extra steps, or missing entirely for some users.

Why some features are missing or altered

The perceived regression in capability stems from the different architectures of the two systems. Google Assistant was engineered to integrate tightly with Android's operating system and various Google apps, allowing for actions like creating a note in Google Keep without unlocking the device. Gemini, in its initial rollout, prioritises its large language model capabilities.

This means the model runs primarily on Google's servers, not entirely on the phone itself. While some features may run locally, many require a cloud connection and a more deliberate interaction. Security protocols also play a role; requests that modify data or access personal information might now require the device to be unlocked. Google has stated that it is working to bring back many of the lost functionalities, but this integration is a complex process given Gemini's different foundation.

What Gemini does differently

Despite the reported losses in some areas, Gemini introduces new strengths derived from its large language model core. It is designed to handle more nuanced and open-ended conversations. Users can ask follow-up questions, request drafts of emails or summaries of long texts, and engage in more creative writing tasks.

Gemini's multimodal capabilities are also a key differentiator. It can interpret images and spoken language in combination. For example, a user could show Gemini a photo and ask questions about its contents. These functions represent a move towards a more sophisticated, context-aware AI interaction than the previous Assistant offered, focusing on understanding and generating complex information rather than just executing predefined commands.

What it changes for the reader

For a reader, the shift to Gemini means adjusting expectations and workflows. Simple, routine tasks that were once quick and seamless may now require more effort, or they might not work as expected. Users accustomed to hands-free convenience for basic commands may find themselves needing to pick up and unlock their phones more often.

Conversely, those who frequently need help with writing, summarising, or engaging in detailed conversations with an AI may find Gemini to be a powerful upgrade. The trade-off is often between the efficiency of old, deeply integrated features and the new, broader generative capabilities. Deciding if the change is an improvement depends heavily on how a user typically interacts with their phone's assistant.