Google has introduced AI Edge Foresight, a new AI-powered meeting assistant designed to record, transcribe and organise conversations locally on a user's device, expanding its portfolio of artificial intelligence tools for workplace productivity.
The application, currently available for Mac, allows users to capture meeting discussions, generate AI-assisted notes and retrieve information from conversations without relying on cloud processing or a continuous internet connection.
Developed by the team behind Google's experimental AI-powered dictation application released earlier this year, AI Edge Foresight uses the company's on-device EmbeddingGemma 2 model, which contains 740 million parameters.
The application is optimised for Apple Silicon and can operate entirely offline, according to Google. This allows users to access meeting notes and related information while travelling or working in locations with limited connectivity.
AI Edge Foresight supports meeting capture across applications, including in-person conversations. Users can take manual notes while the software transcribes discussions and generates structured summaries.
The interface includes a split-screen layout, allowing users to maintain their own notes alongside AI-generated content. The approach is similar to existing AI meeting assistants such as Granola, which combines user-written notes with automated transcription and summarisation.
Google's application also provides access to complete meeting transcripts and includes a conversational assistant powered by Gemma 4. Users can ask questions about previous discussions and retrieve relevant information from recorded meetings.
Beyond transcription, AI Edge Foresight allows users to create a searchable knowledge base by importing documents and other reference materials.
Supported formats include PDFs, Google Docs, Microsoft Office files, plain text, Markdown documents and saved web bookmarks.
The application can use information from these materials alongside meeting transcripts to answer questions during ongoing conversations. For example, users can retrieve details from previously uploaded documents when related topics arise in a meeting.
The local-first approach differentiates the application from AI meeting assistants that depend primarily on cloud-based processing. By running its models on the device, Google is demonstrating how meeting transcription and information retrieval can function without an internet connection.
The launch comes amid growing competition in the AI meeting productivity segment. Companies including Wispr and Calendly have recently introduced their own meeting assistants, while Superhuman, formerly Grammarly, acquired AI note-taking startup Fathom in September.
Google has not announced whether AI Edge Foresight will become a mainstream consumer product or be integrated into its existing Gemini applications.
The release follows the company's broader development of smaller AI models capable of running directly on personal devices.
With AI Edge Foresight, Google is extending these capabilities into meeting documentation and knowledge management, providing an alternative approach to AI-assisted workplace productivity that prioritises on-device processing.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.