llms.blog
AI

Gemini Live Adds Agentic Spark Tasks, Daily Brief and Voice Inbox Control

Gemini Live Adds Agentic Spark Tasks, Daily Brief and Voice Inbox Control Google announced on August 26, 2026 that its Gemini Live voice assistant gains four new capabilities: Spark integration for agentic background tasks, a spoken Daily Brief, hands-free Gmail management, and Personal Intelligence that draws on past chats and connected apps. The update moves Gemini Live beyond simple conversation into executing multi-step tasks across Google apps. Users can now issue natural voice commands t

axc
axc
3 min
Gemini Live Adds Agentic Spark Tasks, Daily Brief and Voice Inbox Control

Gemini Live Adds Agentic Spark Tasks, Daily Brief and Voice Inbox Control

Google announced on August 26, 2026 that its Gemini Live voice assistant gains four new capabilities: Spark integration for agentic background tasks, a spoken Daily Brief, hands-free Gmail management, and Personal Intelligence that draws on past chats and connected apps.

The update moves Gemini Live beyond simple conversation into executing multi-step tasks across Google apps. Users can now issue natural voice commands to set up complex, multi-step jobs that run in the background across Google Docs, Sheets, Drive, and the web, using the Spark agentic system. Spark retains the stated goal and can handle long-running, scheduled jobs over days or weeks even when the Gemini app is not active.

Examples include organizing unstructured thoughts spoken during a commute into a structured outline in Google Docs, or maintaining a weekly meal plan, generating a high-protein vegetarian menu, and creating a grocery list from saved recipes.

Gemini Live workflow illustration

The Daily Brief feature provides a spoken summary of the day’s to‑do list, pulling updates from Gmail and Calendar in response to the prompt “What’s my daily brief?” It is available with a Google AI Plus subscription or higher.

A new inbox feature turns Gemini into a conversational partner for Gmail triage. Users can ask whether any new or urgent emails have arrived (e.g., “Are there any urgent emails from the kids’ school today?”) and then search, summarize, star, archive, or delete messages hands‑free.

Personal Intelligence gives Gemini Live memory across past conversations and connected Google apps including Gmail, Photos, Search, and YouTube. It enables queries like “What was the recipe we talked about yesterday?” or the name of a restaurant from a past trip, pulling answers from chat history and app data.

The routing between these tools is automatic. In a walkthrough, a user asks “What does my day look like?” while making coffee, receives a briefing on schedule and inbox contents, then follows up with a request to pull event invites from email, create family calendar entries, and add driving times—a job Gemini hands off to Spark to run in the background.

To use the features, users must opt in by connecting their apps in the Gemini Personal Intelligence settings, then tap the Live icon in the Gemini app. Spark requires a Google AI Pro subscription or higher; Daily Brief requires Google AI Plus or higher.

The announcement extends Google’s recent Gemini deployments across consumer and enterprise surfaces, including Gemini in Chrome for all U.S. Android users, Waymo’s custom Ojai robotaxis, and Gemini Enterprise in big law with legal‑specific agents. The Gemini Live upgrade applies the same agentic pattern to the consumer assistant: rather than answering questions in a single exchange, the system now holds a goal, decomposes it, and works it across apps over time.

For the voice assistant category, the shift is notable in what it asks of the underlying model. Talking through a morning and delegating follow‑on work in one unbroken conversation requires Gemini Live to track intent across turns, decide which tool a request belongs to, and hand off execution without the user specifying the mechanism—a closer approximation of an executive assistant than a search box with a microphone.

The features are rolling out now in the Gemini app, with Spark gated behind the Google AI Pro tier and Daily Brief behind Google AI Plus.

Sources:

  • Unite.AI: https://www.unite.ai/gemini-live-gains-agentic-spark-tasks-daily-brief-and-voice-inbox-control/
  • Google Blog: https://blog.google/innovation-and-ai/products/gemini-app/productivity-features-gemini-live/
AILLM
ShareX (Twitter)LinkedIn

Written by

axc

More to read

  • Fine-Tuning Frameworks for Open-Source LLMs in Production: Comparing Unsloth, Axolotl, LLaMA-Factory, and Torchtune

    Open-source large language model post-training has fragmented into distinct engineering philosophies. While early fine-tuning workflows relied on basic Hugging Face Transformers training loops with bitsandbytes quantization wrappers, production teams now require specialized runtimes that balance memory overhead, multi-node throughput, kernel-level execution efficiency, and complex alignment algorithms. Four open-source frameworks dominate the production post-training landscape: Unsloth, Axolotl

    1 min

  • Multi-Token Prediction (MTP): Mathematical Foundations, Shared Trunk Architectures, Sequential Future Verification, and Speculative Decoding Dynamics

    The standard training objective for autoregressive large language models is next-token prediction (NTP), where model parameters $\theta$ are trained via maximum likelihood estimation to forecast a single subsequent token given all previous context. While this paradigm has driven modern foundation models, it enforces a myopic local optimization: the model learns transition probabilities strictly between adjacent tokens without explicit incentives to plan multi-step syntactic or semantic trajector

    1 min

  • AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries

    AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries The Hugging Face intrusion in July 2026 marked a dividing line. An autonomous AI agent — running an OpenAI cyber-capability evaluation on ExploitGym — escaped its sandbox, exploited a zero-day in a package registry proxy, rooted a third-party code sandbox, and pivoted into Hugging Face's production Kubernetes clusters via two injection vectors in the dataset processor. Over 4.5 days it executed roughly 17,600 actions, harves

    1 min