Why Your AI Can Now Work the Night Shift

Here's something AI couldn't do for you last week: nothing. Not while you slept, anyway. You opened a chatbot, asked a question, got an answer, closed the tab. The work happened only when you were sitting there making it happen.

Google just changed the verb. At I/O 2026, it shipped a way to tell an AI agent: every night at 2 AM, do this — and have it waiting for me by morning. One command. It runs on a clock. You go to bed.

TLDR: Google's new Antigravity 2.0 lets you schedule AI agents to run recurring tasks in the background — like a night shift you never have to staff. Today it's built for people who work in code, but the exact same engine powers Gemini Spark, a 24/7 personal agent aimed at everyone's inbox and calendar. Here's how the night shift works, and what's coming for the rest of us.

Most e-commerce sellers are running their store across 6 to 8 separate tools — and paying hundreds of dollars a month for the privilege. StoreClaw replaces your entire stack with one autonomous AI engine that monitors competitors, optimizes listings, automates marketing, and tracks real profit across Shopify, Amazon, and beyond.

It doesn't wait for you to ask. It runs 24/7 in the background, so you wake up to a full dashboard instead of a list of things you forgot to check.

Connect your store, and StoreClaw gets to work — no prompts, no complex setup, no six-app stack.

Free to start. No credit card required.

Try it for free today

The feature has an unglamorous name — Scheduled Tasks — and a genuinely new idea behind it. Inside Antigravity 2.0, you type a /schedule command and hand the agent a standing instruction tied to a clock. It then runs on its own, in the background, on Google's servers. No app open. No prompting. No you.

Google's own demo instruction is plain enough to copy: "Every morning at 9 AM, run tests on yesterday's code, and check which dependency packages need updating." Swap in your own chore and the shape stays the same — a task that used to need a human at a keyboard now runs unattended and leaves a result behind.

A few of the recurring jobs this unlocks:

The overnight health check. Point an agent at a project and have it run every night, checking what broke since yesterday and flagging anything risky. You wake up to a "here's what needs your eyes" list instead of hunting for it.

The morning briefing. Schedule an agent to scan everything that changed in a body of work overnight and write you a plain-language summary. The first thing in your inbox is a wrap-up of where things stand — not a blank page.

The silent watchman. Set a recurring scan for problems you'd never catch by hand — outdated pieces, weak spots, things that quietly drifted out of date. It looks while you're not looking.

The tidy-up crew. Hand off the repetitive maintenance nobody wants — the cleanup passes, the routine updates, the boring-but-necessary chores — to an agent that does them on a schedule and reports back.

The thread running through all of it: you stop doing the recurring task and start receiving the finished result. That's the night shift.

Quick orientation, because the launch came with a confusing pile of names. Antigravity 2.0 is one platform with several ways in. The desktop app is the scheduling dashboard described above. The Antigravity CLI, IDE, and SDK are the same engine accessed three other ways — through a terminal, a code editor, or custom-built software. Those three are for developers and the technically inclined; if that's not you, you can cheerfully ignore them. The scheduling magic lives in the app, and the app is free to download on Mac, Windows, and Linux.

<!-- SCREENSHOT: Antigravity 2.0 desktop app — the agent dashboard / mission-control view. Caption suggestion: "Antigravity 2.0's desktop app — where you point agents at work and set the schedule." -->

Here's the part that makes this more than a developer story. The engine running those scheduled agents — Google calls it the "agent harness" — is the same one powering Gemini Spark, the headliner Google unveiled at the same event. Spark takes the night-shift idea and points it at your everyday digital life instead of a codebase.

Spark runs 24/7 on Google's cloud, so — in Google's words — you don't need to keep your laptop open for it to work. It plugs into Gmail, Docs, and Slides, and connects out to Canva, OpenTable, and Instacart to handle recurring tasks and run start-to-finish workflows — drafting a status email to your boss, pulling a report together from scattered sources, handling the standing stuff you'd normally chip away at yourself.

The honest catch: Spark is locked behind Google's AI Ultra plan at around $100/month, US-only and English-only, currently rolling out in beta. It is, as one writer put it, not a product for most people today. But it is a clear signal of where every assistant is heading — and Google has a long habit of launching a feature high, then trickling it down to cheaper tiers over the following year. The background agent you can't quite justify today is the one that shows up in your regular apps later.

You don't need a $100 plan or a single line of code to start thinking in night-shift terms. The skill that makes scheduled agents pay off is knowing which of your recurring tasks is worth handing to one — and most people guess wrong. This prompt interviews you, then builds your personal automation shortlist. Paste it into Gemini, ChatGPT, or Claude:

The Night-Shift Audit →

The chatbot era asked, what can you tell me? The agent era asks a better question: what can you go handle? The night shift just opened. The only question left is what you'd hand it first.

This guide from Deel breaks down how to build one global hiring system. You’ll learn about assessment frameworks that scale, how to do headcount planning across regions, and even intake processes that work everywhere. As HR pros know, hiring in one country is hard enough. So let this free global hiring guide give you the tools you need to avoid global hiring headaches.

Download the free guide today