文章摘要
AI进入办公领域后,传统办公软件操作门槛高、跨应用切换繁琐等痛点凸显,360推出定位AI时代新办公入口的NanoWork(纳米Work),其整合多格式文件编辑、多模态AI创作能力,搭载多智能体协作执行框架可跨工具自动执行工作流,依托360安全能力做全流程风险防控,适用群体覆盖知识工作者、小微商家等。

For over three decades, Office software has been an irreplaceable staple in global workplaces, with its suite of tools defining the modern office ecosystem and creating one of the most robust productivity platforms in software history. But as AI enters the workplace, the long-standing barriers of mastering these tools are beginning to crumble.

Users once spent hours learning software tricks to fit their work around the tools, but now AI agents can understand tasks, call tools, and even operate computers directly. This shift is rewriting the rules of office entry, and 360's NanoWork is positioned as a key player in this new era of AI-powered office solutions.

From Office to NanoWork: Reclaiming the Work Entry Point

Even with AI assistance, users still face persistent frictions: they must switch between apps, move content manually, and handle final adjustments on their own. An AI-generated marketing copy might take minutes to format correctly, or a poster with a typo requires rework instead of easy edits. NanoWork solves this by centralizing all these scattered operations into a single platform.

It integrates common file types like Word, Excel, PPT, and PDF, while also adding AI editing capabilities for images, posters, and videos—emphasizing that every element can be modified. This means people and AI can iterate on the same work continuously, cutting down on copy-paste and repeated exports. The "super canvas" feature lets users adjust images locally, fix text, or tweak composition without starting over, while the "super video editor" turns scripts, storyboards, voiceovers, and clips into a single editable workflow.

What's most impactful here is not just the initial generation, but the ability to revise efficiently. Real work is full of changes, feedback, and adjustments, so a tool that supports ongoing edits is the only one that can fit into a full work process. Industry experts have noted that beyond just models, the execution system is critical for complex tasks. For video work, this means being able to adjust just the voice, replace a single clip while keeping others intact, or update the script without disrupting the rest of the process.

Packing AI Teams into Editors: Reallocating Creative Capabilities

This shift in editing also reflects a broader change in how creative work is structured. Previously, coding was seen as a separate skill, but now it's becoming a new entry point for visual creation. For example, making a product animation used to require designers to build layers and keyframes in complex tools, but now code can directly define element positions, movement, and rendering rules.

NanoWork's video editor takes this a step further, breaking down professional video creation into tasks handled by specialized agents: a script agent for content, a storyboard agent for visuals, a voice agent for audio, and a clip agent for editing, all coordinated by a unified Harness system. This Harness framework acts as the execution backbone, handling task breakdown, tool calls, context management, state saving, and recovery if something goes wrong.

The competition in AI creation is no longer about how impressive the first output is, but how easy it is to revise repeatedly. NanoWork's access to over 100 large models, 500 specialized agents, and 80,000 skill scenarios should be viewed through this lens: the number of models and skills determines coverage, but the efficiency and reliability of the Harness system determines whether these capabilities actually work together. This is a big shift from the past, where companies hired specialized talent and bought software; now, professional capabilities can be organized and called by agents. The barriers that once required whole teams are now being turned into on-demand services, letting small teams and individual creators handle tasks that used to need multiple people.

Connect, Operate, Execute: Letting AI Take Over Workflows

Many AI assistants exist, but the gap is not in more solutions—it's in someone who can actually implement them. An AI might tell you how to schedule a meeting, but you still have to coordinate all the details; it might draft a research plan, but you have to open websites, gather data, and fill in forms. NanoWork addresses this by linking three key areas: tool connection, tool operation, and continuous execution.

First, it connects to popular services like communication apps, calendars, maps, and email, with simplified authorization so regular users can access these external tools easily. But connection is just the start—operation is what matters most. NanoWork can actually use browsers, chat apps, and computer interfaces to do tasks like search the web, fill forms, process messages, and move data between apps. This lets agents handle steps like clicking buttons, inputting text, and reading information, turning multiple apps into a single workflow with more independence and accuracy.

It's not enough for AI to have access to tools; it needs to be able to use them effectively. This is more complex than just chatting with AI—reliable agents need to monitor their environment, check results, retry when needed, and ask for human input when necessary. This requires the same level of engineering expertise as the Harness system. It's easy to make a model click a button once, but hard to keep it on track through dozens of steps, remembering what it did and what comes next.

To solve this, NanoWork adds a cloud office layer for the final stretch: users can assign tasks like data collection, daily reports, or information monitoring to cloud agents, which run even when the local computer is off, letting users check progress or step in when needed. This changes how people work: instead of being tied to their desk, they set goals, define boundaries, and review results.

Security: The Hidden Advantage of Re-entering the Desktop

But with AI handling operations, a new question emerges: how to ensure it doesn't make mistakes? This is where 360's security expertise comes in, a key advantage as it enters the desktop again. When agents have computer access, they could accidentally delete files, send wrong messages, modify important data, or run sensitive actions without proper checks. A large model might make a wrong statement, but an agent could make many wrong actions.

This is where 360's history of security work pays off. Two decades ago, 360 brought free security software to ordinary computers, building expertise in terminal security, system permissions, malware protection, and complex environment engineering. Today, while the risks for agents are new, the core security principles still apply: limiting permissions, isolating environments, tracking actions, and providing recovery for errors.

NanoWork integrates security into every part of the process: managing and checking models, tools, and capabilities at the entry to reduce malicious or untrusted calls; protecting user data with cloud spaces, encryption, and backups; and controlling agent actions with isolated cloud environments, restricted commands, and confirmation for sensitive operations. For businesses, this is more important than adding dozens of models. The stronger the capabilities, the more permissions they need, and the heavier the responsibility for security. This is what sets 360 apart from many new AI companies: its security experience goes back decades, long before the agent boom. For the first time, 360 let people feel safe using computers; now, it wants to make people feel safe letting AI do work on their computers.

Beyond Offices: AI Office Boundaries Are Dissolving

Traditional Office serves knowledge workers, but AI office tools could go far beyond that. NanoWork is already being used in small businesses like cafes, restaurants, and retail stores. A cafe owner has to handle point-of-sale, scheduling, inventory, bills, customer communication, promotions, and social media—all while running the shop. These tasks are small individually, but together they take a lot of time.

NanoWork brings some of these capabilities into one place: using the super canvas to make menus and posters, the video editor to make short videos, connecting to messaging apps for customer communication, and cloud agents to handle periodic tasks like data entry. If it can integrate with point-of-sale, inventory, and membership systems, agents could handle core operations like launching a new product—helping make posters, videos, and ads, and even notifying customers after getting permission.

For small businesses, this is a big deal: instead of paying for a designer or marketer, they can use AI to save money, and a persistent agent could change how the shop runs. This is similar to how companies like Shopify and Square started with small business needs and expanded into management tools. There are millions of small businesses without tech teams, but they have real, frequent needs. If AI can provide creation, customer service, data processing, and analysis at low cost, AI office tools could become common in all kinds of businesses.

Office Entry: A New Era of Competition

Looking back at software history, a pattern repeats: when technology lowers the barrier for ordinary people, the old entry points get redefined, and new industry leaders emerge. NanoWork combines unified editing, multi-model agents, Harness execution, tool connections, cloud work, and security to give 360 a chance to capture the new AI office entry.

Twenty years ago, 360 brought security to Chinese desktops; today, it brings NanoWork to take a spot in the AI productivity era. The shift is clear: the last era was defined by tools like Office, which taught people how to use software. The next era is being defined by AI, which lets software learn to work for people. Microsoft led the last office era; the next one is being written by AI.

以上内容不代表本平台立场,仅供读者参考