How to Use WorkBuddy: A Beginner's Path from Your First Task to Knowledge Bases, Skills, Connectors and Automation

WorkBuddy is an AI workspace built around tasks. This guide goes in the order a beginner needs: creating your first task and checking the result, what the knowledge base, Skills and connectors are each for, when to add automation and messaging-app assistants, and what to watch with permissions and cost.

how to use WorkBuddyWorkBuddy tutorialWorkBuddy beginner guideWorkBuddy skills and connectorsAI task workspace automation

Plenty of people install WorkBuddy, see a long row of entry points (tasks, assistants, projects, experts, skills, inspiration, connectors), have no idea where to start, and end up using it as a chat box. You only need one word to make sense of it: task. WorkBuddy is an AI task workspace. You hand it a specific job, it works on that job inside a task, talking with you, calling tools and producing files, and you review the output. Every other feature exists to make tasks more accurate and less work.

Step 1: Finish one small task first

Once it's installed (there are installers for Windows and macOS), don't try to learn every menu. In the new task bar, describe something small and concrete that you can check right away, such as “turn the PDFs in this folder into a list of key points” or “make a trend chart from this spreadsheet.” While the task runs, you can add requirements in the task conversation, and when it's done you check what it produced in the results view. Going from creating a task to talking to it to reviewing the result once is the core lifecycle of WorkBuddy.

Step 2: Manage tasks instead of running one endless chat

A common beginner habit is asking everything in one conversation until it becomes a mess. A better habit is one job, one task: one for the weekly report, one for the data analysis, one for the content schedule. Related tasks can live together in a project workspace, and you can find old ones in task management. Each task keeps a clean context, and the results are easier to reuse.

Step 3: Use the knowledge base so it works from your own material

The AI doesn't know your company's products, your notes or last week's meeting minutes, so its answers tend to stay generic. The knowledge base fixes that: WorkBuddy can connect to Tencent Docs, the ima knowledge base and the Lexiang knowledge base, so tasks can draw directly on those documents. Start with just the one source you use most, and put in documents you'll actually reference again and again. That works better than importing everything at once.

Step 4: Skills handle “how,” connectors handle “where”

Skills are packaged ways of doing a job, such as browser automation, office document handling, speech transcription, web search or working with Obsidian notes. You can find them in the skill marketplace or create your own. Connectors link WorkBuddy to outside services so tasks can read and write data there. A simple rule: if you do the same kind of job repeatedly and keep explaining the steps, turn it into a Skill; if you need to touch data in another system, look for a connector. Before installing a third-party Skill, check what it actually does, and use a tool like Skill Scanner if you're unsure.

Step 5: Add automation and assistants only once things are stable

When you've run a task by hand several times and the results are consistent, consider automating it, for example sending yourself a news briefing every morning. Another common setup is connecting a WorkBuddy assistant to a chat app you already use. The course covers WeChat, WeChat service accounts, WeCom, QQ, Yuanbao, Feishu and DingTalk, so you can send one message from your phone and have it handle the job. Order matters: get it working by hand, then automate, then open it up in chat apps. Otherwise, when something breaks, it's hard to tell where.

Check permissions, memory and cost before you start

In the system settings, WorkBuddy has two permission modes. When you're starting out, use the more cautious one so it asks for your confirmation before important actions. Memory affects what it carries into later tasks, so it's worth reviewing now and then. WorkBuddy also bills by credits and usage, and automated tasks keep consuming them while you're not watching. Look at the usage and billing pages before going live, then decide how often things should run.

Three mistakes beginners make most often

First, vague task descriptions: “do an analysis for me,” with no data and no output format. Second, building automation right away without ever running the task by hand. Third, dumping every document into the knowledge base, so it ends up citing outdated or irrelevant material. Write specific tasks, go manual before automatic, and keep your sources few and relevant, and WorkBuddy becomes far more useful.

If you'd like to learn it systematically, from the task lifecycle, installation and the main interface to the knowledge base, automation, practice cases, free Skills and assistant integrations on seven platforms, take a look at WorkBuddy: Beginner to Expert (lessons are in Chinese). It has 69 lessons, and the first 8 are free to watch.