In the evolving landscape of workplace productivity, leveraging AI to build a customized workflow is no longer a luxury but a necessity. This guide, tailored for beginners, will walk you through the practical steps to create your own AI workbench — just like the one that boosted work efficiency by 400%.
My AI Workbench: A Glimpse into Customization
Components of a self-built AI workbench:
- Topic Management: Automatically scrapes data from various platforms to identify trending topics
- Library: A web-based system for managing reading lists and notes, with easy retrieval and reading plan formulation
- Schedule Management: A "Delegate to AI" button — one click assigns tasks to AI for automatic execution, with batch processing for multiple tasks
- Experience Review Library: A self-coded writing application for fine-grained content modification
The Right Mindset: Augment, Not Compete
Forget the anxiety-inducing rhetoric about AI replacing humans. The correct approach is to use AI to augment yourself, not compete with it. Just as cars didn't make walking obsolete but expanded our reach, AI should be your tool to achieve more in less time.
Getting Started with TRAE Work
For beginners hesitant about technical hurdles, TRAE Work is an excellent choice — a leading domestic Agent application that's free to use, accessible on PC, web, and mobile, and integrated with the Feishu ecosystem.
Step 1: Automate Repetitive Tasks
The easiest way to start is identifying tasks you repeat weekly and find tedious. Jot them down — these are your candidates for automation.
Example Automations:
1. Growth Daily Report: Previously, checking data across multiple platforms daily was tedious. Now, delegate to TRAE Work:
"Automatically crawl growth data from [platform names] and generate a daily report."
2. Meeting Summary & Action Items: TRAE Work automatically reviews Feishu meeting notes at night, extracts methodologies, creates a knowledge base, and schedules to-dos:
"Read today's Feishu meeting notes, extract key methodologies, create a knowledge base, and schedule action items."
3. Schedule Automation: Reminders for key focus areas daily, or auto-generated weekly/monthly reports:
"Remind me of key priorities every morning and generate a weekly summary."
Step 2: Create Custom Skills
When tasks have a consistent pattern but aren't identical, turn them into Skills — think of them as your personalized AI apps.
How to Create Skills:
- Find Existing Skills: Browse the TRAE Work community for pre-built skills. If none fit, ask TRAE Work to search GitHub or Reddit for similar functions.
- Customize Skills: Once you find a skill, modify it by describing your needs: "I want this Skill to scrape content from [platforms], filter for posts with at least [X] likes, format the output in [specific style], focus on [specific topics], and send the result to Feishu."
Step 3: Build Your AI Workbench
As you accumulate skills, you'll want a centralized hub — the AI workbench where all automations and skills converge. Tell TRAE Work your requirements for the web page, including features and design. Iterate by communicating about UI adjustments and new features:
"I want to build a web page to manage all my viral content skills. It should display titles, engagement metrics, and allow easy filtering. Let's design the UI and add features like a search bar."
Step 4: Mobile AI Office
To truly free yourself from your desk, leverage TRAE Work's mobile capabilities:
- Sync Across Devices: Bind your PC, web, and mobile accounts for cloud synchronization
- Delegate Tasks on the Go: Dictate an idea into your phone, and TRAE Work files it into your idea library or adds it to your calendar
- Remote Document Handling: Command TRAE Work via mobile to edit a document on your office computer and send it: "Edit [document name] as per [client's requirements] and send it to [email address]."
The Human Touch in an AI-Driven World
While AI excels at execution and structured thinking, it can't replace human emotional experiences and real-life experiences. As you offload tasks to AI, you'll find more time for team collaboration, offline meetups, idea exchanges, and personal relationships. AI handles the "work" and "labor", while you focus on "action" — the unique human elements of creativity, connection, and self-expression.
Start small, automate the repetitive, customize your skills, and build a workbench that grows with you. The future of work is not about competing with AI, but about harnessing it to be more human.
常见问题
What's the difference between TRAE Work and regular Trae (the coding IDE)?
TRAE Work is an Agent application focused on task automation and workflow orchestration — it's designed for non-developers to automate daily work tasks. Regular Trae (trae-framework) is an AI coding IDE similar to Cursor, focused on code generation and software development. TRAE Work integrates with Feishu, runs on mobile, and handles things like meeting summaries and data scraping. Think of Trae as your AI programmer and TRAE Work as your AI personal assistant. They complement each other but serve different audiences.
How much time does it take to set up a useful AI workbench?
Start with one automation (Step 1) — that takes about 30 minutes the first time, 5 minutes once you're familiar. After a week of using that one automation, you'll naturally identify 2-3 more repetitive tasks to automate. A basic workbench with 3-5 automations can be built over a weekend. The full workbench described in this article (topic management, library, schedule, review system) took the author about 2-3 weeks of incremental building. The key is starting small and adding incrementally — don't try to build everything at once.
Can I build an AI workbench without TRAE Work?
Yes. The principles apply to any AI automation tool — Claude Code with custom slash commands, Codex with Agent skills, n8n for visual workflows, or even a combination of Zapier + ChatGPT. TRAE Work is recommended because it's free, has mobile support, and integrates with Feishu (popular in Chinese workplaces). If you're outside the Feishu ecosystem, Claude Code + custom scripts or n8n are excellent alternatives. The methodology (identify repetitive tasks → automate → create skills → build workbench) works regardless of the tool.