AI News & Strategy Daily | Nate B Jones
How To Use ChatGPT Work: The Complete Beginner's Guide (2026)
Summary
This guide provides a comprehensive walkthrough of advanced AI workflows using ChatGPT Work and OpenAI Codex, moving beyond simple chat interactions. The focus is on automating complex, multi-step tasks—such as editing documents surgically, extracting structured data from images/spreadsheets, and running scheduled jobs—without requiring coding. Key concepts include selecting the correct execution environment (Cloud vs. Local), defining clear goals, and utilizing connected apps and event triggers to integrate AI into existing enterprise workflows.
Key takeaways
-
Choosing the Right Execution Environment
5:21
Determine if the job requires a cloud environment (for running independently of a closed laptop) or a local environment (if it needs access to local files or applications). Always choose the environment *before* attaching material.
-
Surgical Document Editing
22:09
Instead of using generic prompts like 'improve the doc,' provide precise instructions specifying what parts must change and what parts must remain untouched (e.g., 'Keep the headings, preserve the links, change the explanation in this section').
-
Structured Data Extraction
35:01
When processing photos or messy spreadsheets, define the exact fields needed (e.g., 'merchant column, a date column, a total, a currency, a category') and specify edge cases (e.g., 'if the source doesn't say, leave it blank and mark it for review').
-
Automating Recurring Tasks
53:30
Use scheduled tasks for repeatable jobs (e.g., weekly briefings). Be precise about the cadence and time zone, and always test the first few runs to ensure the instructions are clear enough for reliable repetition.
-
Controlling Complex Workflows
1:00:30
Use the `/goal` command for multi-round assignments that require sustained work, allowing the user to pause, inspect, and correct the task's progress iteratively. For planning, use `/plan` to identify necessary decisions before execution.
Technical details
-
Cloud vs. Local Execution
321s
Cloud tasks can continue running when the laptop is closed, using cloud-available sources and tools. Local tasks require the machine to remain on and available, and local schedules are possible, but the cloud task does not inherit local files.
-
AI Model Selection and Reasoning
3850s
The model (e.g., Astra, Soul, Terra, Luna) determines the system doing the work, while the reasoning effort determines the time/complexity budget. Astra is noted as the strongest option for demanding work requiring careful analysis.
-
Connected Apps and Integrations
3050s
AI can write to external services (e.g., Google Docs, Slack) via connected apps/plugins. When integrating, always ask the AI to check for existing records and read the destination afterward to catch duplicates or unexpected changes.
-
Event-Triggered Tasks
3450s
Tasks can be triggered by supported events (e.g., new emails in Gmail, messages in Slack) rather than just time. It is crucial to define a clear start signal and an end condition to prevent infinite or wasteful processing.
-
Project Management and Context
2800s
Use 'Projects' to group related tasks, sources, and instructions, preventing the need to re-supply background context in every prompt. Distinguish between attaching a file (for one task) and using the 'Library' (for reusable material).
Mentioned resources
- ChatGPT Work
- OpenAI Codex
- Astra
- Google Docs
- Slack
Channel & topics
Watch on YouTube · Back to latest
This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.