Learn how to bring AI into work you already know, decide where it actually helps, and build one workflow you can use again.
Start with a task you understand, then decide where AI belongs and where your judgment still matters.
Start with recurring work you understand, such as a follow-up, report, or handoff.
Choose what AI should handle, what information it needs, and what you will still review.
Try realistic cases, refine the instructions, and save a process you can use again.
What you get
Learn the method by applying it to one familiar task: shape the instructions, test realistic cases, and save the finished workflow to files you control.
Inside the workbook
Follow a worked example to see how a familiar task becomes a clear, reviewable workflow before you build your own.
What you keep
The complete offer
Work through Choose, Build, and Test at your own pace. Keep the finished workflow, and return to the same method when another suitable task comes up. No account, subscription, or installation.
Public checkout remains closed until every activation requirement is verified.
Already have access? Use your recovery code.
External research, bounded context
Published research has found measurable gains when generative AI was applied to specific work in controlled settings. The results make a carefully chosen task worth testing, not assuming.
15%more customer-support issues resolved per hour, on averagePublished study result
A field study followed 5,172 customer-support agents at one company using a generative AI assistant. Agents remained responsible for the conversation and could ignore or edit suggestions.
Context limit: Effects varied substantially. Less-experienced workers gained more, while the most experienced workers saw small gains in speed and small declines in quality.
Generative AI at Work, Erik Brynjolfsson, Danielle Li, Lindsey Raymond, The Quarterly Journal of Economics (2025).
Read the study in The Quarterly Journal of Economics40%less time on the assigned professional writing tasksPublished study result
A randomized experiment with 453 college-educated professionals found that average completion time fell by 40% and independently assessed output quality rose by 18% on short, occupation-specific writing tasks.
Context limit: The tasks were short and self-contained, without the company context or factual verification that much real business work requires.
Experimental evidence on the productivity effects of generative artificial intelligence, Shakked Noy, Whitney Zhang, Science (2023).
Read the study in ScienceThe practical question is not whether AI is universally productive. It is whether one bounded, recurring task has known inputs, a human reviewer, and a manual fallback, then earns an Adopt, Revise, or Stop decision through a small test.
A practical place to begin.