If you’ve been using ChatGPT, Claude, or Gemini for a while, you’ve probably noticed something frustrating: sometimes you spend more time crafting the perfect prompt than you would have just doing the task yourself.
The problem isn’t the AI. It’s that most of us never develop habits around prompting—we treat every conversation like a blank slate. The people who get the most out of AI aren’t necessarily better prompters. They just have systems that make prompting faster and more repeatable.
Here are five habits that will change how quickly you work with AI tools.
1. Keep a Swipe File of Prompts That Already Worked
Stop reinventing the wheel every time you need a summary, email draft, or brainstorm. When a prompt works well, save it somewhere you can copy and paste later—a note, a doc, or even a browser bookmark.
For example, if you’ve crafted a great prompt for summarizing meeting notes, save it as a template: “Summarize the following meeting notes into three sections: Key Decisions, Action Items, and Open Questions. Use bullet points and keep it under 200 words.”
Next time, you just paste your notes and hit send. No thinking required.
2. Start Broad, Then Narrow—Not the Other Way Around
Most people overcomplicate their first prompt. They pack in every detail, constraint, and edge case upfront, then get frustrated when the AI misses the mark.
Instead, start with a simple, clear request. Let the AI give you a first draft. Then refine it with follow-ups like “Make it shorter,” “Add more examples,” or “Rewrite this in a friendlier tone.”
This two-step approach is almost always faster than trying to nail everything in one mega-prompt. Plus, you’ll often discover what you actually want once you see a rough version.
3. Use “Continue” Instead of Re-Explaining
When an AI response gets cut off mid-sentence or stops before finishing a list, don’t rewrite your original prompt. Just type “Continue” or “Keep going.”
The model will pick up exactly where it left off. This works in ChatGPT, Claude, and Gemini, and it’s one of the simplest time-savers that almost no one uses consistently.
If you want to extend an idea without repetition, try: “Expand on the last point with more detail.” The AI remembers the context—you don’t need to remind it.
4. Name Your Outputs to Make Them Reusable
If you’re working on something complex—like a blog post, a report, or a product brief—give each section or version a name in your conversation. For example: “Let’s call this draft version A. Now write version B with a more casual tone.”
Later in the conversation, you can say, “Combine the intro from version A with the body from version B,” and the AI will know exactly what you mean.
This is especially powerful in tools like Claude, where you can reference earlier parts of long conversations without scrolling or copying text. Naming things creates mental shortcuts for both you and the model.
5. End Every Session by Asking “What Could I Prompt Better?”
Here’s a habit that compounds over time: after finishing a task with AI, ask it to critique your prompting.
Try: “How could I have prompted you more effectively to get this result faster?”
The AI will often point out unnecessary details, unclear instructions, or better ways to structure your request. You’ll learn what actually matters in your prompts and what’s just noise. Do this a few times, and you’ll start writing better prompts instinctively.
Speed Comes from Systems, Not Skill
You don’t need to be a prompt engineer to use AI effectively. You just need a few simple habits that eliminate friction and repetition.
Save what works. Start simple and iterate. Use shorthand like “continue.” Name your outputs. And ask the AI to teach you how to work with it better.
These five habits take almost no extra effort, but they’ll cut your prompting time in half—and make working with AI feel less like a chore and more like a superpower.
Want one useful AI tip like this in your inbox every day? Subscribe to the One Two Three AI newsletter and stay ahead without the overwhelm.