Prompt refinement with Smart Loops

DeepWriter’s Smart Loops are your most powerful tool for advanced prompting. Don’t think of prompt generation as a one-and-done task. It's a reiterative process designed to help you fully refine your prompt and unlock better results.

Here’s a quick guide to using Smart Loops effectively:

Step #1: start with a broad stroke, then refine.

Begin with a solid—but not necessarily perfect—prompt. Let DeepWriter generate its initial set of follow-up questions. This helps the system build a general understanding of your goal before diving into the specifics.

For example:

"Write a report on the future of electric vehicles."

After entering your prompt, click the Process button:

You’ll then be given a choice between the Original Prompt and the DeepWriter-Processed Prompt. Since we started with a broad subject, go ahead and select the Processed Prompt. It’s optimized based on your intent and Smart Loop input.

Step #2: answer the follow-up questions thoughtfully

Once you’ve submitted your initial input, DeepWriter will generate up to 9 follow-up questions. Be sure to answer them carefully. The quality of your answers directly impacts the quality of the output. Provide as much detail as you can.

While answering more questions will lead to better results, all questions are optional as well. If you feel stuck on a particular question, just skip it. The goal is to let DeepWriter complete the tasks that you don't want to, but allow you full control where you have a clear plan.

These questions are DeepWriter’s way of understanding your specific needs.

Step #3: the "Re-process" loop: where the magic happens

After answering the follow-up questions, do not proceed with the generation.

Instead, click Re-process.

DeepWriter will absorb your refinements, re-evaluate its plan, and go deeper. Think of this as a collaborative research session—you're guiding its focus with each iteration.

You can repeat this process multiple times.

Pro tip: If the initial output isn’t quite right, don’t start over. Instead, analyze why it missed the mark. Then adjust your original prompt, answer the Smart Loop questions with those insights in mind, and click Re-process again. You can repeat this loop multiple times to fine-tune the output until it converges on your ideal result. Review the Prompt elements sections for more prompt refinment ideas.

Example Scenario:

  • First attempt: Your first draft on electric vehicles might feel too general. Upon review, you realize it lacks specific examples of new battery technologies.

  • Your refinement: Go back to the prompt and add a note like:

“Be sure to include detailed examples of solid-state batteries and next-generation charging solutions.”

  • Next step: re-process the prompt and answer any new follow-up questions to guide the revision further.

Last updated