Get AI summaries of any video or article — Sign up free
Organize your notes using AI (productivity hack) thumbnail

Organize your notes using AI (productivity hack)

Reflect Notes·
5 min read

Based on Reflect Notes's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.

TL;DR

Use Reflect’s AI palette editor (Command J) to run prompts on highlighted note text and replace messy content with structured outputs.

Briefing

AI can turn messy, unstructured note-taking into clean, searchable notes by automatically extracting key ideas, organizing meeting content, improving writing quality, and adding structure like backlinks and tags. The biggest payoff is speed: instead of manually rewriting rambling thoughts, formatting meeting minutes, or remembering what to tag, a set of AI prompts can reshape raw notes into usable outlines.

The workflow starts with “structuring thoughts and rambles.” For voice notes, a prompt in Reflect’s AI palette editor (opened with Command J) can convert a free-form idea into a set of bullet-point key takeaways. The transcript describes using a system prompt that already exists in Reflect, then optionally swapping in a custom prompt that produces more detailed bullets. The result is a concise version of a concept that can be referenced later or turned into a longer piece.

Next comes meeting-note processing, using a custom “executive assistant” prompt. The approach combines quick, in-the-moment bullet notes with a short post-call audio memo (about 30 seconds to a minute) describing what happened and what needs attention. After highlighting the combined text and running the prompt, the AI outputs a structured markdown outline with distinct sections: meeting takeaways (at least three items), action items the note-taker must do, action items for other people, and next steps. For team leads or anyone with frequent meetings, the practical value is turning scattered notes into an organized checklist.

Writing cleanup is handled with a copy-editor prompt. When notes are typed on mobile—where abbreviations, incomplete words, and grammar mistakes are common—the AI can rewrite the text with correct punctuation, capitalization, and overall readability. The transcript emphasizes that this particular cleanup prompt is also a system prompt, so it doesn’t require cloning or customization.

The most “networked” part of the system is automatic backlinks and tagging, aligned with network note-taking. In Reflect, backlinks are created by highlighting entities and using double brackets. A custom prompt can add those double brackets around capitalized entities like places and people (e.g., “Boulder” and “Barbara”), though the transcript notes a current limitation: backlinks may still require a click to fully convert into linked nodes. Tagging works similarly: a custom prompt suggests a few tags based on the highlighted text, and the best practice is to feed the prompt a list of existing tags so it doesn’t invent new ones.

Finally, AI can fill in missing information. A custom prompt can read a note that contains an open question—like estimating drive time from Estes Park, Colorado to Denver International Airport—and then insert the missing detail (about an hour and 40 minutes with traffic, according to the transcript). Across all examples, the common thread is prompt-driven formatting: define a target structure, run it on highlighted text, and transform raw notes into organized, actionable records.

Cornell Notes

AI prompts in Reflect can convert raw notes—voice memos, messy text, and meeting scribbles—into structured, searchable content. Key takeaways can be extracted from rambles into clean bullet points, and meeting notes can be reorganized into markdown sections for takeaways, personal action items, others’ action items, and next steps. A copy-editor prompt cleans grammar, punctuation, and capitalization for faster mobile note-taking. Custom prompts can also add network-style backlinks (double brackets around entities) and suggest tags from an existing tag list. Another prompt can answer questions embedded in notes by finding missing information and inserting it directly.

How does the workflow turn a voice note or ramble into something easier to reference later?

It uses Reflect’s AI palette editor (Command J) with a prompt that extracts “key takeaways” and outputs them as bullet points. The transcript describes running a system prompt that already exists in Reflect, then optionally using a custom prompt that produces more detailed bullet points. The key move is highlighting the ramble/voice-derived text and replacing it with the AI-generated bullet list.

What’s the structure produced by the “executive assistant” meeting-note prompt, and what inputs does it use?

The prompt creates a markdown outline with four categories: (1) meeting takeaways (at least three items), (2) action items the note-taker must do, (3) other people’s action items, and (4) next steps. Inputs include quick bullet points captured during the call plus a short audio memo after the call (about 30 seconds to a minute) describing what happened and what needs attention.

Why is the copy-editor prompt useful for mobile note-taking?

Mobile typing often leads to abbreviations, incomplete words, and grammar mistakes. The copy-editor prompt rewrites the highlighted text with correct punctuation and capitalization and fills in missing words. The transcript notes this is a system prompt, so it can be run without cloning or customizing.

How do backlinks and tags fit into network note-taking in Reflect?

Backlinks are created using double brackets around entities (places, people, projects—anything that starts with a capital letter). A custom prompt can automatically insert those double brackets around entities in selected text (e.g., “Boulder” and “Barbara”). For tags, another custom prompt suggests a few tags based on the highlighted content; the transcript recommends providing the prompt with a list of existing tags so it selects from known categories rather than inventing new ones.

What does “filling missing information” do, and what example is given?

It reads a note to find gaps or questions and then answers them by inserting the missing detail. The example note asks how long it takes to drive from Estes Park, Colorado to Denver International Airport; the AI inserts an estimate of about an hour and 40 minutes with traffic, described as accurate based on a lookup done that day.

Review Questions

  1. When converting a ramble into key takeaways, what output format is used and where is the prompt run?
  2. What categories appear in the executive-assistant meeting outline, and which two note inputs are combined before running it?
  3. What’s the recommended approach to tagging so the system doesn’t invent new tag names?

Key Points

  1. 1

    Use Reflect’s AI palette editor (Command J) to run prompts on highlighted note text and replace messy content with structured outputs.

  2. 2

    Extract voice-note or ramble key ideas into bullet points by running a “key takeaways” prompt, optionally using a custom version for more detail.

  3. 3

    For meetings, combine in-call bullet notes with a short post-call audio memo, then run the executive-assistant prompt to generate a markdown outline with takeaways, action items, others’ action items, and next steps.

  4. 4

    Run a copy-editor prompt on mobile-typed notes to fix punctuation, capitalization, and incomplete words quickly.

  5. 5

    Add network-style structure by using custom prompts that insert double-bracket backlinks around entities, aligning notes through shared people and places.

  6. 6

    Suggest tags from existing tag lists to keep organization consistent and avoid creating new tag categories.

  7. 7

    Use a “missing information” prompt to answer embedded questions and insert the needed facts directly into the note.

Highlights

A single “key takeaways” prompt can turn a free-form voice ramble into clean bullet points that are immediately referenceable.
The executive-assistant meeting prompt outputs a four-part markdown outline: takeaways, personal action items, others’ action items, and next steps.
Backlinks can be automated by inserting double brackets around capitalized entities like places and people, supporting network note-taking.
Tagging is most reliable when the prompt selects from a predefined set of existing tags rather than inventing new ones.
Missing details can be filled in by prompting the AI to answer questions embedded in notes, then inserting the result (e.g., drive time from Estes Park to Denver International Airport).

Topics

  • AI Note Structuring
  • Meeting Notes Automation
  • Copy Editing
  • Backlinks and Tags
  • Filling Missing Info