Skip to content

Evidence-aware article drafting

Symthink can help you turn a Question branch into an externally drafted article without calling AI inside the app. You copy prompts into your preferred AI tool: first an editorial analysis, then an evidence-aware article draft in the same conversation.

This workflow is separate from AI research, which is designed to add new material back into your tree.


Who can use it

Role Where What you get
Author (tree owner) My VisionOptions on a Question Full Analyze reasoning / Create article dialog: analysis + draft prompts
Author viewing own Shared Vision Shared Vision speed-dial on a Question Same full dialog via Analyze reasoning
Someone following the Question Shared Vision speed-dial on the exact Question you follow Analyze reasoning — analysis prompt only
Public viewer / non-follower Shared Vision No article action

Only Question nodes offer this action. Idea and Claim Options menus do not include it.

Following a parent branch does not qualify. You must follow the specific Question you are viewing.


Author workflow (two steps)

Step 1 — Analyze My Reasoning

  1. Open Options on a Question in My Vision (or use Analyze reasoning on your own Shared Vision Question).
  2. Choose Analyze reasoning / Create article.
  3. In the dialog, copy Analyze My Reasoning (Prompt 1) into an external AI chat.

The prompt includes a YAML snapshot of your Question branch. The external AI should identify strengths, editorial gaps, and a heuristic article readiness score.

Important: Readiness percentages are external heuristic judgments for your reflection. Symthink does not calculate, store, or validate them.

Typical readiness bands shown in the dialog:

  • 80–100% — Ready to draft
  • 60–79% — Draft with caution
  • Below 60% — Strengthen the reasoning first

Step 2 — Draft an Evidence-Aware Article

After the analysis completes in the same external conversation:

  1. Return to the Symthink dialog.
  2. Copy Draft an Evidence-Aware Article (Prompt 2).
  3. Paste it into the same AI chat where Prompt 1 ran.

Prompt 2 does not include your branch again. It assumes the AI already has both the branch YAML and its analysis from Step 1. The dialog explains this under Continue in the same conversation.

The external AI should draft an article that does not strengthen claims beyond what your tree supports, preserve Question / Idea / Claim distinctions, insert visible [EDITORIAL GAP: …] annotations when reasoning is insufficient, and list References only from sources actually used.

Symthink does not display or save the article. Copy the result from your AI tool.


Follower workflow (analysis only)

If you follow someone else's public Question:

  1. Open their Shared Vision and navigate to that Question.
  2. Use the speed-dial action Analyze reasoning.
  3. Copy the analysis prompt into your external AI tool.

Use the analysis to decide where you can contribute evidence, ideas, or counterarguments to help the author strengthen their reasoning. Symthink does not send your analysis to the author.

People who follow your Question do not receive the article draft prompt. Drafting remains with the reasoning author.


Branch size limits

Symthink refuses to prepare prompts when a Question branch is too large:

  • Too many nested levels — choose a narrower Question with fewer sub-levels.
  • Prompt too long — choose a narrower Question or reduce nested content.

You will see an actionable message instead of a silently truncated prompt.


How this differs from AI research

AI research Evidence-aware article drafting
Goal Find and add new research to your tree Audit existing reasoning and draft an article externally
Symthink receives the AI response? Yes (paste-back) No
Which nodes? Main card or selected point Question only
Prompt content Card-focused research context Full Question branch as YAML

Use AI research when you want to expand your tree. Use article drafting when you want to evaluate and render what you already have.


Privacy notes (Shared Vision)

  • The action uses the published snapshot visible on Shared Vision.
  • People following the Question see analysis framing aimed at collaborative strengthening, not author-only drafting instructions.
  • Public viewers who do not follow the exact Question do not see the action.