Automation recipe · feedback
Best for
Creators who want to notice actionable production, clarity, pacing, accessibility, or teaching feedback without automatically engaging with abuse, threats, or sensitive disputes.
Recommended action
AI-generated reply
Start in approval mode so you can inspect real matches before anything posts.
The rule boundary
Match one job, leave nearby comments out
AI intent matching is useful because constructive feedback rarely shares one exact keyword. Keep the rule in approval mode and exclude harassment, threats, protected-class attacks, and personal disputes.
| Match type | AI intent matching |
|---|---|
| Starter condition | the commenter is offering specific, good-faith feedback about the video, explanation, production, pacing, accessibility, or teaching approach |
| Action mode | AI-generated reply · Needs approval |
Make it yours
Edit the facts, then preview the reply
The fields below are placeholders for your channel. The preview updates as you type; nothing is posted or saved by this page.
“The captions were hard to follow in the second half, but the explanation was helpful.”
Thanks for pointing out [the specific feedback]. I’ll keep that in mind as I improve future videos.
Add your own values for Feedback theme to enable the draft handoff.
The draft opens in CommentShark's Comment Assistant. Review the rule, accept the product disclaimer, and save it under your own account when it fits your channel.
The draft uses this instruction to guide a fresh reply; the public example above is a starting point you can copy or adapt.
Use real comments to calibrate
What should match and what should not
Examples are guardrails, not a promise that every comment will be classified correctly. Review your own channel language before enabling a rule.
Good matches
The explanation was clear, but the screen recording moved too quickly to follow.
The viewer offers specific, actionable feedback in good faith.Could you add captions to the code demo next time? It would make it easier to pause and read.
This is a concrete accessibility or clarity suggestion.The recipe was useful, though the ingredient amounts appeared too late.
The feedback identifies a specific improvement without attacking the creator.
Keep out
You are an idiot and nobody should watch this.
Insults are not constructive feedback and should not receive this response.I disagree with your political position.
A values dispute needs human judgment and context.Fix this or I will find you.
Threats must not be handled by a generic reply rule.
Why this structure works
A useful answer needs a useful boundary
- It acknowledges the useful part of a comment without automatically agreeing with every conclusion.
- It avoids making promises about future videos that the creator has not approved.
- Approval mode protects the channel from misreading sarcasm, harassment, sensitive topics, or personal disputes.
Setup path
Configure it in three deliberate passes
The recipe page is a planning and preview tool. CommentShark does not post or save anything until you open the editor, review the draft, and save it under your own account.
- 1
Define constructive narrowly
Describe good-faith, specific feedback and list the classes of comments that must stay out of this rule. A broad “negative comments” instruction is too risky.
- 2
Review every suggestion
Keep this rule in approval mode. Edit drafts to reflect the actual feedback and reject replies where acknowledgment would escalate the situation.
- 3
Use the feedback
Record recurring themes in your production notes. The rule is for acknowledging a viewer, not for deciding which criticism your channel should adopt.
Before you enable it
Quick safety checks
- Never use this rule for threats, harassment, hate, doxxing, or personal disputes.
- Do not promise a feature, policy, upload, or correction unless you intend to keep that promise.
- Keep a human review step for sarcasm, controversial subjects, and comments that mention personal circumstances.
Questions creators ask
Frequently asked questions
Should constructive-criticism replies ever be autonomous?
Usually keep them in approval mode. Feedback is nuanced, and the cost of a defensive or tone-deaf public reply is higher than the time saved by removing review.
How is criticism different from abuse?
Constructive criticism describes a specific experience or improvement in good faith. Abuse, threats, identity attacks, and personal disputes need a different moderation or human-review path.
Can the reply promise to fix the issue?
Only when you have chosen to make that promise. The starter instruction avoids commitments so you can decide after reviewing the actual feedback.