Agents Lookup

Use cases / Decision models (Jev)

Filter social comments into reply, hide, and escalate

Comments on your business's social posts are sorted by a decision model into needs a reply, spam to hide, or escalate, so the social manager answers real customers first and reviews hides before they happen.

DifficultyBeginner
Setup time2–4 hours (estimate)
CategoryDecision models (Jev)

The use case

Busy posts collect spam, scam links, questions, complaints, and the occasional angry customer, and the questions get lost. Jev reads each new comment and picks what it needs. Questions and complaints with a confidence of 0.85 or higher go to a reply queue. Comments labelled spam are put in a hide list for the social manager to confirm in one batch. Anything mentioning safety, legal threats, or a serious complaint, and anything unsure, is escalated. Jev does not reply, delete, or block anyone. The social manager decides which comments to hide, answers customers, and handles anything escalated, with Claude drafting replies for them to edit. Test the labels on a month of past comments first, and keep the hide list manual for as long as you run it.

Example workflow

  1. New comments on Facebook and Instagram posts are pulled into a Google Sheet by Make.
  2. Each comment and the post it is on are sent to Jev as the state.
  3. Jev answers a Choice question, what does this comment need, with the allowed answers needs a reply, spam or scam, positive no reply needed, and escalate, and a Noul question, the comment mentions safety, injury, or legal action.
  4. Code places comments in reply, hide list, or escalate tabs based on the answers and confidence.
  5. The social manager confirms the hide list in one batch and handles escalations first.
  6. For the reply queue, Claude drafts responses that the social manager edits and posts.

Tools

Risks and guardrails

A similar pattern in a public build: https://madewithjev.com/builds/reply-guy-filter. See also Decision models.

Related recipe

Repurpose One Video or Webinar into Posts, Newsletter, Clips, and a Blog Outline: An agent turns one long video transcript into LinkedIn posts, a newsletter section, timestamped clip suggestions, and a blog outline in your brand voice, with a claims check and human approval before publishing.

Want it set up for you? Request a quote.

Sources

Written by Agents Lookup. Credits and official pages for the tools this use case names: