Future UGC moderation
If a product accepts user reviews, travel tips, community reports, or customer submissions, agents can use VeracityAPI to flag suspiciously generic, AI-planted, competitor-seeded, or weak-provenance submissions.
Business value
- Protects community quality as UGC volume scales.
- Reduces moderator load by prioritizing suspicious submissions.
- Preserves trust in reviews, tips, and public scam reports.
Agent job to be done
Act as a moderation triage layer. Allow low-risk specific submissions, queue suspicious ones, and reject obvious slop/spam according to local moderation rules.
format: product_reviewintended_use: moderatedomain: UGC moderation / travel tips
When to call VeracityAPI
Run on new UGC submission, edited review, bulk import, or report escalation.
What text to submit
Submission title/body, rating if present, category, destination/product, user-provided metadata, and moderation history. Keep user identity metadata outside the text scoring call unless needed.
Decision policy
- allow: low risk and no policy violations.
- allow or revise: medium risk under default moderate policy, but local product policy may queue it.
- human_review: high specificity/slop/provenance risk, especially for accusations, safety claims, or promotional content.
- reject: combine high risk with spam, duplicate, link abuse, or policy violations.
Request template
curl https://api.veracityapi.com/v1/analyze -H "Authorization: Bearer DOC_KEY" -H "Content-Type: application/json" -d '{"type":"text","content":"Paste content here","context":{"format":"article","intended_use":"publish"}}'Automation recipe
- User submits tip/review/report.
- Moderation agent scores text and applies product policy.
- Evidence spans become moderator notes.
- Trusted submissions publish; suspicious ones queue; obvious spam rejects.
- Aggregate risk by user/IP/source to detect campaigns.
Evidence spans agents should inspect
- review could apply to any product/place
- accusation lacks concrete details
- promotional or astroturf language
- generic AI-sounding community tip
Policy pseudocode
if (result.recommended_action === "allow") continueWorkflow(); if (result.recommended_action === "revise") rewriteWith(result.evidence, result.recommended_fixes); if (result.recommended_action === "human_review") queueForHumanReview(result); if (result.recommended_action === "reject") discardOrRebuild();
KPIs to track
- moderator queue reduction
- false positive/negative moderation audit
- spam campaign detection
- published UGC trust quality
- user report rate
What can go wrong
- VeracityAPI is not a complete moderation classifier.
- Do not auto-reject serious victim reports only because they are poorly written.
- Pair with abuse signals, duplicate detection, user reputation, and human escalation.
Cost and latency notes
Analyze only is $0.005 per 1,000 characters; Analyze + revise with auto_revise=true is $0.010 per 1,000 characters. Both round up to the nearest 1,000 characters. Short captions/emails usually cost $0.005; longer pages or chapters scale linearly by length. Current v0.1 latency is LLM-bound, so batch/concurrent orchestration is recommended for high-volume pipelines.
Agent evaluation checklist
- Does this workflow have a costly failure mode from generic or weak-provenance text?
- Can the agent map evidence spans back to editable source locations?
- Should this workflow fail open, fail closed, or queue human review if VeracityAPI is unavailable?
- Which field drives policy: recommended_action, risk_level, content_trust_score, specificity_risk, or provenance_weakness?
- What local rule should complement the API score?