# Decide when an LLM is needed in a workflow

Judges: one workflow step's input. Questions: noul:needs\_writing; noul:needs\_open\_reasoning. Action: only true cases call an LLM.

Grade: A — A clean typed decision \(yes/no, a pick from a list, or a level on a scale\) that is concrete, repeats routinely, and has meaningful value.. Score: 3.89.

Vertical: Cross-industry.

Function: Agents & dev.

Status: Idea.

Volume: routine.

Value: meaningful.

Risk: moderate.

Evidence: null.

Flags: None.

- [Vertical](/verticals/cross_industry/)
- [Function](/functions/agents_dev/)
- [Security alert categorization system](/use-cases/security-alert-categorization-system/)
- [Cross-domain: sensitive-data leak guard on LLM prompts](/use-cases/cross-domain-sensitive-data-leak-guard-on-llm-prompts/)
- [Phishing triage](/use-cases/phishing-triage/)
- [Keystroke-level sensitive-text guard](/use-cases/keystroke-level-sensitive-text-guard/)
- [Confidential message detector](/use-cases/confidential-message-detector/)
- [PII in free text](/use-cases/pii-in-free-text/)