1. The honesty angle — “AI resume builder that refuses to lie”
The central premise: every AI resume tool that markets rewriting capability has a hallucination problem. Tailoredresponds with five server-side deterministic safety guards (stale-path, no-op, fabricated-date, fabricated-number, fabricated- proper-noun) that filter every proposal before display. Lead with the “30% revenue” example: a proposal containing an invented metric fails the numeric-token check and is silently dropped. Receipts at /guides/how-smart-tailoring-works.