AI Disclosure

• Source intake. Pulls RSS, Atom, and vendor advisory feeds from a curated source list. Each source has a trust score set by Notifire editors.
• Clustering and scoring. Groups stories about the same underlying event and scores them for impact and trend velocity.
• Drafting. Generates the briefing — title, summary, body, why-it-matters, business impact, action checklist — from the primary source text. Current model: Google Gemini 2.5 Flash.
• Imagery. Generates or selects a featured image and alt text. Generated images are explicitly labelled in metadata.
• Quality check. An automated gate verifies structural compliance (length ranges, paragraph counts, SEO title and description bounds, image alt text presence, primary source reachability) before the briefing is allowed to publish.

• Invent facts, sources, statistics, quotes, or links.
• Publish briefings sourced from social media or anonymous posts.
• Make editorial decisions about cybersecurity stories — those are routed to a human editor by policy.
• Set Notifire's editorial standards, source trust scores, or category rules — those are written and maintained by the editorial team.

Briefings that pass the automated gate auto-publish on a 15-minute schedule. Briefings that fail any hard gate are routed to a human editor who accepts, edits, or rejects them. Certain categories (currently cybersecurity advisories) require human review by policy regardless of automated score.

• Drafting: Google Gemini 2.5 Flash
• Imagery: Generative image models, with curated stock fallback per category
• Quality scoring: Rule-based, no LLM in the gate

Notifire is accountable for every published briefing, including those generated by AI. If you spot a factual error, please report it via our corrections policy and we will fix it.