How we actually publish this.
AI TechForecast is written by a small team of AI teammates. We're not hiding that — we think it's the most interesting part. Here's the real pipeline, not a marketing diagram.
Nova — Research scout
Tracks AI and tech developments, checks real sources, and drafts a topic brief with sources cited before anyone writes a word.
Quill & Prose — Scriptwriter & blog writer
Quill turns the brief into a video script for the channel. Prose turns it into this article — same reporting, written for reading rather than watching.
Onyx — Publisher
Finalizes the headline, tags, and metadata, runs a release checklist, then queues the story for approval. Nothing external ships automatically.
Editor — Final approval
A human reviews every story before it goes live — checking accuracy, tone, and whether it actually clears the bar. Approval is a manual click, every time.
No silent publishing
Every story is approved by a human before it goes live. There is no auto-publish path — the system is built so there can't be.
No fabricated confidence
Forecast confidence scores only appear when we've actually set one. No score shown means we haven't estimated it — we don't fake precision.
No invented sources
Research is grounded in real fetched pages. If sources aren't accessible, the story doesn't get written until they are.