Editorial standard v1.0

Automation with a brake pedal.

Choice Ledger uses AI to accelerate research and drafting. It does not use AI as permission to invent experience, sources, prices, ratings, or certainty.

Our labels

Know what kind of page you are reading.

Verified spec

A factual attribute checked against a named source.

Research-based

A comparison built from external evidence without first-hand use.

Hands-on tested

Reserved for products with a recorded test protocol and original evidence.

Publication gates

Every draft has to earn its URL.

01

Source coverage

Material product claims need a reachable source and a retrieval date.

02

Category fit

Every comparison uses decision factors designed for that product category.

03

Claim language

Manufacturer claims stay attributed; research is never relabeled as hands-on experience.

04

Distinct value

A page must add a useful comparison, calculation, framework, or clarification.

05

Freshness

Changing facts receive expiration dates and are queued for rechecking.

06

Editorial approval

Automation can draft, but it cannot move its own work directly into the public index.

Example: mower runtime

Why we do not rank a claim as if it were a lab result.

Battery runtime changes with grass density, height, moisture, slope, cutting mode, drive use, blade selection, temperature, and operator technique. Until the same test is run across every mower, manufacturer runtime belongs in an attributed comparison column—not a performance score.

AI policy

What the system may and may not do.

Allowed
  • Find and organize candidate sources
  • Extract structured specifications
  • Draft comparisons and refresh notes
  • Flag contradictions and stale claims
Blocked
  • Claim personal product use
  • Invent citations, prices, or ratings
  • Publish its own unapproved draft
  • Hide sponsorship or affiliate incentives