Paragon is a small publisher with no face to sell you. These standards are how the work is checked, and how you can check us.
Paragon Editions is a small editorial publisher. We publish two lines of books. The Practice Editions are AI operating playbooks written for one profession each. The Evidence Editions are the Don’t Trust the Hype books and guides, which check popular health claims against the research.
Every title starts from a single profession or a single question. Nothing is written for everyone, and nothing is padded to reach a page count. If a chapter does not help the reader do the work or judge the claim, it does not run.
Each playbook is built from the actual operations of the profession it serves. For a therapy practice, that means intake, scheduling, billing, and follow-up. For a dental practice, front desk, recalls, and reviews. For a fitness coach, onboarding, check-ins, and programming admin.
Every workflow in a playbook is tested in the current versions of the major AI tools before the edition is published. A prompt that only works in one tool, or only on a paid tier, is labeled as such or cut. The language stays plain, because the reader is a practitioner, not a programmer.
Each claim in an evidence check is tested against the primary research where primary research exists: trials, systematic reviews, and the papers the claim itself rests on. Marketing pages and testimonials are never treated as evidence.
Where the evidence is weak, mixed, or missing, the book says so in plain words. No product is endorsed inside an evidence check, and no company pays to be covered or to be spared. Sources are named in the text, so any reader can verify the reading for themselves.
One rule governs every Paragon title. Client and patient information stays out of general AI tools. The playbooks are built to run on templates, scripts, and de-identified material, never on names, records, or case details.
If a workflow cannot be done safely without personal information, it does not go in the book. This boundary is the house rule, and it is printed in the editions themselves.
Paragon editions are dated by year, and the date is part of each title’s identity. AI tools change quickly, and a workflow that was correct last year can be wrong this year.
When a tool or a workflow changes in a way that matters, we revise the current edition and record the revision on the product page. Update notes are written for buyers of record, the readers who already own the edition, so they can see what changed and why.
If you find an error in any Paragon title, write to hello@paragoneditions.com and tell us the title, the page, and what looks wrong. Confirmed errors are corrected in the current edition and logged here.
No corrections have been issued yet.
Paragon publishes under an editorial name, Charles Paragon, so the work stands on its own. There is no founder story to sell and no personality to trust in place of the evidence.
These standards are the accountability instead. The method is public, the editions are dated, the revisions are recorded, and the corrections log sits on this page where anyone can read it.