How we make Unjargoned
Every story is drafted with AI, then checked, edited and signed off by a named editor. Nothing publishes without that sign-off. Here is exactly how it works, and what we do when we get something wrong.
How a story is made
- We watch the sources. Our pipeline reads AI labs, regulators, research, product launches and UK news every 15 minutes, and keeps every item’s original link.
- We pick what matters to you. Reports about the same event are grouped, then ranked for what matters to people who use AI, rather than people who build it, and for UK relevance. Editors start each day with that shortlist.
- AI drafts each story three times. Once for Beginners, once for Intermediates and once for Advanced readers, each with a two-minute “try it” step and the jargon explained.
- A second AI pass checks every claim. It links each claim to a source and flags anything thin or unsupported. A story with an open flag cannot be approved.
- A named editor signs it off. They read all three versions, resolve every flag and approve the story. Their name is on it, and so are the sources.
What AI does, and what people do
AI
- Gathers and groups the news
- Drafts each story at three levels
- Checks claims against sources and flags gaps
Editors
- Choose the stories
- Rewrite, cut and correct the drafts
- Resolve every flag and sign every story off
- Write and approve every tool verdict
We use Anthropic’s Claude models for drafting and checking. No reader’s personal data is ever sent to them. We tell you a story was AI-assisted with the badge under every headline.
Our three levels
The same story is written at three depths. Beginner assumes nothing and explains every term. Intermediate is for people who already use AI for writing and want it to do more. Advanced is for people who build with it, with the technical detail left in. The facts are the same at every level; only the depth changes. You can switch at any time.
Corrections
When we get something wrong, we fix it quickly and say so.
- Tell us at corrections@unjargoned.com, or reply to any briefing. We aim to review every report within one working day.
- We correct all three levels of the story, and add a dated note at the foot of it saying what changed.
- If the mistake was in an email, we correct it in the next briefing too.
- We never quietly change a fact. Typos and broken links we simply fix.
Corrections log
No corrections yet.
How we review tools
- An editor uses every tool in our portal hands-on before we give a verdict, and scores it out of 5.
- Each tool’s page says what it does, how it works, how hard it is to use, what it costs and what happens to your data. An editor checks the page when they sign off the verdict, and we say when we last checked the prices.
- Until an editor signs a verdict off, the tool shows as “being tested”.
- Some tools are partners: if you sign up through our link, the maker pays us a commission, and some give our readers a deal. A partnership never changes a verdict, a score or where a tool ranks, and we label every partner link and offer.
Advertising
Advertising pays for the free briefing. These rules keep it honest:
- Every paid placement is labelled: “Sponsored”, “Together with”, “Featured” or “Category sponsor”. You will always know.
- Ads never sit inside a story, and sponsors never see a story before it is published.
- Sponsors cannot buy, change or see our verdicts on tools, and cannot buy coverage.
- We do not take ads for gambling, payday loans, crypto trading or get-rich-quick schemes, or political ads.
- Offers from AI companies are paid placements. If we earn a commission from a link, we say so next to it.
Advertisers can find our rates on the advertise page.
Contact the editors
Story ideas, questions about how we work, or a source you think we missed: hello@unjargoned.com.