About CoverWire
I started CoverWire because I wanted a better way to keep up with insurance news. I used AI to gather and write up the information I was interested in, then decided to make it public.
The website and stories are created using AI. I'm developing the process as I go, and I want to be open about how it works and where it needs improving.
I'm also building CoverWire's insurance data service. The news site is a way to share useful information while that develops.
If a story gets something wrong, or misses something that matters, please tell me.
David Tucker
Founder, CoverWire
CoverWire is an insurance news publication created using AI. It is a public experiment, founded by David Tucker.
Who publishes CoverWire
CoverWire is published by Dt5 Consulting Ltd, a company registered in England and Wales, number 11293382. Registered office: Crown House, 27 Old Gloucester Street, London WC1N 3AX.
Why CoverWire exists
Insurance news is spread across company announcements, regulators' publications, filings and trade titles. CoverWire tests whether AI can turn those sources into short, accurate, clearly sourced reports every weekday, and it does so in public so readers can judge the results.
What AI does and what people do
AI systems choose stories from the day's announcements, gather the sources, write the articles, check them against those sources, pick an image and publish them. There are no reporters or editors on staff.
David Tucker sets the rules the system works to: what to cover, the house style, the sourcing standard and what must be held back. He reviews how it is working and changes the rules when it goes wrong. He does not read every story before it is published. Longer analysis pieces and the newsletter are published only with his approval.
Some stories are held back automatically rather than published, for example where a story makes a negative claim about a named person or reports a deal that has not been confirmed.
How sources are chosen and checked
Stories are based on sources a reader can check: company announcements, regulators' and courts' documents, official statistics and reporting by other publications, which is credited. Sources are linked at the end of each article. Before publication, each article is checked line by line against its sources by a separate AI step.
A fact is used only if it was found in a source CoverWire could open and read. If a source cannot be accessed, for example because it is behind a paywall, it is not relied on: the story uses another source that can be checked, or leaves the point out.
Facts and analysis
News stories report what the sources say and attribute it. Where an article goes beyond the sources, to explain why something matters or what might follow, it is written as context or analysis rather than as fact.
Corrections and contact
If CoverWire gets something wrong, email [email protected] with the story and what is wrong. Corrections are made in the article with a dated note saying what changed. See the corrections policy and editorial standards.
CoverWire and data
CoverWire is also building an insurance data service: a dataset of UK insurance businesses drawn from public records, developed by Dt5 Consulting. It is in development and is not yet available. The data work does not decide what CoverWire covers or how it covers it, and no company can pay for coverage. Register interest in CoverWire data.
The full process, step by step, is on How CoverWire works. Significant changes to how CoverWire works are recorded on Building CoverWire.