How stories are produced
We collect and archive material from a range of public sources. We remove duplicates, identify the companies and events involved, and check important facts against other available evidence.
Laplace Atlas writes its own stories from that research. It does not publish summaries or rewrites of source articles. Every factual claim must be supported by a recorded source. Links appear as optional further reading at the end.
Confidence
We consider the quality of each source, independent confirmation, consistency, and the precision of the available facts. Stories with weak evidence are not published. We prefer primary sources. When only secondary reporting is available, we treat the information cautiously.
AI and editorial responsibility
AI helps us extract facts, classify material, connect related reports, and prepare drafts. Software handles routine checks such as duplicate detection, dates, links, and company names. We do not treat model output as fact. It must pass our checks before publication. The full approach is set out in the Editorial & AI Policy.
Corrections
When better evidence changes a fact, Laplace Atlas records the correction instead of quietly replacing the earlier account. Readers and subjects can use the published corrections process.
Accuracy audits
Each week, we review a sample of stories for company identity, amounts, dates, investors, unsupported claims, duplicates, and incorrectly combined events.