Skip to content

How we source

A brief is only as good as its sources. Our AI system writes each brief from public sources and tailors it to your role, work, and priorities, following the rules on this page.

We present our work transparently, because leaving things out is the easiest way to mislead, even when every sentence is true.

Sample · one development from a brief

Houthi gains at Bab el-Mandeb raise shipping risks

Houthi forces reportedly captured Perim Island, in the strait, on Friday. Brent slipped below US$105.

Rule 1 Sources NBC News · CNBC · not yet independently confirmed
Rule 2
Potential impact on you · ScrollField assessment, not reported fact

If your lanes touch the Red Sea, check what your carriers’ surcharge terms allow.

Rule 3
Stage · followed since 20 July
  1. Announced, done
  2. Committed, done
  3. Underway
  4. Concluded, not yet
Rule 4

How this was researched Official, market, and shipping sources checked · 14 developments found · 12 left out

Rule 5

Some shipping trackers could not be reached this morning. This brief is partial for freight data.

Sample, not a live brief. The numbers 1 to 5 match the rules.

Our rules, keyed to the specimen

  1. Rule 1 Where facts come from, and how we corroborate them We start with official sources and look for a second, independent one for anything that matters.

    We start where the fact started: governments, regulators, companies, and other official bodies, through their own publications, filings, and databases. We add credible specialist media and research to confirm and explain it.

    If acting on a claim could cost you money, time, or reputation, we look for a second, independent source before it reaches you. Independent means another organization doing its own reporting, not the same news-agency story printed twice. If there is no second source, we drop the claim or tell you so.

    We put each fact next to its sources, with links to the pages they came from. We name each publisher and show which source is official, when there is one, and which confirms it independently.

  2. Rule 2 What happened, and what we think it means Our reading is labeled as ours, kept apart from what was reported.

    We keep the two apart. Our reading is marked as ours: the Potential impact on you section, and any line that begins Our reading. It interprets the sourced facts beside it, so it has no link of its own, and where the evidence is uncertain, it says so. It is the part most likely to miss the mark. When it does, tell us, and the next brief reflects it.

  3. Rule 3 Announced, or actually happening? Each story shows how far it has gotten: announced, committed, underway, or concluded.

    An announcement can matter on its own, but it is not yet a change on the ground. We follow each story from brief to brief and mark its stage, so you can tell what is underway from what has only been announced or committed. The stage says how far events have gotten, not how sure we are of them: the sources line says that.

  4. Rule 4 What we covered and left out Each brief opens with how it was researched: the kinds of sources we checked, and how many developments we left out.

    It also says how many we found in all. We leave one out when it matters too little to your work, or the evidence is too thin. If we missed something that mattered to you, each brief has a place to tell us.

  5. Rule 5 When a scan falls short A failed scan never pretends to be a quiet day.

    Sources go down, paywalls close, and our processing can fail. If any of these happens, your brief says so and names the kinds of sources we could not check.