6,040companies tracked
37,696filings held
723frameworks
332,959control mappings
Cite this data
Signals, The Art of Service. Analysis of SEC full-text search across 8-K, 10-K and 10-Q filings. Accessed 27 July 2026. https://signals.theartofservice.com/
No permission required and no
licence to sign. If you cite it we would like to know, but you do not need to ask.
Rising fastest
Ordered by how surprising the change is at the size it happened, rather than by
percentage, so a large move on a large base is not buried beneath a small move on a small
one.
| Disclosure | Now |
Before | Change | Window |
|---|
| Artificial Intelligence | 208 | 91 | +129% | 8-K, past 7 days |
| AI Governance | 75 | 17 | +341% | 10-K/10-Q, year on year |
| Ransomware | 43 | 13 | +231% | 8-K, past 7 days |
| Operational Resilience | 111 | 59 | +88% | 10-K/10-Q, year on year |
| Cybersecurity Incident | 19 | 3 | +533% | 8-K, past 7 days |
| Data Governance | 101 | 58 | +74% | 10-K/10-Q, year on year |
| Privacy Compliance | 115 | 79 | +46% | 10-K/10-Q, year on year |
| Restatement | 83 | 57 | +46% | 8-K, past 7 days |
| Internal Control | 88 | 63 | +40% | 8-K, past 7 days |
| Business Continuity | 642 | 573 | +12% | 10-K/10-Q, year on year |
Falling fastest
Companies writing something less often is a finding too, and it is one almost nobody
reports.
Machine readable
Updated every morning. No key, no rate limit, no sign-up.
| File | What it is |
|---|
| /daily.json |
Today's edition: every tracked phrase, its counts, its change, and the named
filers |
| /series.json |
Monthly filing counts per phrase back to 2008 |
| /coverage/index.json |
Every framework in the corpus with its control count |
| /sitemap-index.xml |
Every page, including the framework comparisons |
How the numbers are produced, in full
Counts come from SEC full-text search. The fast series counts 8-K filings containing a phrase
over the last seven days against the seven before, and the window slides a full day every
morning. The slower series counts 10-K and 10-Q filings this quarter against the same quarter
a year earlier. Filers are named, and every count is one click from the filing on EDGAR.
Tables are ordered by how far apart two counts are relative to the
variation expected at that scale, not by percentage. Percentage alone would rank a move from
three companies to ten above a move from five hundred and seventy three to six hundred and
forty two.
The caveat we would want quoted alongside any number.
A phrase appearing in a filing usually means the risk section mentions it, not that an event
occurred. Where an 8-K item number indicates an actual event, such as Item 1.05 for a material
cybersecurity incident, the topic pages separate those from routine mentions, and the routine
group is almost always much larger. A headline of the form "N companies disclosed X" is true
and is easily misread.
Nothing here is generated, predicted, scored or characterised. There
are no rankings of what companies are "most worried about", because filing frequency does not
measure worry and we will not imply it does.
Need a cut we do not publish?
By sector, by state, by framework, by company, or a longer history. The data is there and we
would rather it was used. Ask.
The corpus behind it