Intelligence that works while you are not looking.
Patena reads thousands of sources, clusters them into stories, scores risk for the subjects you track, and tells you when something changes.
The dashboard an analyst sees at 14:58 UTC: the last 24 hours of events on the map, each marked by its domain glyph; the five stories with the most coverage; the time scrubber to replay the window. Hold ⌘ and scroll to zoom the map.
The phone feed at 14:58 UTC. Map is the last 24 hours of events on the globe; For you ranks the five stories with the most coverage; Live feed lists them as they came in. Tap a tab to switch.
How it works
Thousands of feeds in thirteen languages, plus structured event sources, read continuously.
Articles about the same event become one story; its coverage is counted, not repeated.
Entities, places and the domain are pulled from each story and put on the map.
Risk is scored for the subjects your organization tracks, with history.
Monitors are rules you write once. Atlas runs them on a schedule against the stream and your thresholds.
In-app, by email or webhook. Each hit can be acknowledged or snoozed.
Situational awareness
The live map, stories ranked by how many sources cover them, and a daily brief for the subjects you track.
Investigation
Entities and places extracted from every article, keyword and semantic search across thirteen languages, risk scores with their history.
Agents
Atlas answers with citations. Background agents run on a schedule, read, notify and draft reports, and ask before they change anything.
Halvard Shipping AS was named in this morning's EU sanctions annex 1. Two of your suppliers list it as a carrier 2. I have not changed any monitor.
Monitors. Say what matters once.
A monitor is a rule your organization writes: these subjects, these places, these domains, this severity, this much evidence, this schedule, these people. Patena runs it against the stream around the clock. Atlas reads every hit before you do, explains it, holds what is not confirmed, and says when the rule should change.
Atlas: the strike meets the rule by a wide margin. I held Tallinn: one T3 source, no corroboration. In the last 7 days this rule would have fired once at severity ≥ 3; the threshold looks right.
Subjects, places with a radius, domains, severity on the four-bar meter, how many sources of which tier it takes to count.
A monitor can fire on score drift: a subject’s risk moves by more than a threshold in a window.
In-app for everyone; email and webhook per analyst, per severity, with quiet hours. Acknowledge or snooze, and Atlas remembers.
“Watch our Baltic suppliers for strikes and sanctions” becomes a full rule. You review, test against the last 7 days, and save.
Twelve domains
Every event is filed under one domain, with its own glyph and hue. A domain is never shown by colour alone.
Patena Agents
Atlas answers questions with citations. Background agents run on a schedule for each analyst, read the stream, notify and draft reports. They ask before they change anything.
One. Halvard Shipping AS was added to the EU annex on Tuesday 1 and appears in OFAC's update the same day 2. Two of your suppliers list it as a carrier 3. I have not changed any monitor; say so and I will add one.
Agents see the same stories, entities and scores as the analyst who owns them. Nothing more.
Every statement carries its sources. An answer without a source is not given.
Hits go where you asked: in-app, email or webhook. Acknowledge or snooze, and the agent remembers.
Agents may draft and propose. Monitors, subjects and settings change only when a person approves.
Atlas uses the same tools as your analysts.
Everything an analyst can do in Patena, an agent can do for them: read the stream, search, score, watch, draft. Within their scope, with their citations, and never past the point where a person has to say yes.
One. Halvard Shipping AS was added to the EU annex on Tuesday 1 and matched OFAC's update the same day 2. Two of your suppliers list it as a carrier 3. Its score moved from 68 to 74.
I have drafted a monitor for the two suppliers and changed nothing. Approve it and I will watch them hourly.
One agent per analyst
Atlas reads your subjects, answers in your language and proposes the next step. The dashboard is where you check its work.
Reads the whole stream
Thousands of feeds and every structured source, through one pipeline, including the private sources only your organization sees.
Runs your monitors
A monitor is a rule you write once: subjects, places, domains, severity, evidence. Atlas runs it every 15 minutes, hourly or daily, explains each hit and skips what you already acknowledged. See monitors
Cites every claim
A sentence without a source is not written. Hover a citation and the outlet, tier, language and time are there.
Asks before it changes anything
Agents draft and propose. Monitors, subjects and settings change only when a person approves.
Getting started
Name it, invite your analysts, set roles.
Companies, places, routes. Set the thresholds that matter.
In-app, email, webhook. Per monitor, per person.
It reads, scores, notifies and drafts. It asks before changing anything.
Questions
Early access, by request.
Tell us what you track and we will set up your organization.