DEFENSE AND NATIONAL SECURITY
SITREPs explain the situation. They don’t explain the people.
MOSAIC turns OSINT (and, when available, your HUMINT) into a Behavioral Briefing™: an evidence-tagged model of how a specific leader or key actor is likely to decide under defined conditions. It feeds the analysis you already run without replacing the analyst.
Request a briefing on a public foreign leader
Leadership files, threat assessments, and network maps stay structural.
Without the person-layer, analysts fall back on mirror-imaging or on the variables easiest to collect.
The collection is funded. The step that turns it into a read on the person isn’t.
Defense work is limited to foreign subjects.
Leadership Analysis
From who a leader is to how they’re likely to decide.
The problem. Leadership profiles describe biography, positions, and affiliations. They tell you who someone is, but they never tell you how they’ll decide when the situation changes. When the file runs out, the assessment fills the gap with assumptions.
What you get. Calls on how the leader is likely to decide under defined conditions: rising pressure, loss of face, a threat to their position. Each call carries its evidence and a confidence level. Where the signal is thin, the call isn’t scored. Each briefing also lists the observable indicators that would confirm or break each call.
Where it goes. Leadership profiles and all-source assessments, indications and warning, deterrence and influence planning, and red-cell or wargame play.
Illicit Networks
The map shows who’s connected. It doesn’t show what they’ll do.
The problem. Network analysis tells you who connects to whom, and the platforms that do it are mature. What it doesn’t tell you is how the person at a key node will act once the network is disrupted. Interdiction ends up designed against structure, not behavior.
What you get. Behavioral reads on the key figures in a network you’ve already mapped. How each is likely to respond to pressure. Who holds real authority, as opposed to the title. How succession is likely to play out if a node is removed. Every call carries its evidence and confidence.
Where it goes. The network analysis your team already runs in existing platforms, plus targeting, interdiction, and disruption planning.
Foreign Agents
Thousands of foreigners work in US companies. Spotting the ones who could pose an IP theft risk is mostly guesswork.
The problem. There are entire task forces dedicated to dealing with foreign influence and the theft of trade secrets on US soil. However, the people tasked with going through each individual or group still do hours of manual work: for any task force this means that the process is slow, hard to audit, and ultimately depending on human bandwidth.
What you get. A breakdown on the patterns of a potential foreign agent’s past, their likely behavior on US soil, the flags to watch out for, and a map of their network. MOSAIC enables task forces to make better, faster decisions on larger groups of foreign nationals.
Where it goes. In the hands of those who make the decisions.
HOW IT FITS
Built to feed your workflow, not to replace it.
API-first. Briefings arrive inside the tools your team already uses, through an API, with MCP support for AI-enabled tools. No new dashboard and no new login.
Analyst in the loop. Every call is contestable. The evidence sits next to it, so the analyst decides what goes into the assessment.
Documented behavior only. No faces, voices, or physiological signals. No protected characteristics inferred.
Accreditation, stated honestly. MOSAIC isn’t FedRAMP authorized or accredited for classified networks today. FedRAMP and CMMC are on our roadmap, and air-gapped deployment is in development. Until then, work is unclassified and built from open sources. We’re pursuing SBIR and OTA pathways.
TSA spent about $900 million on behavior detection. GAO reported that the research behind it put human detection of deception at about chance (GAO-14-159). Anyone in this category has to answer for that.
Our answer: documented behavior, evidence on every call, and a record of where we were wrong.
Test it on someone you already know.
Pick a public foreign leader: a head of state, a minister, or a military commander with a public record. We’ll return a Behavioral Briefing. Judge the method on a subject you’ve already studied.
Please don’t send names connected to ongoing operations or investigations. We’ll take that conversation offline.