Archetypal AI Research
Investigate Archetypal DNA, emergent behavior, alignment risks, and Decision Sources of Record to interpret how agents act.
Portfolio / AI
Product
Archetypal AI
Research emergence. Govern behavior. Preserve the decision.
The company
Research emergence. Govern behavior. Preserve the decision.
Archetypal AI has two connected layers. Archetypal AI Research applies Archetypal DNA, Decision Sources of Record (DSRs), and related methods to interpret emergent agent behavior, alignment risks, and governance needs. Guarded is the operational AI Governance, Audit, Compliance, and Risk Platform, combining a harness, probe, control plane, and browser agent to observe AI use, enforce policy, preserve evidence, and govern decisions across enterprise workflows.
The problem it addresses
Organizations need to interpret emerging agent behavior and turn governance requirements into controls that operate where AI is actually used.
Capabilities
What Archetypal AI brings.
Guarded platform
Bring AI governance, audit, compliance, and risk into operational workflows through a harness, probe, control plane, and browser agent.
Decision evidence
Connect observation, policy enforcement, and review to the evidence needed to understand and govern consequential AI decisions.
Illustrative application
Archetypal AI in context.
A team observes AI behavior, applies policy at a decision point, probes an uncertain outcome, and preserves the evidence needed for review. Archetypal AI Research informs how emergent behavior and alignment are interpreted.
A practical starting point
- Identify the AI systems, organizational policies, decisions, and behaviors that need oversight.
- Apply the relevant observation and control surfaces, with explicit human authority and escalation conditions.
- Review decision records and observed behavior, then revise policies or evaluation boundaries as evidence changes.
What to evaluate
- Coverage of relevant AI interactions
- Traceability from policy to decision evidence
- Handling of uncertain or emergent behavior
The intended outcome
Governance, audit, compliance, and risk operations connected to the decisions AI systems actually make.
This application describes product direction and an evaluation approach, not a completed customer deployment or independently validated performance.
Inside the company
Products and research.
3 products & product families
View company products
5 research projects
View company research
Archetypal LLMsArchetypal AI · Research design
Archetypal LLMs
Language models shaped by organizational policy, decision logic, and structured archetypal feedback.
Archetypal DNAArchetypal AI · Proposed study
Archetypal DNA
Reproducible behavioral tendencies and the mechanisms through which agents retain and transfer them.
Archetypal EmergenceArchetypal AI · Research design
Archetypal Emergence
How tasks, incentives, interaction, and institutional context shape collective agent behavior.
Decision Source of Record (DSR)Archetypal AI · Research framework
Decision Source of Record (DSR)
A durable, reviewable record connecting a decision to its rules, evidence, rationale, and disposition.
MissionSAFEArchetypal AI · Research framework
MissionSAFE
Task-specific safety and control assessment with critical failures kept visible.
Intellectual property
Patents, licenses
and development.
US application 19/438,727
Filed · company-reportedArchetypal LLMs trained on decision logic
US application 19/393,310
Filed · company-reportedDecision Logic Source of Record
Development program · no application number recorded
In development · filing contemplatedArchetypal–NSA derivative patent development
Connected ecosystem
Where the capabilities meet.
Explore the complementary companies that connect Archetypal AI to the wider product journey.
ComputeCompacitas
The verifiable civilization runtime for model governance, scheduling, optimization, and authorized compute federation.
CyberCybersecutitus
A governed AI civilization for secure product assessment, cybersecurity validation, and authorized cyber missions.
CyberDefeated AI
Authenticity, provenance, and attribution infrastructure for the synthetic-media era.
Related perspectives
Studio perspective / AI
From an aha moment to an accountable AI product
Why the studio begins with the problem, the product owner, and the evidence that will make a decision possible.
Read perspective
Research note / AI
Beyond the harness: building with an AI civilization
A model connected to tools is not the same thing as a coordinated environment for product development.
Read perspective