AI-assisted analysis supports security decision-making. It does not replace human security judgment.
AI can make mistakes.
An explanation can be incomplete, misleading, or wrong. Treat it as one input to a decision.
Low risk is not a safety guarantee.
An assessment may miss information or an unfamiliar attack. Continue to follow your organization’s verification procedures.
High risk is not proof of intent.
A suspicious pattern may have an innocent explanation. Investigate context before drawing a conclusion.
Verify important requests independently.
Use an established contact method for payment instructions, account changes, or sensitive information requests.
Keep people in the response loop.
Security teams should investigate serious incidents. AI-supported recommendations do not replace an incident-response process.
Share only what is necessary.
Avoid unnecessarily submitting highly sensitive information. Redact unrelated details and credentials before any future analysis.
OUR APPROACH
Multiple signals. Clear limitations.
Future versions may combine deterministic security rules, threat intelligence, domain and URL analysis, pattern detection, and AI-assisted explanation. Sources and capabilities will be disclosed as they become available.
We aim to show the reasoning behind an assessment, distinguish observed signals from interpretation, and make uncertainty visible. This is decision-support technology—not a promise that every attack can be identified.