Frequently Asked Questions

Hadrian’s "Verified Risks" eliminate the need for manual proof of work by leveraging an AI Orchestrator that autonomously validates vulnerabilities through safe, real-world attack simulations, effectively filtering out the false positives that plague traditional management. This process replaces time-consuming manual investigation with automated evidence generation, providing developers with detailed Proof of Concept (PoC) logs, attack chain descriptions, and specific remediation instructions. By further automating the lifecycle through instant regression testing to confirm successful fixes, Hadrian ensures that security teams can bypass the tedious validation phase and focus entirely on remediating genuine, confirmed threats.

Hadrian refines its AI risk scoring model through a continuous feedback loop where user-driven adjustments to risk severity directly train the machine learning algorithms. When users manually re-categorize a risk based on internal context, the model learns to tailor future assessments, while Hadrian’s in-house hacker team simultaneously reviews these changes to validate algorithmic accuracy. Although analysis indicates that only a fraction of a percent of risks are typically re-categorized—suggesting high initial precision—this combination of automated learning and human oversight ensures the platform constantly evolves to match the specific security reality of each organization.

Hadrian distinguishes between "Potential" and "Verified" risks by leveraging its AI Orchestrator to actively validate theoretical vulnerabilities rather than relying solely on passive detection. While Potential risks are identified through non-intrusive methods like version fingerprinting and represent hygiene issues or theoretical susceptibilities, Verified risks are confirmed threats where the AI has successfully simulated an ethical hack to prove exploitability. This validation process eliminates false positives by providing concrete Proof of Concept (PoC) evidence for every confirmed flaw, allowing security teams to filter out noise and prioritize immediate remediation for genuine, actionable threats over theoretical exposures.

Hadrian prioritizes risks using a proprietary, context-aware machine learning algorithm that goes beyond standard vulnerability scores (like CVSS). Instead of treating every vulnerability equally, the platform assesses the actual danger an asset faces based on "the hacker's perspective."

Hadrian’s Subwiz AI transcends standard dictionary attacks by employing machine learning to generate intelligent permutation alterations of wordlists based on analyzed naming patterns. This predictive approach identifies hidden subdomains absent from public records, which are then instantly validated through SanicDNS—a high-speed resolution tool operating 200 times faster than traditional alternatives. By combining AI-driven prediction with rapid validation, Subwiz uncovers over 10% more subdomains than conventional methods, effectively revealing Shadow IT and assets that evade standard detection.

Hadrian uncovers Shadow IT and hidden infrastructure by combining broad internet scanning, cloud integrations, and proprietary AI to map the complete external attack surface. The platform moves beyond standard technical scanning by utilizing visual fingerprinting to link disparate assets to an organization, while predictive AI models identify unlisted subdomains and dynamic IP addresses. This continuous, event-driven process triggers immediate assessments upon detecting changes , ensuring that even forgotten environments, third-party SaaS applications, and misconfigured cloud resources are detected and validated in real-time.

Hadrian performs asset inventory through a continuous, automated, and agentless process. Unlike traditional methods that rely on manual spreadsheets or periodic checks, Hadrian dynamically maps an organization’s entire digital footprint starting with a single domain or brand name.

Event-based scanning is a dynamic security testing approach where active vulnerability assessments are triggered by specific changes or "events" within an organization's environment, rather than relying solely on static, pre-scheduled batch scans.

Hadrian’s AI agents operate primarily through a central engine or data layer. This system is designed to autonomously mimic the decision-making processes and behaviors of human ethical hackers to discover, validate, and prioritize risks without manual intervention.