--- title: "Home" url: "https://humrik.com/" description: "Understand, Predict and Change Human Risk. Humrik connects behavioral intelligence, governed AI and adaptive learning to build a more resilient workforce." updated: 2026-11-06T15:55:25Z summary: "Understand, Predict and Change Human Risk. Humrik connects behavioral intelligence, governed AI and adaptive learning to build a more resilient workforce." --- # Understand. Predict. Change. **Human risk. Human resilience.** Your people are your greatest potential. Turn everyday behavior into intelligence. And intelligence into a stronger human defense. - BEHAVIORAL DNA: Understand the individual - PREDICTIVE INTELLIGENCE: See beyond the signal - ADAPTIVE DEFENSE: Build safer habits Human-centered. AI-powered. Enterprise-ready. Connected to your stack. Focused on your people. - Microsoft 365 - Google Workspace - okta - CROWDSTRIKE - splunk ## Awareness is a start. Real change is the goal. Connect the dots between what people do, where risk is emerging, and what will actually help. One platform. A continuously stronger human layer. ### Understand — See the complete picture **Every signal. A more human picture.** Bring identity, endpoint, email, and learning signals together. Behavioral DNA reveals the context behind each person’s risk, so your next step is an informed one. - Continuous, individual risk scoring - Behavioral baselines and anomaly detection - Department-level visibility and attribution ### Predict — Get ahead of emerging risk **Spot the pattern. See what comes next.** Connect behavioral anomalies with identity, privileges, and exposure. Surface emerging risk and likely attack paths, with the context your security team needs to prioritize. - Predictive behavioral risk modeling - Identity relationships and attack-path context - Explainable signals for informed triage ### Change — Make safer behavior second nature **The right moment. A lasting difference.** Translate risk intelligence into relevant action. Give each person the learning, practice, or nudge they need, then measure the outcome and adapt the next step. - Personalized 90-second micro-learning - Adaptive simulations and gamified quests - Quiet hours, workload awareness, and outcome tracking ## Meet IRIS. Your next advantage. A governed team of specialist AI agents that turns complex signals into considered action. IRIS connects risk, learning, and response—within the boundaries you set. - **Autonomous, with accountability.** Policy gates, human review, and traceable decisions. - **Personalized, with empathy.** Right-time interventions that respect workload and quiet hours. - **Private, by design.** Tenant-specific models and private inference options. - **Risk triage** — Scores every incoming signal - **Behavioral baseline** — Knows this person's normal - **Blast radius** — Maps what the account can reach - **Policy & governance** — Applies your rules and approval gates - **Containment** — Steps up sign-in, revokes sessions - **Wellbeing guard** — Respects workload and quiet hours - **Adaptive content** — Picks the right 90-second lesson - **Evidence & reporting** — Seals a receipt for every action ### Case 01 · Contain — A lookalike invoice link is clicked - 09:41:02 **Risk triage:** Link leads to a lookalike invoice portal with a credential form. Scored critical. - 09:41:04 **Behavioral baseline:** Usual laptop and city, no new sign-ins since the click. No takeover yet. - 09:41:04 **Blast radius:** This account can approve supplier payments. High blast radius. - 09:41:09 **Policy & governance:** Policy FIN-07: containing a payments user needs a human. Routed to the on-call analyst. - 09:41:47 **On-call analyst:** Reviewed the evidence and approved containment in 38 seconds. - 09:41:52 **Containment:** Session revoked, step-up MFA forced, and the link blocked for all 61 recipients. - 09:42:10 **Wellbeing guard:** Quarter-end close today, so the lesson waits until after 17:00. - 09:42:11 **Adaptive content:** 90-second invoice-fraud comic queued for 17:05, in their own language. - 09:45:08 **Evidence & reporting:** Receipt sealed: 7 agents, 1 human approval, every action traceable. ### Case 02 · Verify — Two countries, 40 minutes apart - 01:44:10 **Risk triage:** Sign-in from Frankfurt 40 minutes after Delhi, on a new browser. Scored high. - 01:44:12 **Behavioral baseline:** Travels often, but has never signed in at 01:44 local time. - 01:44:12 **Blast radius:** Holds export rights for 4,200 customer records in the CRM. - 01:44:15 **Policy & governance:** A step-up check is pre-approved for this pattern. No human needed. - 01:44:16 **Containment:** Step-up MFA pushed to the known phone; CRM export paused. - 01:47:30 **Behavioral baseline:** Confirmed on the known phone: a hotel VPN. Baseline learns the route. - 01:47:31 **Wellbeing guard:** It is 01:47 for them, so nothing else is sent before 09:00. - 01:47:32 **Adaptive content:** Two-minute tip on hotel Wi-Fi and VPNs scheduled for the morning. - 01:50:40 **Evidence & reporting:** Receipt sealed; the false alarm is labelled so it is not raised again. ### Case 03 · Change — A third simulated-phish click - Day 0 **Risk triage:** Third simulated-phish click in 60 days. Real-world exposure is rising. - Day 0 **Behavioral baseline:** Every click came after 21:00 on a phone. Daytime reporting is good. - Day 0 **Wellbeing guard:** Three weeks of overtime. Simulations paused for five days. - Day 0 **Policy & governance:** People-first policy: coaching, not blame. No escalation to their manager. - Day 1 **Adaptive content:** Video modules swapped for a 90-second comic about late-night lures, mornings only. - Day 14 **Risk triage:** Next simulation reported from their phone within 2 minutes. - Day 30 **Evidence & reporting:** Report rate up 3×, risk back in the safe band. Receipt sealed. Illustrative cases. Agent roles reflect how IRIS divides the work; people are pseudonymised and timings are modelled. ## Small moments. Stronger security. Move from annual training to everyday progress. ### Practice for the real world. Adaptive phishing simulations build the instincts to recognize and report social engineering. [Phishing & simulations](https://humrik.com/#platform) ### Learning that meets the moment. Short, relevant experiences, interactive comics, and gamified quests make safer habits stick. [Adaptive learning](https://humrik.com/#platform) ### Show the change that matters. Turn behavioral outcomes into executive reporting, department insights, and verifiable audit trails. [Risk insights & reporting](https://humrik.com/#platform) ## Fits your ecosystem. Elevates your defense. Bring signals from the tools you already use into one human-risk view. Identity, email, endpoints, and collaboration—all with context. - Microsoft 365 — email & identity - Okta — identity - Google Workspace — collaboration - CrowdStrike — endpoint - Slack — collaboration - Splunk — security signals ### Your signals. One perspective. Humrik’s technical specification describes connectors for these systems. Deployment scope and permissions are reviewed during onboarding. - **Microsoft 365 & Azure AD:** email and identity signals - **Google Workspace:** workspace security activity - **Okta:** identity and authentication - **CrowdStrike:** endpoint security telemetry - **Slack:** collaboration signals - **Splunk:** security event context Signal normalization, pseudonymization, and tenant-scoped ingestion support a consistent view across sources. ## People aren’t the weakest link. They’re the most responsive one. ## Enterprise strength. Human at heart. Your people deserve thoughtful security. Your organization deserves control. Humrik is designed for both. - **Privacy built in**: Tenant isolation, pseudonymized signals, and configurable data governance. - **Your policies. Your control.**: Role-based access, step-up authentication, and approval gates for sensitive actions. - **Evidence you can trace**: Cryptographically verifiable audit chains connect actions to outcomes. - **Built for your environment**: Private AI options and integrations across your existing security ecosystem. ## The same signal, framed for what you own. ### Security & awareness — A program that adapts on its own. Stop sending one monthly simulation to everyone. Training and difficulty retarget to the people who need them, automatically. - Simulations retargeted by real click behavior - Lessons assigned the moment a risky action happens - One view instead of three disconnected tools ### SOC & IT — The human layer, without another console. IRIS reads what your identity provider, mail gateway and EDR already log, and hands your SOC a correlated view. - Read API into the signal graph for SIEM and SOAR - Escalations arrive with rule and evidence attached - Containment behind the approval gates you set ### Privacy — Guardrails you can sign off on. Every automated decision logs the policy it followed. Behavior data never leaves without a traceable reason. - Every action logged with its rule and evidence - Pseudonymized signals and tenant isolation - Plain enough to describe in your own privacy notice ### Leadership — One number the board can follow. A single trended human-risk score, backed by evidence, instead of a once-a-year click-rate slide. - One trended score instead of several vendor reports - Time to contain, not just click rates - A pilot before any procurement commitment ## Thinking on human risk. [All resources](https://humrik.com/resources.html) Unlock your people’s potential. Change your risk. See what a more intelligent, more human defense looks like.