What is human risk management (HRM)?
Human risk management (HRM) is a security discipline that measures, reduces and contains the cyber risk created by people's everyday decisions, such as clicking a link, approving a payment or sharing data. It uses behavioral data to decide who needs guidance at the moment of risk, who needs extra protection and who needs training.
For years, organizations handled the human side of security with annual courses and the occasional phishing test. Human risk management keeps those tools but asks a different question. Not "did everyone finish the course?" but "who is most likely to make a costly mistake this month, and what stops that mistake from becoming an incident?"
This guide explains what human risk is, why HRM is gaining ground, the four building blocks of a human risk management program, how to measure human risk, how HRM maps to NIST CSF 2.0, ISO 27001 and NIS2, and how to get started.
What is human risk? Examples in cybersecurity
Human risk is the likelihood that a person's action, or failure to act, leads to a security incident. It is rarely malicious. Most human risk comes from people who are busy, under pressure or facing an attack designed to look like routine work.
Human risk in cybersecurity usually takes one of three forms:
- Split-second actions. Following a link in an urgent message, typing a password into a cloned sign-in page, scanning a QR code on a fake parking notice or approving a changed bank account after a call that sounds like the CFO.
- Everyday habits. Using the same password for work and personal accounts, tapping "approve" on a multi-factor authentication prompt nobody requested, pasting customer data into an unapproved AI tool or leaving phone and app privacy settings on their defaults.
- Things people don't do. Ignoring a suspicious message instead of reporting it, postponing a security update or letting a training session expire.
The third group is easy to overlook. A message nobody reports is an early warning the security team never receives. And because these risks appear in email, chat, text messages, QR codes and the browser, security teams increasingly talk about the human attack surface rather than the inbox alone. In the 2026 Verizon Data Breach Investigations Report, the human element was present in 62% of breaches.
Why human risk management matters now
Three shifts have pushed human risk management up the agenda for CISOs and security awareness leads.
- Attacks are better written and more personal. Generative AI lets attackers produce fluent, tailored phishing in any language and at scale, and voice cloning makes phone fraud more convincing. The spelling mistakes people were trained to look for are disappearing.
- Attacks have moved beyond email. Text messages, QR codes, Microsoft Teams and Slack messages, shared-document links and fake browser pages reach employees directly, often outside the controls built around the email gateway. In the 2026 Verizon DBIR, 41% of social engineering breaches involved channels other than email.
- Training metrics are no longer enough. A completion rate shows that people sat through a course, not that they behave more safely. Boards, auditors and regulators increasingly ask a harder question: is our human risk actually going down?
Analysts reflect the same move. In 2024, Forrester published its first evaluation of the market as human risk management solutions rather than security awareness and training, and Gartner calls the wider goal a security behavior and culture program (SBCP), listing it among its top cybersecurity trends for 2025. In both framings, training becomes one input of a broader program.
The building blocks of a human risk management program
A mature human risk management program combines four building blocks that feed each other: measure, guide, protect and train. Measurement decides where guidance and protection go, and every interaction produces new data to measure.
management
Measure behavior, not just completion
Start from what people actually do. Useful signals include how employees respond to phishing simulations, whether and how quickly they report suspicious messages, whether they interact with real threats that reach them, how they perform in training and whether their apps and devices are configured securely. Combined, these signals produce a risk score per person, team and organization that updates continuously.
Example: a finance assistant who reports every simulation within minutes and a new hire who has clicked twice in one month need very different things. Measurement is what lets you tell them apart.
Guide people in the moment of risk
Most mistakes happen in a few seconds in the inbox or a chat window, not in a training module weeks earlier. In-workflow guidance helps employees at that exact moment: instant feedback when they report something, an explanation of why a message is suspicious and a safe way to check a link they are unsure about.
This matters because most employees are neither clickers nor reporters. They hesitate, do nothing and learn nothing. Example: an employee receives a "shared document" link in Teams and is not sure it is genuine. Instead of guessing, they open it in an isolated environment, get an explanation of the warning signs and report it with one click. One real interaction becomes a learning moment.
Protect the people who need it most
Human risk is not spread evenly. Repeat clickers, new hires and high-value targets such as executives and finance teams often carry a larger share of it. Human risk management lets you apply proportionate controls: stronger protection for the highest-risk groups and a lighter touch for people who consistently make good decisions.
Example: for a group of repeat clickers, every incoming email is opened in an isolated environment, so a mistake cannot reach company systems while their coaching continues.
Train foundations and compliance
Behavior change still needs knowledge. A structured curriculum of short, practical sessions builds an understanding of how attacks work, compliance courses on topics such as GDPR and AI literacy give auditors the evidence they ask for, and timely threat alerts explain new techniques as they appear. The difference in an HRM program is that risk data decides who needs which training and when. Security awareness training stays essential; it simply stops being the only lever.
How to measure human risk: KRIs, KPIs and behavioral risk scores
Measuring human risk starts with separating two kinds of metrics. Key performance indicators (KPIs) show whether the program is running. Key risk indicators (KRIs) show whether risk is changing. Most awareness reports are built on KPIs alone.
| Metric type | What it tells you | Examples |
|---|---|---|
| Program KPIs | Is the program running as planned? | Training completion, simulation coverage, participation in courses |
| Behavior KRIs | Are people acting more safely? | Report rate, time to report, share of repeat clickers, interaction with real threats, secure settings adopted |
| Outcome KRIs | Is risk turning into fewer incidents? | Incidents caused by human error, compromised accounts, time to contain a reported threat |
Click rate deserves a warning. It mostly measures how people respond to simulations, it rises and falls with the difficulty of the simulations sent, and it says nothing about the large group of employees who never clicked. Report rate and behavior around real threats say far more about resilience. Two independent studies presented at Black Hat 2025 (ETH Zurich and the University of Chicago) reached a similar conclusion: simulations alone barely moved long-term behavior.
A behavioral risk score combines these signals into one number per person, weighted by how recent and how serious each action was. Benchmarked by team or department, it shows where risk sits today and lets you segment people automatically: first-time clickers, repeat offenders, new hires or anyone whose score drops below a threshold. Phished calls its version the Behavioral Risk Score™.
Human risk management frameworks: NIST CSF 2.0, ISO 27001 and NIS2
No major framework uses the term human risk management, but each expects people-related controls that are documented, ongoing and effective. A human risk management framework gives you the evidence for them.
| Framework | Relevant requirement | What HRM adds |
|---|---|---|
| NIST CSF 2.0 | Awareness and Training category in the Protect function (PR.AT): all personnel receive awareness and training so they can do their work with cybersecurity risks in mind (PR.AT-01), and people in specialized roles get training for their specific tasks (PR.AT-02) | Proof that training changes behavior, and targeting by risk instead of one program for everyone |
| ISO/IEC 27001:2022 | Clause 7.3 (awareness) and Annex A controls 6.3 (information security awareness, education and training) and 6.8 (information security event reporting) | Report-rate data for event reporting and continuous measurement for performance evaluation |
| NIS2 (Directive (EU) 2022/2555) | Article 20(2): management bodies must follow cybersecurity training and entities are encouraged to train employees regularly. Article 21(2)(g): basic cyber hygiene practices and cybersecurity training | Documented, recurring and risk-based measures rather than a single annual session |
Sources: NIST Cybersecurity Framework 2.0, ISO/IEC 27001:2022 and Directive (EU) 2022/2555 (NIS2).
Human risk management vs security awareness training
Security awareness training (SAT) is a set of activities, mainly courses and phishing simulations, usually judged on completion and click rates. Human risk management is the program around those activities. It measures behavior across channels, targets effort by risk, guides people in real time and adds protection where training alone is not enough. SAT answers "have people been trained?" HRM answers "is our human risk going down?" Most organizations do not replace SAT; they make it one input of HRM.
Security awareness training
Activities: courses and simulations
Measures completion and click rate
Mostly the email inbox
Human risk management
A program around those activities
Measures behavior and risk over time
Every channel people use
How to start a human risk management program: 5 best practices
- Set a baseline. Gather what you already have: simulation results, report rates, incidents linked to human error and training completion. Note the blind spots, such as no data on text messages or chat.
- Decide which risks matter most. Payment fraud for finance, credential theft for IT admins, data handling for HR. Agree three to five KRIs with leadership before you change anything.
- Make reporting easy and worthwhile. One-click reporting with instant feedback is the fastest way to turn hesitant employees into an early-warning network.
- Segment and act on risk. Use risk scores to define groups and match each group with the right mix of coaching, protection and training. Review the groups every month.
- Report outcomes, not activity. Show leadership the trend in your KRIs, incidents avoided and time saved for the security team, then adjust the program on that basis.
If a full rollout is not realistic yet, start with one high-risk group, such as repeat clickers or executives, prove the effect and extend from there.
How Phished approaches human risk management
Phished built its AI-native human risk management platform around a single target: no incidents caused by human error. More than 7,500 organizations and 2 million users rely on it to measure, guide, protect and train their people across the channels attackers actually use.
- In-workflow Guidance & Protection. Phishing simulations adapt to each employee's role, language and skill level. Every report gets immediate feedback, and simulations are removed from the inbox once they end.
- Phished Assistant. When employees doubt an email or link, they open it in an isolated environment, a digital silo, so even a malicious link cannot reach company systems. AI analysis coaches them in real time in email, the browser and other channels such as chat, text messages, QR codes, and voice and video calls. Isolation applies a zero-trust principle to the human layer.
- Risk dashboard with the Behavioral Risk Score™. Employees are sorted automatically into risk categories such as first-time clickers, repeat offenders, expired training sessions, new hires and the C-suite. Admins set the protection mode per group, including turning on the Phished Assistant for one group first.
- Zero-Incident Mail™. For the highest-risk groups, every incoming email is opened in an isolated environment automatically, with no action needed from the employee.
- Automated incident response. AI analyzes every reported email and resolves around 99% of them automatically, which keeps the queue clear for the IT team. The roughly 1% that remains are genuine threats, and they reach admins with a recommended response.
- Phished Academy. A structured curriculum with Bronze, Silver, Gold and Platinum levels, compliance courses and threat alerts on new attack techniques.
Protect the full human attack surface
Train your people, guide them at the moment of contact and stop the attack before it runs. See it in your own environment.
FAQ
Frequently asked questions about human risk management
What is human risk management in cybersecurity?
In cybersecurity, human risk management means treating people's decisions as a measurable risk that can be lowered, not as a training box to tick. It combines behavioral data, real-time guidance, extra protection for high-risk groups and targeted training, and it judges success by lower risk rather than by completed courses.
What are some examples of human risks?
Human risks fall into three groups. Split-second actions, such as following a link in an urgent text or approving a payment after a spoofed call. Everyday habits, such as password reuse or pasting data into unapproved AI tools. And omissions, such as not reporting a suspicious message or letting security training expire.
What is the difference between human risk management and security awareness training?
Security awareness training delivers courses and phishing simulations and is usually tracked through completion and click rates. HRM is the wider program: it measures behavior across channels, targets effort by risk, guides people in the moment and adds protection. Training becomes one input of HRM rather than the whole approach.
How do you measure human risk?
You measure human cyber risk by tracking behavior rather than activity: report rate, time to report, repeat clicks, interaction with real threats and adoption of secure settings. These signals are combined into a behavioral risk score per person and team, and tracked over time as key risk indicators next to outcomes such as incidents caused by human error.
Is human risk management required by NIS2 or ISO 27001?
Neither uses the term human risk management, but both require people-related controls. NIS2 Article 21 includes cyber hygiene and cybersecurity training, and Article 20 requires training for management bodies. ISO/IEC 27001:2022 includes awareness, training and event reporting controls. An HRM program provides ongoing evidence for these requirements.
What is a human risk management platform?
A human risk management platform is the software that runs an HRM program day to day: it collects behavioral signals, keeps a risk score per person and team, and turns that score into guidance, protection or training. For a side-by-side view of the main vendors, see our ranking of the top 10 HRM platforms.
Human risk management glossary
- Human risk management (HRM)
- A security discipline that measures, reduces and contains the cyber risk created by people's decisions, using behavioral data, real-time guidance, risk-based protection and targeted training instead of relying on training completion alone.
- Human risk
- The likelihood that a person's action or inaction, such as clicking a phishing link, approving a fraudulent payment or not reporting a suspicious message, leads to a security incident. Most human risk is unintentional.
- Security awareness training (SAT)
- Courses and phishing simulations that teach employees to recognize and avoid cyber threats. SAT is usually measured by completion and click rates and forms one input of a human risk management program.
- Human attack surface
- All the ways attackers can reach and manipulate people in an organization, including email, text messages, phone calls, QR codes, chat platforms, collaboration tools and websites, rather than the email inbox alone.
- Behavioral risk score
- A score that combines behavioral signals, such as reporting, clicks, training results and device settings, into one measure of how well a person or team can withstand a security incident, updated over time. The higher the score, the better.
- Behavioral Risk Score™
- Phished's behavioral risk score. It is updated continuously from each employee's actions across the Phished platform and lets admins benchmark people against their department and organization and segment groups by risk.
- Key risk indicator (KRI)
- A metric that signals whether risk is rising or falling, such as report rate or the share of repeat clickers. In human risk management, KRIs complement activity-based KPIs such as training completion.
- Report rate
- The share of employees who report a suspicious or simulated message instead of ignoring or opening it. A rising report rate is one of the strongest signals that human risk is going down.
- In-workflow guidance
- Security coaching delivered at the moment of risk, inside the tools employees already use, such as instant feedback on a report or an explanation of why a link is suspicious. Also called just-in-time guidance.
- Isolation (digital silo)
- A security approach in which a suspicious email, link or web page is opened in a fully separated environment, so that malicious content cannot reach the user's device or company systems.