Phishing benchmarks show awareness training sharply reduces risk

At a glance

What this is: This is a phishing benchmark report showing that South America’s initial click risk is high but drops materially with sustained security awareness training.

Why it matters: It matters because phishing is still an identity-adjacent entry path into accounts, and the data shows that human risk can be reduced when controls are continuous rather than one-off.

By the numbers:

• AI-generated content now appears in 82.6% of phishing emails, according to Knowbe4.

Context

Phishing remains a governance problem because the attacker’s first control objective is not malware delivery but human trust. In this case, the article focuses on South America, where baseline susceptibility is unusually high and improves materially only when training is sustained over time.

The identity angle is straightforward: phishing is often the entry point to account compromise, credential theft, and downstream access abuse. For IAM, PAM, NHI, and human identity programmes, the question is not whether awareness matters, but whether it is treated as a continuous control with measurable reduction in exposure.

Key questions

Q: What reduces phishing risk when awareness scores are already high?

A: Risk falls when awareness is continuous, measured, and tied to real user behaviour rather than completion rates. Organisations need repeated simulations, rapid feedback, and clear reporting paths so caution becomes a habit under pressure. A strong programme also targets high-risk business actions, because the most damaging phishing outcomes often come from account recovery, approvals, or credential reuse.

Q: Why do non-email phishing campaigns increase enterprise risk?

A: They increase risk because they bypass the controls most organisations built around email, while still targeting the same identities and the same SSO-connected applications. That creates more entry paths, less visibility, and a wider blast radius when one account or session is compromised.

Q: What are the signs that phishing awareness training is not working well enough?

A: Training is failing when employees still click, reply, or escalate suspicious messages without checking basic details first. Other warning signs include repeated responses to urgent language, weak reporting rates, and inconsistent use of verification steps for unexpected links or attachments. If people can describe threats but do not change behavior, the program is informative but not operationally effective.

Q: Should organisations prioritise awareness training or stronger email controls first?

A: They should do both, but prioritisation depends on the failure mode. If messages are getting through at scale, email controls need attention first. If users still engage with tested lures despite filtering, sustained awareness becomes the better investment. Mature programmes combine both because phishing succeeds when technical and human defences are treated as separate problems.

Technical breakdown

Why phishing benchmark scores change after training

Phish-prone Percentage, or PPP, measures how likely users are to click or interact with a phishing lure. The report’s central mechanism is behavioural conditioning: repeated exposure to realistic simulations and feedback reduces click propensity over time. That matters because the attack is not defeated by a single moment of awareness. It is weakened when users learn to slow down, verify sender intent, and report suspicious prompts before they become credential capture events. For identity teams, PPP is a proxy for the human side of authentication risk, not a complete security metric.

Practical implication: track phishing susceptibility as a living control outcome, not a one-time training completion metric.

How AI-generated phishing changes the control problem

AI-generated phishing raises the fidelity and scale of social engineering, making messages more plausible, better localised, and faster to produce. That shifts the problem from obvious spam to convincing identity deception, where the content mimics internal language, trusted brands, or urgent workflow requests. Traditional filters still matter, but the article’s point is that native email defenses are no longer enough on their own. The control gap is not only technical detection. It is the organisation’s ability to keep people alert when messages are polished enough to pass casual scrutiny.

Practical implication: pair email security with recurring training that helps users challenge believable but fraudulent requests.

Why continuous awareness works better than periodic campaigns

A one-off awareness campaign decays quickly because attacker pressure is continuous while human memory is not. The article shows that sustained training produces much lower phishing-prone percentages after months and a year, which is what you would expect when security behaviour is reinforced at the point of risk. This is less about memorising rules and more about operational habit formation. In IAM terms, it reduces the odds that a social engineering event turns into valid account access, token theft, or fraudulent approval through a user-controlled channel.

Practical implication: run awareness as an ongoing control with reinforcement, simulation, and reporting loops, not as an annual checkbox.

Threat narrative

Attacker objective: The attacker’s objective is to obtain trusted user access that can be converted into account compromise, fraud, or further intrusion.

1. Entry begins with a convincing phishing lure that imitates a trusted workflow, brand, or internal request.
2. Credential access occurs when the target clicks, submits credentials, or approves a malicious interaction.
3. Impact follows when the attacker uses the stolen access to move into accounts, data, or downstream business processes.

Breaches seen in the wild

• CoPhish OAuth phishing via Copilot Studio: Datadog showed Copilot Studio agents on a Microsoft domain can front OAuth consent phishing and forward stolen tokens; no victims reported.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Phishing is now an identity control problem, not just an email problem. The report shows that the most effective reduction in risk comes from changing user behaviour over time, which means the governance boundary sits squarely inside identity programmes. Awareness is not a soft control here, because a single click can translate into valid authentication, delegated access, or fraudulent approval. Practitioners should treat phishing resilience as part of identity assurance, not a standalone awareness metric.

Continuous reinforcement matters more than campaign volume. The article’s benchmark pattern shows steep improvement only when training persists long enough to alter behaviour. That tells us the governance failure is episodic treatment of human risk, where security teams measure completion instead of resilience. The named concept here is human trust decay gap: the lag between user caution during training and user susceptibility under real attacker pressure. Practitioners should design for decay, not for awareness events.

AI-generated phishing compresses the defender’s response time. When phishing content becomes more convincing, the cost of hesitation rises and the value of user verification increases. That shifts the security model toward layered verification, friction at high-risk actions, and faster reporting paths. The article validates that technical controls alone do not close the loop when social engineering quality improves. Practitioners should assume persuasion quality will keep rising and plan identity controls accordingly.

Regional and sector benchmarks should shape policy, not just reporting. The South America figures and the consumer-services result show that risk is not uniform across the enterprise footprint. That means training intensity, simulation design, and executive attention should follow observed susceptibility rather than generic global averages. The right operating model is risk-tiered awareness governance. Practitioners should align training cadence and escalation thresholds to measured click behaviour, not to org charts alone.

From our research library:

• The IBM/Ponemon 2025 Cost of a Data Breach Report found that phishing-initiated breaches cost an average of $4.8M each.

What this signals

Phishing resistance should be managed like any other identity assurance outcome. The useful metric is not whether employees have been trained, but whether the organisation can show sustained behavioural change in the populations most likely to be targeted. That is especially important where phishing can lead directly to account takeover, delegated access abuse, or fraudulent approvals.

Human trust remains a fragile control surface. As phishing content becomes easier to tailor and more believable, organisations need to assume that user judgement will fail under pressure unless controls add friction at the right moment. The stronger design pattern is to verify sensitive actions, not to rely on instinct alone.

For practitioners

• Adopt continuous phishing simulation Run recurring simulations that measure click, report, and credential-submission behaviour over time, then segment results by region, function, and business unit.
• Tie awareness to identity risk metrics Use phishing susceptibility as an input to identity governance and access review prioritisation, especially where user compromise would expose privileged or sensitive workflows.
• Harden high-risk approval paths Add step-up verification for payment changes, account recovery, delegated approvals, and external login events so a phished user cannot easily authorise sensitive actions.
• Improve reporting and triage loops Make suspicious-message reporting fast and visible, then route reports into SOC and mail security workflows so active campaigns are contained before they spread.

Key takeaways

• Phishing benchmarks are strongest when they measure behaviour over time, because click susceptibility can fall sharply with sustained training.
• AI-generated phishing makes social engineering more convincing, which raises the value of layered detection and user verification.
• Identity teams should treat phishing resilience as a governance control that protects authentication, approvals, and downstream access.

Key terms

• Phish-prone percentage: Phish-prone percentage measures the share of users who click or otherwise respond incorrectly during simulated phishing tests. It is a behavioural metric that helps security teams baseline susceptibility, target training, and track whether awareness efforts are improving real-world judgment over time.
• Security Awareness Training: Security Awareness Training is the practice of teaching people how to recognize and respond to security risks in daily work. It covers phishing, password hygiene, data handling, social engineering, device safety, and reporting procedures. Effective training changes behavior, supports policy compliance, and reduces human error as an attack path.
• AI-generated phishing: Phishing content created or heavily assisted by artificial intelligence to improve grammar, tone, timing, and personalisation. The goal is to make a malicious request look like ordinary business communication, reducing the visual cues people traditionally used to spot fraud.
• Human Trust Decay Gap: Human trust decay gap is the period between improved caution after training and renewed susceptibility under real-world pressure. It describes why security programmes need continuous reinforcement, because attacker success often returns as user vigilance fades between campaigns.

Deepen your knowledge

The NHI Foundation Level course covers NHI governance, human identity, and secrets management through the industry's only accredited NHI security programme. It helps practitioners connect identity controls to the real behaviours that drive compromise risk.