Article At A Glance
- KnowBe4 has launched AIDA Orchestration, the eighth AI-powered agent in its AIDA suite, making it the first fully autonomous agent for human risk management.
- The new Custom SAPA AI Agent replaces generic security assessments with environment-aware evaluations tailored to your organization’s specific policies, technologies, and security stack.
- KnowBe4 is the only platform built to train both humans and AI agents — a critical distinction as AI becomes deeply embedded in business operations.
- This milestone marks ten years since the beta launch of AIDA, backed by 15+ years of user behavior data powering these intelligent agents.
- There’s a key reason why prompt engineering and social engineering are now considered the same human risk problem — and it changes how you should think about your entire security posture.
Cybersecurity training just got a major upgrade, and if your organization is still running the same generic annual phishing tests, you’re already behind. KnowBe4, widely recognized as the world’s leading human risk management platform, has officially launched AIDA Orchestration — its eighth AI-powered agent — alongside the new Custom SAPA AI Agent, marking a definitive shift in how organizations defend against modern threats.
This isn’t just another product update. It’s the result of a decade of AI development and over 15 years of behavioral data, all converging into a platform that now handles cybersecurity training with a level of precision and autonomy that wasn’t possible even two years ago. KnowBe4 now trains both humans and AI agents, addressing a gap that most security platforms haven’t even acknowledged yet.
KnowBe4 Just Changed Cybersecurity Training Forever
The launch of AIDA Orchestration as a fully autonomous agent is a significant line in the sand. Where previous AIDA agents handled specific tasks — generating phishing templates, delivering remedial training, refreshing knowledge — the Orchestration Agent ties everything together. It uses AI to automate protection measures end-to-end, reducing risk to both humans and the AI agents now operating inside your organization.
What makes this moment different from past product launches is context. AI is no longer just a tool your IT team uses. It’s embedded in workflows, making decisions, processing sensitive data, and interacting with employees. That shift created a new attack surface, and KnowBe4 built AIDA specifically to address it.
Why It Matters: Most security awareness platforms were built to train humans against phishing emails. KnowBe4’s AIDA suite was built to manage risk across an entire human-AI workforce — a fundamentally different and more complete approach.
8 AI Agents Now Live in the Market
With the launch of the Orchestration Agent, KnowBe4 now has eight specialized AIDA agents actively in market. Each one targets a specific layer of human risk, and together they create a defense system that responds faster than any manual security process could.
- Orchestration Agent — Fully autonomous risk management automation
- Remedial Training Agent — Triggers targeted training after risky behavior
- Template Generation Agent — Creates realistic phishing simulation content
- Callback Template Generation Agent — Generates vishing and callback phishing templates
- Knowledge Refresher Agent — Delivers timely micro-learning nudges
- Policy Quiz Agent — Tests employees on internal security policies
- Recommended Landing Pages Agent — Personalizes post-click training experiences
- Custom SAPA Agent — Tailors security awareness assessments to your specific environment
Why This Launch Matters for Your Organization Right Now
Every organization using a traditional security awareness program is operating with a blind spot. Generic training doesn’t account for your specific technology stack, your internal policies, or the unique ways your employees interact with AI tools. That gap is exactly where attackers find their way in.
The Custom SAPA Agent directly addresses this. By generating assessments built around your actual environment — your controls, your policies, your security stack — it produces results that are measurable, relevant, and actionable. It’s not testing employees on hypothetical scenarios. It’s testing them on the real risks inside your organization.
What Is AIDA and Why Does It Exist
AIDA stands for Artificial Intelligence Defense Agents. It is KnowBe4’s suite of AI-powered agents, each designed to detect, respond to, and reduce human risk inside organizations. AIDA operates within KnowBe4’s broader HRM+ platform, which combines security awareness training, real-time coaching, crowdsourced anti-phishing, and AI defense into a single unified system.
| Component | Function |
|---|---|
| Security Awareness Training | Educates employees on phishing, social engineering, and security best practices |
| SecurityCoach | Delivers real-time coaching based on detected risky behaviors |
| AIDA Agents | Automates risk detection, training delivery, and threat response |
| Cloud Email Security | Filters and defends against email-based threats |
The Threat That Made AIDA Necessary
Cybercriminals didn’t stop evolving when AI became mainstream — they accelerated. Phishing attacks became more convincing. Deepfakes made voice and video impersonation trivially easy. AI agents inside organizations became new targets for prompt injection attacks. The old model of once-a-year security training simply couldn’t keep pace with threats that now adapt in real time.
How AIDA Sits Inside the KnowBe4 HRM+ Platform
AIDA isn’t a standalone tool bolted onto an existing platform. It’s deeply integrated into HRM+, which means every agent has access to behavioral data, training history, policy documentation, and real-time threat intelligence. That integration is what allows the Orchestration Agent to function autonomously — it already has the context it needs to make smart decisions without waiting for human input.
The 8 AIDA Agents and What Each One Does
Each of KnowBe4’s eight AIDA agents was built to solve a specific, real-world problem in human risk management. Understanding what each one does makes it easier to see how they work together as a complete defense system.
1. Orchestration Agent: Security Training Administration From Hours to Seconds
The Orchestration Agent is the newest and most advanced of the eight. It is the first fully autonomous agent in the AIDA suite, meaning it doesn’t just recommend actions — it executes them. It automates the coordination of protection measures across both human employees and AI agents operating within your organization, compressing workflows that previously took hours of administrator time into automated responses that happen in seconds.
2. Remedial Training Agent
When an employee clicks a simulated phishing link or exhibits other risky behavior, the Remedial Training Agent automatically assigns targeted training relevant to that specific failure. It removes the manual step of an administrator reviewing logs and assigning courses, making the response immediate and personalized.
3. Template Generation Agent
Creating realistic phishing simulation content at scale is time-consuming and requires staying current with evolving attacker tactics. The Template Generation Agent uses AI to produce fresh, contextually relevant phishing templates that mirror real-world threats, keeping simulations believable and effective rather than stale and predictable.
4. Callback Template Generation Agent
Vishing — voice phishing — and callback phishing attacks have surged in recent years because they bypass email filters entirely. The Callback Template Generation Agent creates realistic pretexts and scripts for these simulations, helping organizations test whether employees would hand over credentials or sensitive information over a phone call or through a fake IT helpdesk interaction. This is one of the most underused but critical simulation types in a complete security awareness program.
5. Knowledge Refresher Agent
Security awareness fades fast. Research consistently shows that without reinforcement, employees forget most of what they learned in a training session within weeks. The Knowledge Refresher Agent solves this by delivering timely, bite-sized micro-learning nudges automatically — triggered by behavior, time elapsed, or emerging threats — so retention stays high without requiring administrators to manually schedule follow-up campaigns.
The Forgetting Curve Problem: Without reinforcement, employees can lose the majority of newly learned information within days. The Knowledge Refresher Agent is specifically designed to interrupt this decay cycle with targeted, well-timed content that keeps security top of mind throughout the year — not just during annual training windows.
The agent doesn’t just push random refreshers. It uses behavioral data from the HRM+ platform to determine which topics each employee needs to revisit most, making every nudge relevant rather than generic. That specificity is what separates this from simply sending a monthly security newsletter.
For security teams managing hundreds or thousands of employees, this level of individualized reinforcement would be operationally impossible to deliver manually. The Knowledge Refresher Agent makes it automatic and scalable without sacrificing the personalization that makes reinforcement actually work.
6. Policy Quiz Agent
Most employees have access to a company security policy document they’ve never fully read. The Policy Quiz Agent changes the dynamic by turning your organization’s actual internal policies into interactive quiz content, testing employees on the specific rules and procedures they’re expected to follow — not hypothetical scenarios pulled from a generic content library.
Example Use Case: An organization updates its acceptable use policy for AI tools. The Policy Quiz Agent ingests the updated document and automatically generates quiz questions to verify that employees have understood the new requirements — all without an administrator manually building a course from scratch.
This is particularly valuable during periods of organizational change — mergers, regulatory updates, new technology rollouts — when policy compliance gaps tend to appear fastest. The agent keeps training in sync with your actual governance documents in real time.
The outcome is a workforce that’s tested against your real policies, not someone else’s. That alignment between training content and internal standards is something generic security awareness platforms structurally cannot provide.
7. Recommended Landing Pages Agent
What happens after an employee clicks a simulated phishing link matters just as much as the simulation itself. The Recommended Landing Pages Agent personalizes the post-click training experience, serving content that’s directly relevant to the specific type of attack the employee just fell for rather than a one-size-fits-all warning page.
- Matches landing page content to the phishing technique used in the simulation
- Delivers immediate, contextual learning at the exact moment the employee is most receptive
- Reduces the chance of repeat failures by targeting the specific knowledge gap exposed
- Adapts recommendations based on the employee’s training history within HRM+
- Removes the need for administrators to manually map simulations to follow-up content
The timing here is deliberate and important. An employee who just clicked a fake invoice link is in a uniquely teachable moment — they’re aware they made a mistake and are actively paying attention. Generic landing pages waste that window. Personalized, attack-specific content converts it into a genuine behavior change opportunity.
This agent works in close coordination with the Template Generation Agent, ensuring the simulation and the follow-up experience are coherent and contextually matched end to end.
8. Custom SAPA Agent
The Custom SAPA (Security Awareness Proficiency Assessment) Agent is the most recently launched addition to the AIDA suite alongside the Orchestration Agent. It moves security awareness measurement away from generalized benchmarks and into environment-specific assessment, generating evaluations built entirely around your organization’s unique policies, technology stack, and security controls.
The Custom SAPA AI Agent Explained
Every organization has a different technology environment, a different risk profile, and different internal policies governing how employees are expected to behave. Yet most security awareness assessments ask the same questions regardless of context. The Custom SAPA Agent was built specifically to close that gap — making assessments as unique as the organizations taking them.
What SAPA Stands For and What It Replaces
SAPA stands for Security Awareness Proficiency Assessment. In its standard form, it measures how well employees understand security concepts and behaviors. The Custom SAPA Agent represents an evolution of this concept — moving from broad, generalized testing to assessments that are explicitly tailored to what your organization actually uses, enforces, and requires.
Before this agent existed, security teams either accepted generic assessment results that didn’t reflect their specific environment, or they invested significant manual effort building custom content themselves. The Custom SAPA Agent eliminates both compromises by automating the creation of environment-aware assessments without sacrificing the specificity that makes them meaningful.
How It Tailors Assessments to Your Organization’s Controls and Policies
The Custom SAPA Agent ingests information about your organization’s internal security policies, the specific technologies your employees use, and your existing security stack. From that input, it generates assessment content that tests employees against scenarios and knowledge requirements that are directly relevant to their actual working environment. If your organization uses a specific endpoint detection platform, enforces particular data handling procedures, or has recently rolled out new AI tools, those specifics are reflected in the assessment — not replaced by generic industry scenarios.
Why Generic Security Assessments Are No Longer Enough
A generic security assessment can tell you that an employee understands what phishing is. It cannot tell you whether that employee knows how to correctly handle a suspicious file according to your organization’s specific incident response procedure. That distinction is the difference between measuring general awareness and measuring actual operational readiness — and only one of those numbers tells you whether your organization is truly protected.
KnowBe4 Trains AI Agents, Not Just Humans
The most significant and underreported aspect of this launch is that KnowBe4 has extended its training and risk management framework beyond human employees to include AI agents themselves. As organizations deploy AI tools that interact with employees, process sensitive data, and make autonomous decisions, those agents become targets — and they need to be hardened against manipulation just like humans do.
What It Means to Train an AI Agent Against Social Engineering
Training an AI agent isn’t the same as training a human employee. Humans are susceptible to emotional manipulation, urgency, and authority cues. AI agents are susceptible to prompt injection — carefully crafted inputs designed to override the agent’s intended behavior and make it act in ways that serve an attacker’s goals. The attack surface is different, but the underlying principle is identical: an adversary is trying to manipulate an entity into taking an action it shouldn’t.
KnowBe4’s position is that this is fundamentally a human risk problem, just expressed in a new form. When an employee is socially engineered, they become an unwitting insider threat. When an AI agent is prompt engineered, it becomes the same — a trusted system doing an attacker’s bidding. The AIDA suite addresses both vectors under a single risk management framework rather than treating them as separate problems requiring separate tools.
The Orchestration Agent plays a central role here. By automating protection measures at the human-agent interaction level, it monitors and manages the boundary where employees and AI systems meet — the exact point where prompt injection attacks are most likely to be introduced. This is an area of enterprise security that most organizations haven’t begun to address, and it represents one of the fastest-growing attack surfaces in the modern threat landscape.
Prompt Engineering vs. Social Engineering: The Same Human Risk Problem
The framing that separates prompt engineering from social engineering is becoming increasingly outdated. Both are manipulation techniques. Both exploit trust. Both aim to make a system — human or AI — take an action it was never supposed to take. The only difference is the target. Social engineering exploits human psychology. Prompt injection exploits AI instruction-following behavior. KnowBe4’s unified approach treats them as two expressions of the same underlying vulnerability: the human risk problem.
A Decade of AI Innovation Behind This Launch
The launch of eight AIDA agents didn’t happen overnight. It’s the result of ten years of iterative AI development that started long before “AI” became a boardroom buzzword. KnowBe4 began exploring AI-driven security awareness tools during a period when most organizations were still debating whether phishing simulations were worth the effort at all.
That head start matters enormously. The behavioral data, the model refinements, and the platform integrations built over a decade give AIDA a depth of capability that a newer entrant simply cannot replicate quickly. When the Orchestration Agent makes an autonomous decision about how to respond to a detected risk, it’s drawing on years of accumulated intelligence — not just a freshly trained model operating on limited context.
AIDA’s Beta Began in 2016, 10 Years Before This Milestone
- KnowBe4 launched the beta version of AIDA in 2016, making 2026 the platform’s ten-year AI development milestone
- Eight specialized agents are now active in market, each targeting a distinct layer of human and agentic risk
- The platform has evolved from early automation experiments to fully autonomous risk management with the Orchestration Agent
- Each iteration incorporated real-world behavioral data from KnowBe4’s massive global customer base
- The 2026 launches — AIDA Orchestration and the Custom SAPA Agent — represent the most significant capability jump in the platform’s history
What started as a beta experiment in AI-assisted security training has become the most comprehensive human risk management platform available. The ten-year arc from beta to fully autonomous orchestration is not just a product story — it’s a reflection of how dramatically the threat landscape has changed and how consistently KnowBe4 has built ahead of it.
Most security awareness platforms are reactive by nature. They add AI features in response to market demand. KnowBe4 built AIDA proactively, which is why it now has eight agents in market while most competitors are still announcing their first. That difference in timeline translates directly into a difference in capability, reliability, and depth of integration.
The beta period also meant years of real-world testing at scale. The agents customers use today aren’t theoretical — they’ve been refined through actual deployment across organizations of varying sizes, industries, and risk profiles. That operational maturity is embedded in every automated decision AIDA makes.
15+ Years of User Behavior Data Powering These Agents
KnowBe4 has been collecting user behavior data from phishing simulations, training completions, policy interactions, and security incidents for over fifteen years. That dataset is one of the most extensive in the cybersecurity industry, and it’s the foundation on which every AIDA agent was built. When the Knowledge Refresher Agent decides when and what to send an employee, it’s drawing on patterns observed across millions of user interactions — not a generic algorithm.
This scale of behavioral data creates a compounding advantage. Each new simulation run, each training completion, each risky click adds to the intelligence the platform uses to make better decisions. For organizations joining KnowBe4’s platform today, they’re immediately benefiting from fifteen-plus years of collective learning that their own organization could never have generated independently.
The practical implication for security leaders is significant. You’re not deploying an AI system that needs months to learn your environment before it becomes useful. The baseline intelligence is already there, shaped by the largest behavioral dataset in human risk management. Your organization’s specific data refines and personalizes that intelligence further from day one.
Data Advantage in Practice: When the Custom SAPA Agent generates an assessment tailored to your organization’s environment, it doesn’t start from zero. It draws on over 15 years of behavioral benchmarks to contextualize your results — showing not just how your employees performed, but how that performance compares to meaningful industry and behavioral baselines built from real-world data at scale.
KnowBe4 Is the Only Platform Built for the Hybrid Human-AI Workforce
Every major analyst firm is now tracking the rise of agentic AI in enterprise environments. AI tools are being granted access to sensitive systems, communicating with employees, and making decisions that have real business consequences. Yet most security awareness platforms are still built around a model where only humans need to be trained and protected. KnowBe4 is the only platform that has formally extended its risk management framework to include AI agents as both protected entities and potential risk vectors — positioning it as the only complete solution for the workforce that actually exists in 2026, not the one that existed five years ago.
Frequently Asked Questions
Organizations evaluating KnowBe4’s AIDA suite consistently ask the same core questions: What exactly does AIDA do, how many agents are live, and what makes the Custom SAPA Agent different from what they’re already using? The answers below cut through the marketing language and address each question directly.
These FAQs are particularly relevant for security leaders, IT administrators, and risk managers who need to evaluate whether AIDA fits into an existing security program or represents a full platform transition.
What Is KnowBe4’s AIDA and What Does It Do?
AIDA — Artificial Intelligence Defense Agents — is KnowBe4’s suite of AI-powered agents built to detect, respond to, and reduce human risk inside organizations. It operates within the KnowBe4 HRM+ platform and handles tasks ranging from generating phishing simulation templates to autonomously orchestrating end-to-end risk management workflows. AIDA is designed to protect both human employees and the AI agents now operating inside modern organizations.
How Many AI Agents Does KnowBe4 Currently Have in Market?
KnowBe4 currently has eight specialized AIDA agents in market following the launch of the Orchestration Agent and the Custom SAPA Agent. The eight agents are: Orchestration, Remedial Training, Template Generation, Callback Template Generation, Knowledge Refresher, Policy Quiz, Recommended Landing Pages, and the Custom SAPA Agent.
Each agent addresses a specific layer of human and agentic risk. They are designed to work independently within their domain and collaboratively as part of the broader AIDA suite — with the Orchestration Agent serving as the coordinating layer that connects their outputs into a unified, autonomous risk management response.
What Is the Custom SAPA AI Agent?
The Custom SAPA AI Agent is KnowBe4’s newest assessment tool, delivered within the AIDA suite. SAPA stands for Security Awareness Proficiency Assessment. The Custom SAPA Agent differs from standard assessments by generating evaluation content tailored specifically to an organization’s internal security policies, technology stack, and security controls — rather than relying on generic industry scenarios that may not reflect an organization’s actual risk environment.
The agent marks a shift from generalized security awareness measurement to environment-aware proficiency testing. It’s particularly valuable for organizations with complex, highly specific security requirements — regulated industries, organizations managing proprietary AI tools, or any business that has invested in custom security policies that generic training platforms cannot account for.
Can KnowBe4 Train AI Agents as Well as Human Employees?
Yes. KnowBe4 explicitly positions its platform as one that trains and protects both humans and AI agents. As AI becomes embedded in business operations — accessing sensitive data, interacting with employees, and making autonomous decisions — those agents become targets for manipulation through prompt injection attacks. KnowBe4’s AIDA suite addresses this by managing risk at the human-agent interaction level, securing the boundary where employees and AI systems meet.
This dual focus — training humans against social engineering while protecting AI agents against prompt engineering — is what distinguishes KnowBe4 from platforms that only address one side of the equation. In a workforce where AI agents are functioning colleagues rather than background tools, this distinction is no longer theoretical. It’s a live operational security requirement for any organization that has deployed agentic AI.
How Long Has KnowBe4 Been Developing AI-Driven Security Tools?
KnowBe4 launched the beta version of AIDA in 2016, making 2026 the platform’s ten-year AI development anniversary. This timeline predates the mainstream AI adoption wave by several years, giving KnowBe4 a significant head start in terms of model refinement, behavioral data accumulation, and real-world operational testing at enterprise scale.
The platform’s AI capabilities are also supported by over 15 years of user behavior data collected across phishing simulations, training interactions, and security event responses. This dataset is one of the most extensive in the cybersecurity industry and directly powers the intelligence behind each of the eight AIDA agents currently in market.
The practical result of this timeline is a platform whose AI capabilities are operationally mature rather than experimentally promising. Organizations deploying AIDA today are not beta testers — they’re adopting a system that has been refined through a decade of real-world deployment across a global customer base spanning organizations of every size and industry vertical.
KnowBe4 AIDA Development Timeline:
Milestone Year Significance AIDA Beta Launch 2016 First AI-assisted security awareness automation Initial Agent Releases 2020–2023 Specialized agents for training, templates, and coaching Custom SAPA Agent Launch 2026 Environment-aware proficiency assessments introduced AIDA Orchestration Launch 2026 First fully autonomous human risk management agent 8 Agents In Market 2026 Complete human and agentic AI risk management suite
For security leaders evaluating long-term platform partners, this development history is a meaningful signal. AI tools built over a decade of iteration with a massive behavioral dataset behind them are fundamentally more reliable than point solutions assembled quickly in response to the current AI hype cycle.
KnowBe4 continues to develop its AI capabilities with a clear trajectory — more agents, deeper automation, and an expanding focus on the risks that come with organizations deploying their own AI agents at scale. For any organization serious about comprehensive human risk management in an AI-augmented workforce, KnowBe4’s platform is built for exactly the environment you’re operating in today.