KnowBe4 personalizes training by risk score, role, and history. It does not detect which apps each employee actually uses. Kinds does. Here is the precise difference.
Kinds Security vs KnowBe4: Personalization
The short version: Both platforms personalize, but they personalize different things. KnowBe4 tailors training and phishing to each user's risk score, role, and past behavior, driven by its AIDA agents. Kinds Security tailors them to the actual apps each user works in every day, detected automatically from Microsoft 365, Google Workspace, and Okta. KnowBe4 does not offer app detection as of July 2026. That is the one precise difference worth understanding.
Does KnowBe4 personalize phishing simulations and training?
Yes, substantially. KnowBe4's AIDA agents (Ongoing Training, Phishing, and Orchestration) pick and schedule content per user based on risk score, training history, job role, and Phish-prone Percentage. Its SmartRisk Engine analyzes 316 indicators. Per KnowBe4's knowledge base, AI-Selected templates are "chosen based on a user's training history, phishing events, and performance metrics."
The AI-Selected template model also keeps track of the last 10 emails sent to each user, a behavior KnowBe4 introduced on April 8, 2026, so people do not see the same template twice in a row. On top of that, admins can use placeholders like first name, job title, and manager name inside templates. Those fields have to be populated in the console by the admin ahead of time. This is real, mature personalization. It just runs on behavior and role, not on what software the user actually touches.
Does KnowBe4 detect which apps each employee actually uses?
No. As of July 2026, nothing in KnowBe4's documentation describes reading a given user's real SaaS footprint from Microsoft 365, Google Workspace, or Okta and assigning phishing simulations branded to those specific apps. Its personalization is risk-based, role-based, and history-based. The apps in a KnowBe4 template come from the template library, not from the user's actual stack.
In practice that means the person in your finance team who lives in QuickBooks and the developer who lives in GitHub can receive the same generic bank-alert template. It might be well chosen for their risk level. It is not chosen because of the tools they use, and real attackers increasingly do target the tools people use.
How does Kinds Security personalize?
Kinds connects to Microsoft 365, Google Workspace, or Okta over OAuth and reads which apps each person actually uses. It then auto-assigns phishing simulations branded to those exact apps. The QuickBooks user gets a fake QuickBooks notice. The GitHub user gets a fake GitHub alert. No admin mapping, no placeholder setup.
The same signal shapes training. Because assignment is automatic and per person, there is no campaign builder step where an admin guesses which template fits which group. The directory sync also adds and removes users automatically, so personalization stays current as the team changes.
How do the two approaches compare?
Kinds Security | KnowBe4 | |
|---|---|---|
Main personalization signal | The real apps each user works in, read from Microsoft 365, Google Workspace, or Okta | Risk score, role, training history, Phish-prone Percentage |
Template selection | Auto-branded to apps the user actually uses | AI-Selected from the library, based on history and performance; tracks the last 10 emails per user (since April 8, 2026) |
Name and role merge fields | Filled automatically from directory sync | Admin-populated placeholders in the console |
Risk analytics | Focused on behavior inside training and simulations | SmartRisk Engine with 316 indicators; deep and mature |
Admin work required | Connect OAuth once, then automatic | Configure groups, placeholders, and campaigns; AIDA automates selection on SAT Advanced |
Bottom line
If you want deep risk analytics across a large program, KnowBe4's SmartRisk Engine and AIDA agents are genuinely sophisticated, and its 4.6 out of 5 Gartner Peer Insights rating reflects that. If you want each employee tested against the apps they actually use, without an admin building the mapping, that capability is specific to Kinds. Try it free for 21 days, no demo required.
FAQ
Does KnowBe4 have AI-driven personalization? Yes. AIDA agents select and schedule content per user based on risk, role, and history. It is one of the platform's strengths.
Can a KnowBe4 admin recreate app-based targeting manually? Only roughly. An admin can assign app-branded templates to groups they define, but the platform will not detect each user's real apps for them.
What does Kinds need to enable app detection? An OAuth connection to Microsoft 365, Google Workspace, or Okta. Users, apps, and assignments then sync automatically.
Does personalization matter for results? Attackers impersonate the tools people already trust. A simulation branded to a tool the person uses daily is a fairer test of the real thing.
