Deploy It
Must customers pay for F5 Premium Support?
No, it's a $0 SKU included with each order
A top global AI leader who recently became a customer of F5 AI Guardrails
NVidia
The most common semantic attack meant to evade inherent AI safety controls, sometimes through a series of prompts
Jailbreak
What is the product which enables offensive security testing?
F5 AI Red Team
A competitive advantage for F5 AI Guardrails is that it can protect models from which vendor or vendors?
Virtually Any
What F5 buying programs are available for customers to purchase Guardrails and Red Team?
Both Term Subscription and FCP
This AI security threat arises when an attacker provides "developer input" which the model mistakes for instructions
Prompt Injection
This language is required to write a new Guardrail
English
How does a customer size the Services Package needed for their deployment?
Select the same size as the product SKU (Small, Medium, Large)
F5 AI Red Team metric for the overall security score for a model
CASI score (0-100)
F5 AI Guardrails can prevent this type of compliance risk: an LLM sharing private information it should not
Data or PII Leakage
Upcoming product to be announced at AppWorld which automates the creation of Guardrails
Remediate
This item is available to customers to automate the setup of Guardrails in a Kubernetes environment?
A Helm Chart or OpenShift Operator
Name either the consistently worst or the best rated model rated on F5 Labs, in terms of inherent security
Anthropic (best)
Grok (worst)
Customers sometimes deploy F5 AI Guardrails to comply with this European regulation
EU AI Act
This built-in guardrail is pre-configured to forbid prompts on certain subjects
Restricted Topics Scanner
Cybersecurity leaders often ask about how F5 can protect against AI threats outlined in this popular risk framework
OWASP Top 10 for LLM Applications
Unlike F5, most competitors only provide this half of the solution for AI Security
Defensive Security
Name a type of damage which could impact a customer by failing to implement AI Security?
Reputational damage, compliance penalties, data loss, intellectual property leakage, or direct financial impact
This area of the product allows customers to experiment with any prompt to see how various Guardrails would interact with actual input
Playground