In the world of AI-enhanced applications, trust and safety matter more than ever. Whether you're building internal tools or customer-facing experiences, ensuring that your AI doesn’t amplify offensive or harmful content is critical.
Enter Now Assist Guardian, a lightweight but powerful layer of protection built into the ServiceNow platform. In this episode of Did You Know, Willem Zeiler gives a rapid-fire walkthrough of one of its key features: the offensiveness filter.
With Guardian turned off, offensive input flows straight to the model. But once Guardian is enabled, any flagged content is either blocked or logged—depending on your setup—before it ever reaches the AI. It’s a simple switch, but it delivers serious peace of mind.
Guardian doesn't stop there. It can also:
-
Defend against prompt injection attacks
-
Block sensitive topics like workplace safety from being mishandled by AI
-
Filter personally identifiable information (PII) to protect your users and business
These filters make Guardian a must-have for any responsible AI implementation on the Now Platform. And the best part? It takes under five minutes to get started.
So if you’re building with Now Assist, make sure your content pipeline is secure. As Willem says, "It’s the simple safeguards that help maintain trust and compliance."
