# Blocked Word Action: decide what happens when a forbidden term appears

> Blocked Word Action turns detection into enforcement, giving each tenant a predictable, policy-aligned response when a forbidden term appears.

**Category:** Prompt Injection
**Author:** NeuralSeek Team · **Published:** June 9, 2026
**Canonical:** https://neuralseek.ai/ai-grounded/blocked-word-action
**Section index:** https://neuralseek.ai/ai-grounded

Blocked Word Action is one of NeuralSeek's Prompt Injection guardrails — part of the platform's 118 individually configurable, fully auditable controls. In regulated, high-volume AI, the difference between a system you can trust and one you merely hope works comes down to specific, tunable controls exactly like this one. Here is what Blocked Word Action does, why it matters to the business, and how to set it for your own environment.

## What it actually does

This control defines the action taken when a blocked word is detected — whether the request is refused, the term is removed, or another response is triggered. It connects detection to a concrete, predictable consequence.

## Why business teams care

Detection without a defined action is just an alert; the action is what actually protects the experience. Making it configurable lets each tenant align the consequence with its own policy.

## How to tune it in practice

Choose a strict action for terms that are never acceptable and a softer one for borderline vocabulary. Keep the action consistent with the rest of your content policy so behavior stays predictable.

## Common failure modes it prevents

Adversaries probe production AI constantly, hiding instructions in user input, pasted text, and retrieved documents to hijack how the model behaves. Blocked Word Action closes that gap directly. By making the behavior an explicit, enforced control rather than something left to chance, it converts a latent risk into a managed, observable event — one that surfaces in the audit trail instead of in a customer complaint or a compliance finding.

## Where it fits in the stack

It runs early in the request pipeline, screening input before it can reach the model or contaminate retrieval. Because it lives in NeuralSeek's governance layer rather than inside any single model, the control holds identically whether a request routes to OpenAI, Anthropic, Gemini, Llama, Mistral, IBM watsonx, or an in-house model.

## Defense that updates with the threat

Attack patterns evolve, so this control is curated from real production attacks and refreshed as new techniques emerge — the protection you ship with keeps pace with the adversaries you actually face.

> A blocklist is only as useful as the action you attach to it.

## The takeaway

Blocked Word Action turns detection into enforcement, giving each tenant a predictable, policy-aligned response when a forbidden term appears.

---

From NeuralSeek's AI Grounded — practical, web-verified guidance on building governed, grounded enterprise AI. NeuralSeek is the model-agnostic, governed AI platform you own: any LLM (swap with no rebuild), your data in your own tenant (cloud or on-prem), 118 guardrails enforced before any action, one container that runs anywhere.