A company deploys a generative AI customer-service model through Azure AI Foundry and needs the system to automatically block outputs containing hate speech or self-harm content before those responses ever reach a customer, with severity thresholds they can tune per category. Which capability should they enable on the deployment?

The correct capability is Azure AI Content Safety filters attached to the model deployment. These filters attach directly to the deployment and evaluate generated text against categories like hate speech and self-harm, blocking or flagging content according to configurable severity thresholds before it is returned to the user, which is exactly the enforceable, pre-delivery control the scenario requires. The lower temperature setting is incorrect because temperature adjusts the randomness and creativity of the model's word choices; it does not classify content into harm categories or block anything, so lowering it offers no reliable safety guarantee and may only change writing style a little. The system prompt instructing the model never to produce harmful content is incorrect because it is a soft instruction the model may still fail to follow under determined or unusual inputs; it is better than nothing, but it is not an enforced technical safeguard with adjustable severity thresholds. The Azure Monitor alerts that notify admins after harmful content is sent are incorrect because alerting after the fact means the harmful content has already reached the customer, which fails the requirement to block it before delivery. Only the dedicated content-filtering service provides the enforced, tunable, pre-response gating the company needs.