video1.0<iframe src="https://www.loom.com/embed/727a7f183c0f4ee58f2c6b4593563f6a" frameborder="0" width="1334" height="1000" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>10001334Loomhttps://www.loom.com10001334https://cdn.loom.com/sessions/thumbnails/727a7f183c0f4ee58f2c6b4593563f6a-9313caf7d1984ac9.gif500.113This Loom tours the new Zoopla AI gateway capabilities, focusing on policy templates and custom policy deployment. It shows how admins connect an AI Gateway project to GitHub to update a core template, then create providers, a Root team, and policy templates including API key authentication, a model block list (for example blocking 4.0 mini), an Akamai AI firewall configuration, and editable semantic caching (such as 3600 seconds, 0.4 tolerance). The author then creates an app called Tennis Chat, demonstrates semantic cache hits and misses in the playground, and observes request and cost usage (about 1.4 cents on two requests) along with observability and cache analytics.