--- title: Bringing It all Together description: Learn how Citrix helps organisations deliver secure access, protect applications, and support hybrid work through a unified platform for modern workplaces. image: https://www.cxanz.com/hubfs/AUS-CXA-080-Blog-6.jpg --- [![CXANX+Citrix LockUp Black](https://www.cxanz.com/hs-fs/hubfs/CXANX+Citrix%20LockUp%20Black.png?width=300&height=40&name=CXANX+Citrix%20LockUp%20Black.png "CXANX+Citrix LockUp Black")](https://www.cxanz.com/) [![CXANZ+Citrix LockUp Black RGB](https://www.cxanz.com/hubfs/Australia/Projects/CXANZ/AUS-HSM-CXANZ/Brand%20Assets/CXANZ+Citrix%20LockUp%20Black%20RGB.svg "CXANZ+Citrix LockUp Black RGB")](https://www.cxanz.com/) - [Home](https://www.cxanz.com) - [Partners](https://www.cxanz.com/partners) - [Customers](https://www.cxanz.com/customers) - [Blog](https://www.cxanz.com/blog) - Resources - [Find a Partner](https://www.cxanz.com/psp-partners) - [Contact Us](https://www.cxanz.com/meet-the-team) - [Home](https://www.cxanz.com) - [Partners](https://www.cxanz.com/partners) - [Customers](https://www.cxanz.com/customers) - [Blog](https://www.cxanz.com/blog) - Resources - [Find a Partner](https://www.cxanz.com/psp-partners) - [Contact Us](https://www.cxanz.com/meet-the-team) # Bringing It all Together By [**Wayne Baker**](https://www.cxanz.com/blog/author/wayne-baker) on Nov 10, 2026, 9:51:43 AM ![\Bringing It all Together\](https://www.cxanz.com/hs-fs/hubfs/AUS-CXA-080-Blog-6.jpg?width=1920&height=1080&name=AUS-CXA-080-Blog-6.jpg) # From the keyboard to the model: the complete governance picture *Six posts in, here is the whole picture in one place: the complete architecture — every layer, from the user's keyboard to the model it calls. It's meant to work either way — a reference if you've followed the series, the whole argument if you've only just arrived.* When I published the first post in this series, I asked a question that tends to produce the sort of pause in customer conversations usually reserved for budget overruns and surprise audits: **how are you governing what actually happens inside the browser today?** As I noted then, the silence is telling. What I did not fully show in that first post — because I was examining one part of the architecture rather than the entire stack — is just how far that governance question really travels. It does not stop at the browser. It follows the work all the way down to the data layer and the foundational AI model beneath it. And across many of the organisations I speak with in ANZ, the gap between where governance currently ends and where the work actually goes is not trivial. In some cases, it is large enough to have its own weather. So let’s have a look at the full picture we’ve pulled together over this blog series: ![](https://www.cxanz.com/hs-fs/hubfs/image-png-Oct-13-2026-08-44-07-2777-PM.png?width=740&height=730&name=image-png-Oct-13-2026-08-44-07-2777-PM.png) **Figure 1: End-to-end governance, from the user to the model** **Access Layer:** Citrix Workspace with Citrix SecurAccess ZTNA **Browser:** Chrome Enterprise **Experience:** Citrix Workspace **AI governance:** NetScaler AI Gateway Each layer covers risks the one above it cannot see. **Citrix SecurAccess ZTNA** controls who gets in. **Citrix SecurAccess with Chrome Enterprise** governs what happens in the browser once they are there. **NetScaler AI Gateway** governs what the application sends onward to the model — the portion that sits beyond the reach of SecurAccess with Chrome Enterprise and conventional network controls. And **Citrix Workspace** does what it has always done: it secures the work itself. **What this actually closes** The first post in this series focused on the browser as the ungoverned surface. That has not changed. The browser is still where much of the enterprise risk now sits, and most security architectures were built before it became the primary place where work happens. But the governance problem does not stop there. Nor, in most real environments, does the application estate conveniently consist only of modern web applications. There is usually a legacy layer still doing useful work in the background, much like an ageing mainframe or a senior accountant: not fashionable, but uncomfortably important. In customer conversations, I tend to frame it this way: the old security model governed the network and the device. The current model governs the network, the device, and the browser. The complete model governs all of that, plus the path between the application and the model it is calling, while still bringing legacy applications into the same operating picture. That matters, because traditional ZTNA is generally very good at controlling access to modern applications, but much less persuasive when asked to provide a coherent control plane across browser activity, AI interactions, and legacy application delivery at the same time. What the complete architecture changes is not simply the breadth of control, but the quality of accountability. For the first time, an organisation can answer the question regulators and boards are beginning to ask with increasing regularity: **what did your users and AI systems do, to what data and on whose behalf?** Just as importantly, it can do so without pretending the rest of the environment stopped existing the moment AI arrived. That answer now exists in the single architecture. **Every type of worker, one platform** The architecture above serves every worker type in the organisation, and that does not change depending on which layer you happen to be looking at. The same identity, the same policy model, and the same audit trail apply regardless of which of the six worker types is in session. Including the ones that aren't human. ![](https://www.cxanz.com/hs-fs/hubfs/image-png-Oct-13-2026-08-44-18-5591-PM.png?width=940&height=551&name=image-png-Oct-13-2026-08-44-18-5591-PM.png) **Figure 2: Six worker types, one governance model** The hybrid user is usually the largest population, and very often the one current architectures handle least well. The reason is not especially mysterious: unmanaged devices. If you are not sure which persona your architecture is failing first, start there. In most environments, that is where the gap stops being theoretical and starts becoming operational. **The thread running through all of it** Citrix has spent three decades operating as the vendor-agnostic layer between the user and whatever application they need — Windows applications, SaaS, virtual desktops, published applications — regardless of where those applications live or who built them. That is not marketing language dressed up for a conference slide. It is, quite literally, what the platform does. **NetScaler AI Gateway** extends that same principle one layer deeper. It sits in the path between the application and the model it is calling, applying consistent governance to that request regardless of model, provider, or cloud. The same organisation might be calling **Azure OpenAI** from one application, **Amazon Bedrock** from another, and an internally hosted model from a third. **NetScaler AI Gateway** applies the same prompt inspection, the same model catalogue checks, and the same audit trail across all of them. That is not a capability many security vendors can offer at this layer, because most are constrained by the boundaries of their own estate. **Purview** governs the Microsoft surface. **Bedrock Guardrails** governs what sits within AWS. The vendor-agnostic position is not a new Citrix argument hastily assembled for the age of AI. The vendor-agnostic position is precisely the Citrix position, and it has been for decades - well before AI governance became part of the enterprise vocabulary. **The series, if you've joined late** This is the sixth post in a series that started with a question about the browser. Here's the short version of where the argument went. **The browser is the new endpoint — series summary** **Blog 1** The browser has become the enterprise's most critical and least governed surface. The case for treating it as an endpoint in its own right — and why ZTNA controls the door but not the room. **Blog 2** The three gaps the current model doesn't close: visibility inside the browser, control over what happens once access is granted, and identity — [43% of SaaS access happens via personal credentials](https://go.layerxsecurity.com/hubfs/Browser_Security_Report_2025.pdf) the corporate identity stack can't see. **Blog 3** The Citrix answer: three planes working together (access, browser, experience), six worker types served from one platform, and why the hybrid user — the largest population in most organisations — is the one most architectures fail on. **Blog 4** The “day in the life of Sarah” — the questions that reframe the security discussion, the objections that come up, and why "what is the endpoint for a user on a device you don't control?" is the question that tends to change the room. **Blog 5** The layer beneath the browser: why prompt-level governance is a different problem from browser-layer DLP, what it means to govern the model-call path consistently across providers, and the four questions most organisations still can't answer. **Blog 6** This post. The complete architecture from keyboard to model, every layer named, the full governance story in one diagram. --- The question I asked in the first post, “how are you governing what actually happens inside the browser today?” turns out to be the narrow version of the one that really matters. The fuller question is this: across every AI application you run, every model it calls and every cloud those models sit in, can you account for who — or what — authenticated, from what device, what data their prompts carried, and what it all cost, as a single picture rather than a different answer from each vendor? The browser was simply where that question first became impossible to ignore. For most organisations I work with, the answer is still only partial: controlled in some places, visible in others, and full of gaps that have not yet been properly mapped. The architecture above is designed to close those gaps. Not all at once — most organisations will need to approach this in layers, starting wherever the pain is most immediate — but the complete model is there to work towards. Previous Blog [← NetScaler AI Gateway](https://www.cxanz.com/blog/netscaler-ai-gateway) [![CXANX+Citrix LockUp White](https://www.cxanz.com/hs-fs/hubfs/CXANX+Citrix%20LockUp%20White.png?width=350&height=46&name=CXANX+Citrix%20LockUp%20White.png)](https://www.cxanz.com/) Cloud Software Group and any trademarks, service marks, trade names, and logos of Cloud Software Group (including Citrix and NetScaler) are owned by Cloud Software Group, Inc. and/or its affiliates and are used by Dicker Data under license from Cloud Software Group, Inc. Dicker Data Ltd is a strategic partner of Cloud Software Group, Inc. and is not a corporate subsidiary or agent of Cloud Software Group, Inc. and/or of its affiliates. Copyright © 2026 Dicker Data Ltd. All rights reserved. ABN 95 000 969 362 - [Privacy Policy](https://www.cxanz.com/privacy-policy) - [LinkedIn](https://www.linkedin.com/company/cxanz) - [YouTube](https://www.youtube.com/@CXANZCitrix) ```json { "@context" : "https://schema.org", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "Wayne Baker", "url" : "https://www.cxanz.com/blog/author/wayne-baker" }, "dateModified" : "2026-11-09T22:51:43.911Z", "datePublished" : "2026-11-09T22:51:43.000Z", "headline" : "Bringing It all Together", "image" : [ "https://www.cxanz.com/hubfs/AUS-CXA-080-Blog-6.jpg" ], "mainEntityOfPage" : { "@id" : "https://www.cxanz.com/blog/bringing-it-all-together", "@type" : "WebPage" }, "publisher" : { "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://www.cxanz.com/hubfs/assets/logos/CXANZ-DD-AU.svg" } } } ```