11 Dec 25 
Producing creative at volume is not one clever tool, it is a stack of connected layers. Here is what actually sits under a working creative automation setup, from design tokens at the bottom to batch rendering at the top.
When people picture creative automation, they usually imagine a single magic tool that spits out a hundred ads. The reality is less glamorous and more durable: a working setup is a stack of connected layers, each doing one job, and the value comes from how they fit together rather than from any one clever piece. Understanding the stack matters because it is the difference between a fragile setup that breaks the moment your needs change and a solid one that scales. Here is what actually sits under creative automation, from the foundation up.
The Foundation: Design Tokens
At the bottom of the stack are design tokens, the atomic pieces of your brand expressed as reusable values: the exact colours, the type scales, the spacing units, the logo rules. Tokens are the foundation because they turn your brand from a look a designer applies by hand into a set of defined values a system can reference. Everything above depends on this layer, because automation can only stay on-brand if the brand is encoded as something a machine can read and apply consistently. Without tokens, every automated asset is a fresh chance to drift, which is exactly the design-system-meets-automation groundwork that makes scaling brand-safe rather than brand-diluting.
The Middle: Templates and Composition
Above tokens sit templates: the structured layouts that define how the pieces arrange, with slots for the variable content and the tokens wired in so every template is automatically on-brand. Templates are where flexibility and consistency meet, because a good template can produce endless variations while holding the brand steady, since the movable parts are defined and the fixed parts are locked. This is the layer that lets you produce a hundred variants without a hundred designers: the template carries the structure, the tokens carry the brand, and the only thing that changes per asset is the content dropped into the slots.
The Content Layer: Copy and Assets
Feeding into the templates is the content: the headlines, the images, the specific messages that vary from one asset to the next. This is where variation actually comes from, and increasingly where AI does real work, generating and adapting the copy and imagery that populate the templates. It is also the layer where human judgment matters most, because the content is what makes an asset land or fall flat, and a template full of weak content is still weak. Connecting a prompt-driven pipeline into this layer is how you feed volume in without hand-making every piece, while keeping a person on whether the content is any good.
The Top: Batch Rendering and Output
At the top is batch rendering, the step that takes the templates, the tokens, and the content and produces the finished assets at scale, in every size and format you need, all at once. This is the part that looks like magic from the outside, a hundred finished ads appearing, but it only works because everything beneath it is in place: rendering is just the final assembly of a stack that was built correctly underneath. If the tokens are solid and the templates are sound and the content is good, batch rendering turns them into volume effortlessly. If any lower layer is shaky, rendering just produces shaky assets faster.
Why Thinking in a Stack Matters
The reason to see this as a stack rather than a single tool is resilience and scale. When a need changes, you adjust the relevant layer rather than rebuilding everything: a brand refresh updates the tokens and flows upward automatically, a new campaign format is a new template, a different message is new content. Each layer can improve independently, and the whole system holds. This is also where knowing the line between generated and hand-crafted work lives, deciding which layers run automatically and which get a human hand, deliberately rather than by accident. A team that understands its stack can scale creative without losing control of it, because they know exactly which layer to reach for when something needs to change.
If you want a creative automation stack built to scale without breaking your brand, that is exactly the kind of system we build.