Your Brand Guidelines Are Already Fiction

Getting brand strategy to actually connect to content creation has always felt like Doc Brown trying to jam those two cables together before lightning hits the clock tower. Months of careful thinking, stakeholder alignment, and positioning work all captured in a beautiful PDF that starts becoming fiction the moment it’s saved. It’s driven me mad at times that it can be so hard to take these dozens, even hundreds of pages of documentation and keep them current as the world changes, and actually applicable at the moment someone’s making something.

Not because the thinking is wrong. Usually the thinking is sharp, reflecting months or years of positioning work, careful messaging rationale, market understanding and deliberate design choices. Executives going to the mountain top to align on a vision and transmute that into cascading waterfalls of information that would align the entire org and capture market share in a net made of magic. So, the thinking itself is usually pretty, pretty, pretty good.

The problem is the container. Static brand guidelines documentation is a photograph of a decision. It captures what someone intended at a point in time, but it has no mechanism to surface and enforce that intent at the moment of creation.

The craftspeople aren’t the problem

Skilled Art Directors and Designers can interpret and even extend these guidelines in the context of doing their work, and much of the visual content work that brands need done requires their conceptual thinking and design process. Strong copywriters and editors can nail the voice and tone, and align content to key pillars and weave in brand-approved proof points. Over time, the right people can meld together in such a trusted way that the brand essentially becomes whatever they’re expressing and creating. And that’s a beautiful thing when that happens, especially if it works. So, let’s agree that actual masters of these crafts can be trusted to properly interpret the brand guidelines and apply them to their work, with or without the use of generative AI.

They’re not the problem. It’s the rest of us, those who simply can’t be trusted to marry the visual and verbal identity to communicate ideas. In a moment where most in-house design teams are underfunded and being told to do more with less, that means many folks in other marketing functions who aren’t the aforementioned craftmasters are either having to backburner ideas that they want to put out to the market or just going for it and creating abominations that ultimately embarrass the brand team.

Ever had a scenario where the demand gen team needed to iterate and get fresh creative out the same day, and the brand design team was slammed on the final push to get everything ready for the user conference? And, then, the demand gen team did their best in MS Paint or whatever but ran some ads the CEO saw, who then asked the brand team why they were putting those terrible ads out in paid social? That system’s not fair to anyone.

The moment of creation is everywhere now

The moment of creation is now happening everywhere, constantly, often with a human barely in the loop if at all. And beyond some hex codes if you’re lucky, most folks don’t know how to guide generative AI tools in staying on-brand.

Marketing teams with AI-curious folks are generating content across dozens of tools. And they’re exploring every format under the sun, often chaining together multiple tools for a single content piece. Every one of those tools will happily use your brand’s sparingly-used orange as a background fill, slap a second typeface on a headline, or deploy your accent color in a context that contradicts the intent behind it. The tool doesn’t know the intent. It was never given the intent. It was given a hex code.

Many of us have long sought ways to solve the underlying challenge of marrying the messaging house with the visual system—hello DAM, hello PPT template factory—even built full-on brand design software, and it’s been just plain tough for human reasons or technical limitations or both to create and maintain a living, breathing, updatable source of truth for both, in one place, that any individual could actually apply in their work.

What governance metadata looks like

What actually changes the game is when the source of truth isn’t just what the value is, but how it should be used, where it should never appear, and what intent it serves. The brand constraints travel with the value. Here’s a simple example—a semantic color token that describes intent, not just a raw value:

"action.primary": {
  "$value":   "{color.primitive.orange}",
  "type":     "color",
  "use":      "CTAs, interactive elements, linked text, one highlighted phrase per section.",
  "never":    "Background fill. Error states. More than one dominant instance per section. Full sentences of text.",
  "text-use": "Linked text and single highlighted phrases only — never full paragraphs or body copy."}

Compare that to the typical brand guideline: “Use Brand Orange sparingly as an accent color.” Again, a good designer knows what to do with that. And maybe the non-designers have great templates for most things. That’s still absolutely important.

But the contrast matters. Same intent. Radically different in terms of usefulness. The token version can be consumed by an AI generation tool, a design system component, a QA script, or a human reviewer, and the constraints are clear in every case. The PDF version requires a human to interpret “sparingly,” remember the context, and catch brand drift after it’s already been produced.

This is the layer many brand teams are still missing. A machine-readable governance layer that sits between the brand decisions and the plethora of tools generating brand content.

Compiled Brand Truth

The approach we like for our own brand and clients is portable, made up of specifications and structured files—markdown, JSON, and a few well-organized databases—that can easily be handed off and dynamically updated to always reflect the current messaging and visual identity for use in most tools and formats.

Think of it as a compiled Brand Truth. A single reference that includes both the visual identity manifest—design tokens with governance metadata, compliance checks, pattern libraries—and the verbal architecture: narrative elements, application rules, messaging hierarchies with deployment conditions. Prior to compilation they’re both alive as independent, updatable governing sources of truth for their respective domains, and importantly, they cross-reference each other. When the visual identity changes, the verbal layer knows. When a key message gets updated or a key proof point is contested, and those changes are approved by whoever is in charge of such things for your brand, compile the changes and push to your favorite shared file destination for folks and machines to grab from when firing up the content engine.

I don’t think this is the only way to solve the problem. But the problem is real: brand identity stored in static, split-brain containers that aren’t referenced while AI tools generate at scale.

Interface agnostic by design

As things evolve and the best tool for each job tends to shift each week, it makes sense to have, as part of your strategy, an interface agnostic approach to on-brand generative AI. The compiled brand system should be portable, not locked into one system.

I think it’s pretty cool to think about individuals on teams being better equipped to tell a well-informed story and communicate it visually while staying on brand. Further, to have a reliable way to update that governed source of truth when what was true yesterday gets disrupted by better robots, zombies, or inflation. And, to actually then apply it to whatever anyone on the team is making.

I’d love to connect with other folks who are working on this stuff and share ideas. That’s the premise of this series. Over four posts, we’ll walk through the structural problem, the architectural solution, the proof that it works, and a practical blueprint for building it with tools you already have.