CMSwire recently ran an article telling CMOs their AI problem isn’t really an AI problem. It said it’s a governance problem: organizations govern technology investments function by function, while the systems those investments create inevitably interconnect. Customers experience the combined result. Nobody’s accountable for the combination.
It’s a good article. It’s also not a new argument.
We’ve read this before
Ann Rockley and Charles Cooper made a version of this case in Managing Enterprise Content. Then, the systems in question were content management platforms rather than AI models. They made a structural diagnosis. Organizations build technology investments in silos. So, accountability stays tied to function. The resulting chaos—duplication, inconsistency, inexplicable and unfindable content—isn’t a technology failure. It’s a governance failure disguised as a technology one.
Kevin Nichols made the same case with more architectural precision in Enterprise Content Strategy: A Project Guide (disclosure: I reviewed the book before its publication).
All of this occurred years and years before anyone was framing this as an AI-era discovery.
Content operations, governance processes, and the systems that wrap around both were already described as foundational concerns. Not emerging ones. Foundational.
Scriptorium’s maturity model, built on Rahel Bailie’s work from 2011, has been telling clients the same thing for over a decade.
Most enterprise content operations sit at the lowest maturity level, fractured across marketing, technical communications, support, and training, each with its own terminology, its own tools, and no shared owner.
Organizational hierarchy not only influences content governance, but in most companies, it is the default governance model.
The gap isn’t information. It’s authority
None of this is obscure. The people writing about enterprise content governance in 2003, 2011, and 2015 weren’t wrong. And they weren’t early. They were right but largely ignored—because the argument stayed where arguments about content usually stay: adjacent to the org chart, not embedded in it.
That’s the actual gap. It’s not a knowledge gap but an authority gap. The people who understood the operating model problem were rarely the people with the authority to fix it. They wrote about the interconnection. Someone else owned the individual systems. The distance between diagnosis and authority is where the problem has lived for twenty years.
I haven’t just read about this. I spent that time building content practices inside organizations that had exactly this problem—watching a taxonomy decision nobody owned become a CMS migration nobody wanted, watching governance get treated as a rulebook to consult instead of a structure to build. The literature was right. Unfortunately, it wasn’t sitting in the room where the decisions got made.
AI didn’t create the problem. It exposed it
Here’s what’s actually changed. AI doesn’t create the interconnection problem. It did make the cost of ignoring it visible faster, and at a scale that’s harder to paper over.
A content taxonomy no one owns was survivable when the consequence was inconsistent search results. It’s a different problem when that same ungoverned structure is feeding a model that’s deciding what a customer sees, in what order, and why. These decisions now have to be explained after the fact, sometimes to a regulator.
That’s no longer hypothetical. Governments are starting to write it directly into law, because “we govern each tool separately” doesn’t survive contact with systems that touch real people’s lives at scale.
The EU AI Act’s deployer obligations are the clearest example to date. The Act doesn’t treat AI tools as isolated products each with self-contained risk. It splits obligations between the organization that builds a system and the organization that deploys it. Organizations can’t discharge their own accountability just by pointing at the vendor. They have to understand how the system behaves inside their own data and decision flows: what’s feeding it, who it affects, where the outputs go next.
That’s the same argument we’ve been making about content platforms for twenty years. The technology has changed. The failures have not. If you don’t understand how your own systems connect, you don’t actually control what they do. Eventually, someone with real authority is going to ask you to prove that you do.
The fix is not new
Architecture comes before tooling. Ownership comes before rollout. Build the model around how systems actually connect—not how the org chart says they should.
That’s not an AI governance insight. It’s a content governance insight that AI has made impossible to keep ignoring.
The operating model was always the problem. The technology just got expensive enough to make everyone finally look at it.
Cross-posted at truxell.net. Want to learn more and decide if this is the right path for your organization, schedule a free consultation.