What others are getting right about AI and content systems

Issue 23

From other smart minds

Last issue was the longest issue of Model Thinking yet. There was a lot there, and to be honest, I’m still recovering from pulling all that together.

So I thought I’d do a round up of some valuable writing by others in related spaces.

AI-generated painting of a cowboy on horseback swinging a lasso overhead. On the ground are a number of loose papers. — Rounding up some valuable writing. Cattle roundups are a spring thing, but content roundups can happen any time.

The Coming Collapse of Corporate Knowledge: How AI Is Eating Its Own Brain

Michael Iantosca writes about how human validation and input is crucial to maintaining the artificial intelligences (AIs) that are garnering all the buzz today.

The greatest threat to artificial intelligence and every AI-driven business initiative is not bias, regulation, or even adoption. It is obsolete, incorrect, and incomplete content; the decaying foundation on which most large language models and local vector stores are built.

Noz Urbina has coined a term, Truth Collapse, which Iantosca hits on: “A silent catastrophe is unfolding: the slow, systematic decay of truth inside the digital knowledge ecosystem.”

Iantosca continues:

Internal knowledge graphs degrade, models retrain on synthetic sludge, documentation ecosystems become self-referential echo chambers, engineers waste hours debugging lies, and users lose trust. We are not heading toward artificial intelligence, but toward automated ignorance.

The remedy involves content operations.

The solution begins with humility and systems thinking. Keep humans in the loop, automate verification, track provenance, filter training data, and measure content debt.

Misinformation in AI platforms is a systemic problem

Michael Andrews has been writing a series of articles on AI-related content problems, and his Oct. 17 article gets to similar points as Iantosca, but content strategists can see the opportunity that hides underneath Andrews’s article.

Most chatbots rely on text that’s scraped from sources they don’t control. OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude, and similar AI platforms take online content that was developed by others. That content was not developed with chatbots in mind. Chatbots are an unintended audience.

Online content that’s scraped by bots was intended for human readers on a specific website, who could see the context of the text’s intent and meaning.

While Andrews points the finger at how AI platforms go about getting their information, content professionals can see that they are in a precarious position without accurate, helpful chatbot content and that they can achieve a competitive advantage by structuring and crafting content geared for chatbot consumption.

Content systems vs. pattern libraries

Content design is lagging, according to Scott Pierce. “Voice and tone” documents, usage examples, templates, and links to brand guidelines aren’t enough for where content needs to be.

Content design is stuck at the pattern library phase while product design has moved to systematic thinking. And this gap is costing us.

Pierce looks further at how content design could learn from their product design peers.

Design systems evolved beyond component libraries because designers recognized a crucial truth: consistency comes from shared thinking, not identical outputs.

It’s the age of AI that shows that content design lacks systems, Pierce says.

AI is forcing content designers to articulate what we’ve always done intuitively. It’s making implicit expertise explicit. It’s revealing where our “systems” are really just vibes.

Take a look at the article to see the six pieces of what Pierce calls “a real content system.”

From Content Management to Intelligent Content

Over on LinkedIn, Marc Salvatierra shares about WebMD Ignite and suggests that its principles are principles that any content system should have.

- From tags → relationships

- From pages → knowledge objects

- From static → dynamic, personalized experiences

Controlled Vocabularies, Part I

If you’re struggling to move towards a fancy knowledge graph, Jessica Talisman writes about implementing a controlled vocabulary as your first step.

A controlled vocabulary is the simplest reliable agreement a team can make about language. It is a curated, finite list of approved terms, each with one intended meaning and a few rules for how those terms appear.

Talisman goes on to talk about how controlled vocabularies bring consistency, which pays off in content platforms and content management systems (CMSs), in product catalogs, in customer support and customer relationship management, in compliance programs, cross-company and cross-industry data exchanges, in AI implementations, and in evaluation.

The “Done” Delusion: Why Data Modeling Should Be a Program, Not a Project

Back in Issue 2, I wrote briefly about how content management isn’t a project but a program (or even a product). Joe Reis sees the same thing in data modeling, and he writes a more in-depth article about it.

In the Netherlands, data and information modeling is treated as continuous work in progress, with no specific notion of “done” or ending. He framed it perfectly: they don’t treat modeling as an initiative or a project. They treat it as a program.

The point is that modeling is ongoing.

Data modeling is more akin to a program. Data modeling isn’t a task you complete. It’s a core, perpetual business function, like finance or operations. It’s not one and one, but a living, breathing activity. This simple reframing helped me see data modeling not as a one-time setup, but as a continual activity of sense-making.

“Organizations fail to understand that content is their product.”

Managing Enterprise Content: A Unified Content Strategy by Ann Rockley

Top of mind

Things that are bouncing around in my head as I synthesize a range of ideas

A new venture

Last week, I launched a new website to validate the market for a product idea that combines my tech industry experience with my content expertise and my passion for emergency services.

I’ve served as a public information officer (PIO), I’m a certified firefighter and emergency medical technician, and before that I was an award-winning journalist. Through all those experiences, I saw a real gap.

PIOs are hardworking and care a lot, but many weren’t trained as communicators. Yet they’re suddenly expected to be their agency's voice. I'm bringing my content and tech experience to help solve that.

Next year, I hope to launch Practical PIO built specifically for those PIOs. It will be a tool to help them quickly generate effective, professional news releases.

You’re really not the audience for the product, so I’m not trying to pitch you on it. I’m simply sharing it because I’m really excited about it. I welcome your moral support. (And if you have friends or family in emergency services positions, feel free to share it with them!)

Screenshot of the Practical PIO website with a headline that says "For the voluntold public information officer: Communicate confidently — Screenshot of the Practical PIO website I launched last week.

Playoff time!

7 columns of fireworks and colored smoke shoot up from the soccer pitch inside a soccer stadium.

Major League Soccer has just wrapped up the regular season, and “my” team, Austin FC, is headed to the playoffs for the second time in its 5-year history. Granted, the bar to make MLS playoffs isn’t too high, and Austin is very much in the middle of the league rankings. Nonetheless, I’m excited for some post-season action.

And tonight is the start of MLB’s World Series! I don’t have a team in the competition, but I’m a fan of the game (who usually sides with the National League team).

I don’t know that you can beat Freddy Freeman’s walk-off grand slam from last year when it comes to chills, but if you’re at all a fan of baseball, savor the chance to watch the generational phenomenon of Shohei Ohtani and cross your fingers for one more Scherzer-Kershaw pitching matchup.

Trying to sort through the CMS market?

I launched a service to help CMS buyers get personalized, expert recommendations and a CMS implementation readiness assessment.

If you’re considering a new CMS—or are already neck-deep in one you want to optimize—I’d love to help. Head over to Choose Your CMS to get started.

John Collins

Thanks for reading!

Did someone forward you this email? Subscribe here

If you’re already a subscriber and you found value in something here, tell your friends and colleagues to subscribe!

Welcome to the 8 new subscribers who joined us since the last issue of Model Thinking.