Home/Notes/Boye CMS Experts

Finding My Tribe: What the Boye CMS Experts Meeting Showed Me About AI, Taxonomies, and the Future of Content

9 September 2026 Jean A. Alexandre CMS & AI · Field Notes

Every once in a while, you step into a room where you do not need to explain why content modeling matters, why agentic AI changes enterprise search, or why information architecture is the real backbone of GenAI. Yesterday at the Boye & Company CMS Experts gathering in Boston, I felt like I found my tribe.

The Context: On September 8, 2026, Boye & Company convened its North American CMS Experts peer group at the New England Art Center in Boston's SoWa Art + Design District. Hosted by Matt Garrepy of CMS Critic, the meeting brought together eighteen CMS leaders, technologists, analysts, and enterprise buyers to examine agentic AI, structured content, and the shifting digital experience landscape.

Stepping Into the Room

There is something grounding about gathering in a creative space like the New England Art Center, surrounded by paintings and sculpture, to talk about technical architecture. Gallery owner Claudia Fiks opened the session with a reflection on how AI and 3D printing are expanding artistic frontiers, which proved to be an ideal stage for what followed.

In tech consulting, we spend a lot of time translating concepts for clients who just want their websites to work. But stepping into a room full of dedicated CMS and DXP practitioners who have lived through the evolution from monolithic CMS platforms to headless, composable, and now agentic systems felt instantly electric. You do not have to qualify your enthusiasm. You are among people who care deeply about the same foundational questions.

"Fly in the AI" and the Inversion of Digital Experience

Matt Garrepy kicked off the morning with his "Fly in the AI" presentation (originally delivered at CMS Connect). Matt laid out five key patterns he is tracking as artificial intelligence shifts from an external feature to the underlying substrate of digital experiences.

The central point that resonates most strongly: machines are becoming a primary audience for our websites. For twenty-five years, content management systems were built to render HTML for human eyes. Today, an escalating percentage of discovery, summarization, and query resolution is handled by AI agents, answer engines, and LLM crawlers.

If machines are your new audience, traditional SEO and glossy presentation layer tricks lose their efficacy. What matters instead is semantic structure, machine-readable clarity, and high-fidelity source information.

Taxonomies, Ontologies, and Library Science in the GenAI Era

One of the absolute standout sessions was delivered by Erin Wise, Associate Director of Information Management at Harvard Business School. Erin brought the formidable discipline of library science, taxonomies, and ontologies directly into the conversation around generative AI retrieval.

There is a widespread fallacy in tech right now that you can take messy, unstructured corporate documents, dump them into a vector database with naive chunking, slap a Retrieval-Augmented Generation (RAG) prompt on top, and get reliable answers. Anyone who has deployed that in production knows it eventually breaks down.

Erin demonstrated with clarity that structured taxonomies, controlled vocabularies, and formal ontologies are not legacy academic relics; they are the guardrails that make enterprise AI systems dependable. When an AI chatbot has structured metadata and relational ontology to anchor its retrieval step, hallucinations plummet and answer precision skyrockets. It was a masterclass in why good information architecture has never been more relevant.

Connecting the Dots: OKF and Cerebras Knowledge Bases

Erin's presentation opened up a lively group discussion, and I was grateful for the chance to share some of the research and implementation work we have been driving at JA Design Solutions.

Specifically, we discussed the Open Knowledge Format (OKF), an open, plain-markdown architecture that uses strict YAML frontmatter, standardized semantic types, and explicit markdown link graphs to give AI agents version-controlled, durable context. Because OKF relies on clean markdown without proprietary lock-in, agents can navigate repository-scale knowledge graphs naturally, whether running in local CLI environments or automated workflows.

I also drew parallels to Cerebras' knowledge base development and high-performance inference architectures. When you study how top-tier teams build knowledge infrastructure for ultra-fast, high-precision retrieval, the conclusion is always the same: clarity of structure and human-authored curation beat opaque black-box processing every single time.

Unfiltered Insights Over Dinner

While the presentations were exceptional, the open, candid conversations over dinner and drinks were just as valuable. I thoroughly enjoyed talking with Andrés Rodríguez Perea (MilliporeSigma), Lauren Lucibello (Mirion), Mark A. Raymond (Grove), and co-host Matt McQueeny (Earnfluencing).

We dug into the practical realities of managing digital estates at scale: what enterprise product owners actually battle day-to-day, the nuances of regional public-sector engagements, and Mark's eye-opening benchmark research revealing the steep disconnect in traditional vendor partner programs. That kind of unvarnished peer feedback cannot be replicated on webinars or sponsored conference panels.

Late-Night Sci-Fi and a Necessary Apology

Later in the evening, the conversation shifted from content graphs and answer engines to another shared passion: favorite science fiction shows and books. I had a great time talking sci-fi with Bill Rogers (ai12z), Derek Barka (SilverTech), and Matt Garrepy.

That said, I owe Matt an apology. He mentioned loving both the books and the television series. I haven't read the books yet, but I really enjoyed the show, and I told him with certainty that they were coming back for another season. Turns out I was mistaken: The Expanse is continuing its narrative through the 12-issue comic omnibus (The Expanse: Dragon Tooth by BOOM! Studios), not another season of the show. Consider the record officially set straight, Matt!

Why Communities Like Boye & Company Matter

In a world overflowing with vendor-hyped webinars and generic tech conferences, Boye & Company's peer groups stand apart. The vendor-neutral format, small room size, and rigorous practitioner focus create an environment where people can speak candidly, challenge assumptions, and learn from each other's scars.

I walked away from Boston energized, inspired, and deeply grateful to Matt Garrepy, Matt McQueeny, Janus Boye, and every attendee in the room. I will be spreading the word about Boye & Company and already look forward to future meetings.

If you build or manage content platforms and are serious about where AI, information architecture, and digital experience are heading, find your way to this table.

Thinking through AI knowledge bases or content architecture?

We help businesses, non-profits, and educational institutions build clean, structured knowledge systems and practical AI integrations that deliver accurate results without vendor lock-in.