“Documentation has always mattered. In 2026, it became infrastructure.”

The State of Docs 2026 report just came out, and I was featured discussing three critical areas for product and design leaders: Purchase decisions and business impact, Docs tooling, and AI and documentation consumption.

The fundamentals haven’t changed, and that’s actually good news.

The teams winning at customer experience enablement aren’t the ones with the fanciest platforms or the most advanced AI, but the ones with solid information architecture. I’ve seen it firsthand in my consulting work. Organizations spend thousands on tooling and still struggle because their content organization is fundamentally broken. Some expect AI to fix it. It won’t.

Information architecture is the foundation everything else rests on. Many teams are building on sand. When a user lands on a page in isolation, they don’t carry context from pages they haven’t read. AI won’t either. Yet I keep seeing organizations solve this backwards: they buy better tools, become “AI-first,“ and wonder why the output still feels fragmented. The answer is simpler and harder than they expect.

The right tool is only right if your information architecture can support it. When your content is well-organized and your pages are self-contained with full context, everything else (tooling, retention, AI implementation) works exponentially better.

The unsexy work of structuring your content properly is actually the competitive advantage.

I’d highly recommend exploring GitBook’s The State of Docs 2026 report. Tal Gluck and the whole GitBook team have done a fantastic job putting this resource together and synthesizing not just expert interviews but thousands of survey responses into a cohesive, useful guide.

Borrowing from the report’s conclusions, I’ll wrap up with one of the recommendations for documentation leaders.


Invest in information architecture before AI features
Every AI-powered documentation feature depends on well-structured content underneath. Adopt an IA framework, make pages self-contained, invest in structured metadata. Teams that skip this step end up with AI features that surface bad answers — and bad answers erode trust faster than no answers at all.
Invest in information architecture before AI features
state of docs gitbook 2026 contributor
The State of Docs brings together insights from documentation experts from across the industry

Visit → The State of Docs 2026 report

Delfina Hoxha

Author

I’m Delfina Hoxha, the founder of Little Language Models, an information architecture consultancy in Vienna helping companies with thousands of users increase clarity and conversions.

I share weekly case studies, practical techniques, and everyday examples that help readers create outstanding products.

Follow to not miss upcoming weekly IA insights.