AI hasn’t killed the website but it has exposed weak content foundations | Tech Radar

Overview

News, deals, reviews, guides and more on the newest smartphones

News, deals, reviews, guides and more on the newest computing gadgets

Details

Start exploring exclusive deals, expert advice and more

Unlock and manage exclusive Techradar member rewards.

AI hasn’t killed the website but it has exposed weak content foundations

AI exposes weak content foundations behind websites

When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

Unlock instant access to exclusive member features.

Get full access to premium articles, exclusive features and a growing list of member rewards.

Claims that AI is making websites obsolete misunderstand where the real problem sits, because from a technical point of view AI still depends on structured sources of truth.

Websites, APIs and content platforms remain the foundations that models draw from, and what AI is exposing is not the redundancy of the web but how poorly much of it has been built.

CEO and co-founder of enterprise headless CMS Storyblok.

For years, engineering teams have worked around legacy CMS decisions that prioritized speed of publishing over clarity of structure, with content designed for pages and campaigns rather than for reuse or interpretation by machines.

Testing AI is not like testing software and most companies haven't figured that out yet

Fields were left loosely defined, taxonomies changed without oversight and meaning was inferred rather than set out explicitly, decisions typically made to meet delivery deadlines and seldom revisited once platforms were in use.

In a traditional search environment, those weaknesses were largely hidden. Pages ranked, users clicked through, and humans filled in the gaps themselves. Even outdated or poorly structured content could still perform adequately if it matched search intent closely enough. The burden of interpretation sat with the user, not the system.

AI-driven discovery removes that safety net. When models ingest content across an organization's entire digital estate, they look for consistency, context and authority across everything they can access. Weak schemas and page-centric designs turn into noise, making it harder for systems to distinguish core information from supporting material or outdated content.

Engineering teams see this quickly once content is fed into AI systems. Poorly defined fields obscure distinctions between core information and supporting content, while inconsistent labelling undermines precision once material is processed by downstream systems.

Content that once functioned well enough for a website audience starts to break down when treated as data. What was previously internal technical debt becomes visible to users, customers and partners.

Much of the current discussion around AI readiness focuses on metadata and optimization layers, but for developers this misses the point. Content usability for automated systems depends on how it is modelled, not on tagging or late-stage optimization.

Structured, modular data with explicit schemas and defined relationships provides a more stable basis, supported by predictable, versioned APIs rather than assumptions that shift between releases. Keeping meaning independent of presentation reduces the need for transformation when content is reused across sites, applications and AI systems.

CIOs don’t need more AI—they need AI that actually understands their business

What engineering teams need to change in how websites are structured and governed

The role of the CMS is shifting within many architectures. Instead of functioning solely as a publishing layer, it increasingly acts as a point of control that enforces consistency before content is distributed more widely. Content models tend to apply tighter constraints, not to restrict delivery, but to reduce variability once material is reused.

Validation, versioning and provenance are treated as system concerns rather than editorial ones, making it clearer which information is current and approved when content moves between platforms.

As reuse expands across sites, applications and services, governance tends to sit closer to engineering ownership, with fewer informal processes filling structural gaps. In practice, problems arise less from AI itself than from adding new tools onto stacks whose assumptions have not been revisited.

Relevance and discovery are increasingly handled at the data layer, with enrichment, validation and distribution managed through automated workflows. This reduces the amount of custom integration code required between systems. The gain is not simply speed, but greater predictability and lower maintenance over time.

Across modern content operations, development focus has moved away from maintaining individual systems and towards designing structured content flows. Engineering effort is increasingly directed at structure, ownership and performance, rather than one-off integrations. Consistent workflows tend to determine how well platforms perform as content is reused across channels.

The quality of the underlying structure becomes critical. Content models, ownership and workflow design shape how reliably systems operate. As websites feed into a wider set of platforms that consume and reuse content, weaknesses in those foundations become harder to contain.

We've tested and reviewed the best web hosting services.

CEO and co-founder of enterprise headless CMS Storyblok.

You must confirm your public display name before commenting

1 Invincible season 4 episode 5 just made a big change to one of the Prime Video show's new heroes — and I can't see everyone being okay with it

2'The Apple graveyard': 7 mythical Apple products from the past 50 years that never saw the light of day

3 How Apple accidentally destroyed the record business — and why I wish we'd stuck with i Pods

4 Iran’s internet shutdown proves we need to go beyond Starlink and VPNs — this could this be the solution

5'You’re holding it wrong': 11 iconic phrases that define Apple's last 50 years, from genius highs to embarrassing lows

Tech Radar is part of Future US Inc, an international media group and leading digital publisher. Visit our corporate site.

© Future US, Inc. Full 7th Floor, 130 West 42nd Street, New York, NY 10036.

Key Takeaways

• News, deals, reviews, guides and more on the newest smartphones
• News, deals, reviews, guides and more on the newest computing gadgets
• Start exploring exclusive deals, expert advice and more
• Unlock and manage exclusive Techradar member rewards

• AI hasn’t killed the website but it has exposed weak content foundations