AI Didn’t Kill the CMS. It Promoted It.

Tina Nelson, Product Strategy Director at Optimizely, explains why AI didn’t kill the CMS, but allowed it to evolve. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

Not long ago, brand websites followed a predictable playbook: build core pages, optimize for SEO, and publish. The CMS behind them served as a system of record, storing content and pushing it live, but offered little visibility into what actually drove pipeline or revenue. In talking with marketing teams, they’re doing everything right by the old playbook: more content, tighter SEO, faster publishing cycles, yet their organic visibility is still eroding. The platform isn’t the problem. The model is. What they’re optimizing for no longer matches how buyers find them.

Today, that model is breaking down as users turn to AI to surface answers directly, often without having to click through pages. According to Capital One Shopping data, more than half of shoppers now use GenAI instead of traditional search for recommendations, fundamentally changing how content is discovered and visibility is earned.

At the same time, AI is accelerating the pace of content creation, increasing both opportunity and noise. Every interaction generates signals about what’s working, but when that data is scattered across systems, teams are left asking a familiar question: What actually moved the needle?

The answer starts with the CMS—but not the one teams are used to. Traditional platforms were built to publish for human audiences. Now, content has to be interpreted, selected, and surfaced by AI systems before it ever reaches a customer.

In this environment, publishing is no longer the CMS’s primary job. Its role is to shape how content is understood, connect it to performance data, and influence what AI systems prioritize in discovery. When discovery is intermediated this way, the CMS goes beyond supporting visibility to actively determining it—with direct impact on pipeline, conversions, and revenue.

The CMS is Now a System of Influence, Not a Record

Customer journeys are increasingly starting with AI, whether through chatbots or AI-generated search overviews. Google reports that AI Overviews now reach more than 2 billion monthly users, and what was once a multi-click journey is quickly becoming a zero-click journey. In fact, only 8 percent of users who encounter these overviews click through to a website, according to Pew Research Center.

Most CMS platforms weren’t built for this reality. They were designed to publish content, not adapt it—and certainly not to influence how AI systems interpret that content. Producing more content or doubling down on traditional SEO doesn’t solve this gap. Without a way to learn from performance and act on it, teams generate volume without improving results.

To stay visible, content needs to be structured, context-rich, and continuously informed by performance signals so AI systems can extract, interpret, and surface it accurately. That requires a CMS that identifies which content resonates with specific audiences and why, so teams can turn those insights into updates that keep it relevant and easy to find over time.

But visibility is only the first step. When users reach your site, the CMS determines what happens next, shaping experiences based on engagement signals such as drop-off points, content interactions, and conversion paths. Teams can refine experiences in real-time, resulting in a more personalized journey that reduces friction and guides users toward meaningful action.

This ability to connect insight to action determines whether content is included in AI-driven discovery or ignored entirely. Without it, teams produce more without improving impact. With it, content becomes a driver of measurable business outcomes.

There is a silver lining. The traffic that does reach your site now arrives with higher intent — we’re seeing AI-referred traffic convert at 23 percent higher rates than organic traffic did before. The visitors who make it through have already been pre-qualified elsewhere. That makes what happens when they arrive even more important.

One premium consumer brand faced this challenge head-on during a redesign of its digital presence. Rather than simply migrating content to a new platform, their team rebuilt how content was structured, tested, and continuously improved. They created modular blocks of storytelling content in Optimizely’s CMS, and tagged and organized them so visitors encountered different combinations on each visit. The goal wasn’t freshness for its own sake. It was to test which content combinations drove the most engagement, then use those signals to keep experiences relevant over time.

The results reflected a CMS doing more than publishing: engaged session time increased, organic traffic improved, and conversion rates rose. Publishing was the starting point, not the finish line.

3 Capabilities a CMS Should Enable 

Having a CMS isn’t enough to compete in an AI-first discovery environment. Many platforms still treat content as something to publish and measure later, forcing teams to stitch together performance across disconnected tools.

What matters now is whether your CMS can influence what gets seen and connect that visibility directly to how buyers move from discovery to decision. To do this, it needs to support these critical capabilities.

1. Connected Insight Across the Full Buyer Journey

Most teams still evaluate content in silos: page views, engagement rates, or conversions tied to a single asset. But in isolation, those metrics don’t show whether content is being surfaced in AI-driven discovery or how it influences the broader buying journey.

A modern CMS should unify those signals into a single working environment. If a guide or landing page drives engagement and pulls users into product or conversion experiences, that full path should be visible—without exporting data into separate analytics or BI tools.

When content, experimentation, and outcomes live in one system, teams can move from fragmented reporting to a shared, actionable view of performance and adjust in real-time.

2. Built-In Experimentation

The teams that get this right don’t treat testing as a phase that happens after launch. They treat every piece of live content as a hypothesis. The shift from “publish to measure” to “publish and learn” sounds small, but it changes how a content team is organized, what they prioritize, and what the CMS needs to do.

Your CMS should flip that model by making experimentation a native part of content creation and delivery. Within the same interface used to build pages, teams should be able to test full experiences, like whether a story-driven page with embedded products outperforms a product-first layout, or how different content structures influence engagement.

The shift here is cultural as much as technical. A test-and-learn mindset prioritizes insight over assumptions, especially when results challenge those assumptions. Failed tests provide direction, helping teams learn from real performance, move on from what doesn’t gain traction, and double down on what does.

3. Continuous Optimization While Content is Live

Publishing is no longer the finish line—it’s the starting point. The value of content comes from how effectively it can be improved once it’s live.

Most systems rely on predefined personalization rules or periodic updates, which quickly fall out of sync with real user behavior. A modern CMS should allow teams to adjust content directly within the experience, refining messaging, reorganizing components in a visual editor, or updating embedded elements based on live performance.

AI accelerates that process by identifying patterns across content, such as where engagement drops or conversions stall. Its advantage isn’t just generating more content, but shortening the loop between insight and action. Teams can relaunch content in the same environment, optimizing content while it’s still driving engagement.

Content That Earns Its Place in Discovery

Many organizations struggle to improve content performance because content, testing, and data live in separate systems, each owned by a different team. As a result, it’s difficult to connect what’s created to what actually drives results.

That disconnect isn’t sustainable in an AI-driven environment, where content is evaluated and surfaced before a user ever reaches your site. Visibility is no longer guaranteed; it has to be earned continuously. This fundamentally changes the role of the CMS. Rather than serving as a static system for storing pages, it’s the system that determines how your content is understood, surfaced, and acted on.

The CMS was never just a container for content. It was always a lever for growth; most teams just didn’t have the tools to move it. Now they do. The question is whether they’ll use it before their competitors do.