The Future of CMS

Content management systems emerged in the mid-1990s when websites outgrew the manual handling of HTML files. They helped organizations manage publishing, permissions, and increasingly complex digital experiences.

Many evolved into digital experience platforms (DXPs), combining content management with marketing tools. Later, the headless CMS separated content from presentation to support more channels. Today, AI assistants and agents are adding another dimension: systems that can interpret content and perform tasks on behalf of editors and customers.

What does this mean for the future of CMS? Building effective customer journeys increasingly means making your information discoverable, trustworthy, and usable by both people and the AI systems assisting them.

Fragmented Channels and AI-driven Discovery

Customer journeys already span websites, apps, search engines, social platforms, and physical locations. AI-powered search and assistants add new entry points. A potential customer may encounter your organization through an AI-generated comparison or recommendation before ever visiting your website.

Your CMS therefore needs to support more than web publishing. Content may appear in search results, apps, self-service portals, videos, or podcasts. An AI agent may also visit your site to compare options, fill in a form, or help a customer make a booking.

This shift has brought answer engine optimization (AEO) and generative engine optimization (GEO) into the conversation. AEO focuses on visibility in answers; GEO focuses on visibility and representation in generative responses. The terms overlap, and neither replaces sound SEO.

For CMS teams, the practical priority is to publish useful, accessible information that answers real questions and can be accurately understood and referenced.

Read more: Enonic: A flexible platform for the future »

Building Customer Journeys for People and Agents

A CMS should help you build experiences that solve tasks for your target audience. Customers still need to understand your products, assess their options, and complete actions such as buying, subscribing, or requesting help. Increasingly, they can delegate parts of that journey to an AI agent. Your content and services need to support both direct interaction and assistance.

A resilient foundation combines reusable content, accessible interfaces, and well-defined integrations. Measurement must also evolve beyond page views and search rankings.

Google Search Console now offers dedicated generative AI visibility reports, rolled out globally in August 2026, while Bing Webmaster Tools provides AI citation insights. Use these alongside referral traffic, conversions, and task completion: appearing in an AI answer and winning a customer are different outcomes.

No platform can guarantee that it will meet every future requirement. However, the following principles provide a practical foundation for adapting to new channels, AI-assisted discovery, and agent-driven interactions:

Content First—and Ready for AI Discovery

At the foundation of your customer journey is content. Whether someone reads a guide, compares products, or asks an assistant about your services, the underlying information needs to be accurate and consistent.

Model important facts, relationships, languages, ownership, and review dates explicitly, so teams can maintain a reliable source of truth across channels.

Atomic content design supports reuse by separating content into meaningful components. This is a modeling choice, rather than a requirement to publish tiny pages for AI. For AEO/GEO, focus on original expertise, clear answers, evidence, and current information.

Keep public pages crawlable and indexable, and ensure any structured data matches their visible content. Google requires neither special AI schema nor an llms.txt file for inclusion in its AI search features, and no formatting trick guarantees a citation.

Don’t miss: Atomic content design: interview with Eirik Hafver Rønjum »

Headless CMS and Governed AI Agents

Headless delivery makes structured content available through APIs, enabling websites, apps, and other services to draw from the same source. AI applications can similarly retrieve approved content to ground their responses.

The Model Context Protocol (MCP) provides a complementary way for applications to expose tools to AI systems, such as searching a content repository or creating a draft. A content API alone does not define which actions an agent may perform.

Inside the CMS, AI is moving from isolated text generation toward assistance with multi-step editorial tasks. An agent could identify outdated content, propose revisions, prepare translations, and submit changes for review.

These workflows can be augmented with permissions, version history, audit trails, and approval rules for improved security. Headless can also support full websites, provided the solution supplies navigation, URL management, preview, and page composition that editors need.

Hybrid CMS and WebMCP

A hybrid CMS combines structured content and API delivery with website capabilities such as routing, preview, and visual page editing. This remains useful when organizations need both reusable content and editorial control over web experiences.

The next consideration is how the resulting website supports customers using agents. Accessible navigation, labeled forms, and clear action results already help both people and automated assistants.

WebMCP adds an emerging option: websites can expose explicit tools to browser-based AI agents through JavaScript or annotated HTML forms. For example, a booking site could offer tools for checking availability and preparing a reservation. Unlike server-side MCP integrations, WebMCP operates in the context of a web page.

As of October 2026, it remains a draft Community Group specification, with a Chrome origin trial available for testing. It is not a finalized W3C standard or a guarantee of AI-search visibility. Enforce authorization and validation on the server, and request user confirmation for consequential actions.

When assessing a CMS for these requirements, look beyond labels such as headless, hybrid, or AI-powered. Useful capabilities include:

• Explicit content types that editors and developers can define, validate, and reuse
• A content-oriented foundation, rather than exclusively page-oriented storage
• Rich content models with references, nested fields, localization, and clear ownership
• Searchable content and documented APIs, with permissions, versioning, and controlled access for AI tools
• Website capabilities including preview, routing, image delivery, accessible interfaces, and support for adding agent tools where useful

***

The future of CMS combines content management with AI-assisted work, AI-mediated discovery, and services that agents can use. Structured content supports reuse; AEO/GEO brings attention to visibility in answers; MCP and emerging WebMCP capabilities can support controlled actions. Each addresses a different requirement, and their value depends on the experience you need to deliver.

Begin with the foundations: maintain trustworthy content, preserve editorial accountability, and make your website accessible. Then test AI agents on bounded tasks, measure discovery and customer outcomes, and evaluate WebMCP where supported.

A CMS that helps you adapt these foundations gives you a stronger starting point as customer journeys continue to change.

First published 19 May 2021, updated 5 October 2026.