When the CMS becomes context, everything changes

Last month we wrote about discovery moving from keywords to context. That shift runs deeper than the front door: a structural shift from managing content to managing context across the whole experience. 

The market is moving swiftly in this direction. Contentful has been folded into Salesforce; Contentstack now talks about a "context economy", Kontent.ai about an "agentic CMS", and Sanity calls itself a content operating system. Even the consumer tools have got there: ChatGPT's new memory, rolled out on 4 June, now holds a living picture of each user rather than a saved list of facts. 

The content management system is becoming a thin layer on top of something more important: how well a brand has organised what it knows. But the platform pitches skip the interesting part. Context can be engineered to make you sound more like everyone else, or more like yourself. As fuel for faster personalisation, it deepens the sea of sameness; as a carrier for your knowledge, voice and authority, it becomes the thing that sets you apart. 

Organise what you know before you automate what you do.

The instinct will be to point an agent at your existing content and let it run, and we would resist that. The brands that pull ahead first do the less glamorous work of structuring what they actually know: the facts, how they relate, and where the gaps and contradictions sit. Then anything built on top has something coherent to draw on, and your knowledge becomes an asset to engineer rather than a library to search. 

Your voice is part of the context, not a layer on top.

Tone, point of view and authority are not decoration added at the end; they travel inside the information itself. When you hand an AI your brand to work with, the way you have written your guidance teaches it how to sound as much as the rules you set. Engineer expression in deliberately, or the output will quietly smooth you towards the average. 

Build context, don't just retrieve it.

There is a real difference between a brand that has organised what it knows and one that lets an AI rummage through the filing cabinet and hope. The first stays consistent; the second falls apart the moment two sources disagree. As these systems begin holding more context about your customers, being able to see and govern what they draw on becomes part of the build, not an afterthought. 

This is the first in a short series on what the move from content to context asks of a business: the context itself, the workflow around it, the interface agents reach it through, and the outcomes it should serve.