Contentful has answered the question marketing teams keep asking about agents, and the answer is to stop asking it. In How to prepare your website for AI agents without rebuilding everything, the post rejects the worry that agents make the website obsolete: “The agentic shift doesn’t make your website less important. It raises the standard for the content layer behind it.” The distinction it draws is useful: generative AI helps you understand, automation executes predefined steps, and an agentic system takes an outcome and decides how to get there. “The shorthand we keep coming back to is the move from ‘help me understand’ to ‘help me get it done.’”
There are no figures, and the argument rests on one sentence: “your content layer is your context for agents.” People need story and reasons to trust; agents need “accurate, structured, machine-readable information they can query and reason over.” The instinct to build a second thing for the new audience “is expensive and wrong: two sources of truth become two versions of the truth.” The sharpest example is about stale data. “A person will squint at two different prices and work out which is probably right. An agent won’t. It will either pick wrong or decide your source can’t be trusted.” The advice is to start with one customer journey, then check that content is structured and composable, then do the hygiene work: content modeling, Schema.org JSON-LD, a deliberate robots.txt and semantic markup. Governance, it says, “is the dimension we see teams underestimate most.”
The catalog reads the content layer the post is selling. The Contentful provider page lists 8 API pages, and an agent reading content would go through the Contentful GraphQL Content API or the Contentful Preview API, while the marketing workflows the post says agents are already changing write through the Contentful Content Management API, with Contentful Webhooks carrying the changes out. On the Agent Readiness score, the MCP server, agent skills, the event surface and the full identity layer are lit, delegated identity, protected resource metadata and dynamic client registration.
The Kin Score is 44.1, developing band, with discoverability at 56.7 and contract quality at 52.9 against developer ergonomics at 39.8 and contract governance at 0.0. The Agent Readiness score is 59.9, agent-ready, well above the composite. Error semantics and OpenAPI examples are unlit. The post’s governance list asks for “visibility into where AI has made changes on your behalf, and a human in the loop where the stakes justify it.” That is the right list, and the agent layer on Contentful’s record is better built than most to support it. The contract-governance facet at zero is the other half of the post’s own argument: the content layer is the context, and the description of the API that serves it still needs the same care.