Author-it launches AION JSON format for enterprise AI

Author-it has launched AION, a structured JSON publishing format for large language models, retrieval-augmented generation pipelines and enterprise AI agents. The format is included in every Author-it Cloud deployment at no extra cost.

AION is now generally available as a standard output format across the cloud service. It sends structured content and metadata to downstream systems through HTTP Post and is intended for AI-driven tools such as chatbots and knowledge services.

Author-it, which develops a component content management system, said the new format requires no changes to its underlying product architecture. Its platform is built on a relational database, with each content component stored as a governed object carrying a persistent identifier, modification history, release state and template type.

That design allows the company to add an AI-focused output format by drawing on content structure and metadata already held in the platform. Each AION output includes content hierarchy, resolved variable values, author identity, modification timestamps, template classification, persistent object IDs, folder path and library provenance, along with the full content text.

Content problem

The launch targets a problem many companies face when trying to use internal documentation in AI systems. Key information is often stored in PDFs, word processing files and file-sharing repositories, making it difficult to preserve metadata, track provenance and apply governance when that material is passed to language models.

Author-it argues that structured content with attached metadata can improve retrieval quality in AI workflows. By making content type and governance data explicit, organisations can filter material before similarity search and trace answers back to a specific content component and author.

It added that this approach can reduce the need for AI systems to infer context from surrounding text. That may lower context window use, as well as the cost and delay involved in processing large volumes of material.

Built in

AION is available to all Author-it Cloud customers as part of the 2026.R1 release. The addition comes with no extra subscription charges, which the company described as central to the rollout.

Founded in 2000, Author-it sells content management software for content creation, review, translation and publishing across multiple channels. Its customer base includes organisations in manufacturing, software and utilities, with a focus on regulated industries where content control and auditability matter.

The company argues that those long-standing product choices are now relevant to enterprise AI adoption, particularly for businesses trying to limit inaccurate responses from chatbots and other automated systems. In those settings, the source and status of a document can be as important as the text itself.

Adrian Winks, chief executive of Author-it, linked the launch to the platform's original design. "We built Author-it on a relational database because structured, governed content was always the right way to manage content at enterprise scale. We didn't build it for AI - AI didn't exist when we started. But the architecture we chose turns out to be exactly what reliable enterprise AI requires. When our team built AION, there was nothing to rearchitect. The content intelligence was already in the database - it always has been. Every other vendor in this space is retrofitting. We're not. That's not a claim - it's a consequence of 25 years of building the right way," he said.