Give AI a source
it can read and verify in confidence

What an AI says cannot be controlled. What it answers from can be. AI systems are already shaping how organisations are seen. Their answers cannot really be influenced; what organisations can do is give those systems a reliable source. Apparat creates that source: a structured, machine-readable record of an organisation’s official public information, drawn from the documents it has already published, and built to be found, read and checked by anyone.

The web wasn’t built for AI systems. The information AI systems use today is scattered across websites, documents, directories and databases. It can be incomplete, outdated or contradictory. It is not always clear who published it, when it was updated, or which version is current. AI systems need information that is easier to identify, interpret and verify.

A new information layer for AI. Apparat builds Data Fidelity Registers, a structured record of an organisation’s public information, drawn from the documents it has published and quoted from them word for word. Every fact is dated, linked to its source, and can be verified at any point, by anyone. The register is published on the organisation’s own domain, at an address that belongs to it, in a form AI systems can read and anyone can check.

Want to know more? Learn how it works, read the frequently asked questions, or see the DFR specification. And to start a conversation about a register, do get in touch.