Contract databases require shape

Contracts are made up of multiple documents. Information sits across those multiple documents and often conflicts with or overrides information in other documents (e.g with amendments).

Simply "extracting" a bunch of fields from documents without first grouping them into the correct contract is therefore context-blind. Major errors will result. The problem is that correctly grouping documents is very hard and very resistant to AI and other forms of automation.

The same is true for individual data points across those documents. Document A might be an order form discussing the Commencement Date and Initial Term, but renewal information is in Document B. Without first recognising those two documents as belonging to the same contract, you’d never be able to calculate Renewal Dates or Renewal Notice Deadlines. The data points need to know how they should interact with each other in order to produce the information you actually care about.

Accuracy alone is not enough. We simulate the ontology of contracts. We then do the hard work of organising your documents and data points into that ontology. We even invented a programming language to make that possible.

We do all of this because any contract database without its data in the correct shape will eventually fail. We’ve also never met a customer who wants or knows how to do it themselves.