The document is not the deal

Every transaction I have watched up close has the same shape: a small number of real decisions, buried under a very large number of documents. The documents are not the deal. They are the residue of the deal — the paperwork left behind by decisions someone already made, often years earlier.

The trouble is that the residue is the only evidence a buyer has. So diligence becomes an exercise in reconstruction. You read a hundred contracts to answer one question about revenue concentration, and then you read them again because the second question arrived after you finished.

Software should be doing the reconstruction. Not summarising — reconstructing. The difference matters: a summary tells you what a document says, while a reconstruction tells you what the business is, and cites the documents that prove it.

That framing changes what you build. You stop building a reader and start building a model of the company that happens to be sourced from files. Questions get asked of the model, not of the pile.