Hey everyone, 

disclosure: I build software, not title. One of the things I keep hearing was title work is a natural fit for it, so I tried to write down an actual workflow.

It is narrow on purpose. It does not search the record and it does not examine title. It assumes the search is already done, and it uses AI to draft the review workpapers from the package: lay the conveyances into a chain and flag the breaks, match releases to their instruments, pull the exceptions and every prior-instrument reference, reconcile the legal descriptions, and build a name run sheet. Everything it produces is a draft you verify against the recorded images.

I got a few feedback already but I am still an outsider to the craft, so I am after what I still have wrong. Where does it overstate? What would it let someone skip that they should not? What is missing? 

The guide is free, nothing to sign: https://docanalyzer.ai/guides/review-title-search-package 

Grateful for any correction, and happy to answer anything.