Much like a mathematically-inclined rabbit, contracts are fuzzy and logical at the same time.
They are written in legalese. Legalese is much closer to natural language than the cold, hard logic of programming languages. It’s fuzzy. Fuzziness is good because it gives you a level of flexibility and expressive power (and sometimes deliberate ambiguity!) that actually makes contracts work in a commercial setting.
If you only focus on data points, codification, and AI-generated summaries and ignore the fuzz, you’re going to miss something important and it’s going to cost you. Like whether a global pandemic is covered under a force majeure clause.
On the other hand, contracts contain concrete strands of logic that weave through the swirl of fuzz. A very common example of this is the set of rules an auto-renewing contract lays out around its commencement, initial term, and renewal conditions.
If you only focus on search and “extracting” data points without also explicitly modelling the rules to calculate implications, you’re going to miss something important and it’s going to cost you. Like missing a renewal notice deadline because it didn't appear as a date in the contract.