The concept of a moat has long been used as a business metaphor for whatever protects you from competitors. It’s now turning up in AI discussions. Specifically, I’ve seen the moat invoked as—to quote Kyle Bahr—“the thing that protects lawyers from displacement by rapidly advancing AI.”
The suggestion is that this moat is based on “judgment.” What is judgment? For one take on that, see Lev Loukhton’s post (here).
But as usual, my focus is on what the moat is for purposes of contract drafting. I suggest that before you reach the moat of judgment, you encounter an outer moat: expertise.
Mainstream contract drafting is a mess. I’ve devoted my career to chronicling the dysfunction. AI has gorged itself on the dysfunction, so that’s what it regurgitates. For example, Claude recently trotted out for my benefit that canard, “tested” contract language—the notion that because courts have told us what traditional contract language means, it’s risky to change it. (See this blog post about that.)
So the endless salvos of AI redlines I hear about presumably use dysfunctional traditional contract language and rely on addled conventional wisdom. Changing that requires expertise. That includes expertise in how to say clearly and concisely whatever you want to say in a contract. For that, the resource is my book A Manual of Style for Contract Drafting, which is the only comprehensive guide to the building blocks of contract language.
But do enough people want “better”? Someone at an AI company told me their clients say better drafting isn’t something they want from their AI provider. That’s not surprising. Since forever, we’ve drafted contracts by copy-and-pasting, on faith, from precedent contracts and templates of questionable quality and relevance. If all you know is copy-and-paste drafting, you don’t know what “better” looks like, so you don’t think to ask for it. Still, indifference to quality is no way to run a business, let alone a profession.
Judgment is indeed relevant, but as the inner moat. You need it in spades for deciding what deal terms belong in a contract. But here too, we see the effects of copy-and-pasting. I’ve created highly customizable contract templates, most recently for Adams Contracts. (Go here for a two-minute demo video.) They feature a comprehensive decision tree that gives the user a broad range of options for deal terms, so users can exercise their judgment to full effect. But a common response from users has been, “Too many questions!” Because all we know is the modest menu of options that copy-and-paste drafting offers us, we mistake a feature for a bug.
So if quality matters, the need for expertise and judgment should prevent a human-free AI takeover of contract drafting. That would explain why the only discussion I’ve encountered of AI contract drafting involves repurposing AI-assisted contract review. Because that means starting from someone else’s draft of questionable quality, it’s just copy-and-pasting by another name.
But for a moat to function as such, what lies behind it must be worth protecting. If lawyers and clients settle for copy-and-pasting, whether of the old-fashioned kind or the AI kind, we’ll have drained the moats ourselves. If we want to bring real quality control to contracts for the first time ever, we need a human-driven overhaul, with AI playing whatever sidekick role makes sense. The moats of expertise and judgment make that the only sensible option—if we care enough about quality.
(Attribution for the photo of Bodium Castle: WyrdLight.com, CC BY-SA 3.0, via Wikimedia Commons.)
