Here’s a real-world example of what I mean.
A client was preparing to sign a contract worth roughly $400,000 in year one and potentially $800,000 to $1 million in year two. The client drafted it using AI.
Did the draft include the basic business terms the client gave the AI? Yes. Price. Scope. Term. Payment. A few other key points. The “essential terms” were there.
But that’s not the same thing as having a good contract. Because the real work in contract drafting is not just writing down the deal everyone hopes will happen. It’s planning for what happens if things don’t go according to plan.
- What if performance is delayed?
- What if the scope expands?
- What if the customer refuses to pay?
- What if the other side terminates early?
- What if the deliverables are disputed?
- What if one party needs leverage before litigation?
- What if the contract is silent on the thing that later becomes the fight?
That’s where the AI draft fell short. Not because AI is useless. It isn’t. It’s incredibly useful. But AI can only respond to the questions it is asked. And a non-lawyer usually does not know which questions matter until it is too late.
That is the part clients underestimate.
A lawyer is not just typing legal language. A lawyer is issue-spotting. Stress-testing. Allocating risk. Thinking through remedies. Building leverage. Filling gaps the client does not know exist.
Which brings me to the uncomfortable question: Why would anyone bind themselves to a contract worth $1 million-plus and not spend the necessary legal fees to make sure the contract actually protects them?
AI can help generate a draft. It cannot replace judgment. And in contracts, judgment is often the difference between a document that looks fine and a document that actually works.
