Is ChatGPT good enough for contract drafting in a law firm?
Is ChatGPT good enough for contract drafting in a law firm? For brainstorming a clause idea or cleaning up a sentence, it can help. For producing a first draft your partners would send, or holding the firm’s positions across matters, it is usually not enough on its own. The useful question is not whether the prose sounds fluent, but whether the tool starts from your precedents and keeps what the firm learns.
Where a general chat assistant helps
General assistants such as ChatGPT are strong at language. They can suggest alternative wording, summarise a long mark-up, or turn rough notes into a cleaner paragraph. Many transactional lawyers already use them that way, carefully, without pasting confidential client material into a public tool.
That kind of help is real. It is also not the same as drafting from the firm’s own templates, preferred liability positions and clause history. Fluency is not the same as fit.
Where it falls short for firm work
It does not know your precedents unless you paste them in. A polished clause that ignores the firm’s standard carve-outs still needs a heavy rewrite. Pasting templates into a chat window for every matter does not scale, and it creates its own confidentiality questions.
It forgets what the firm decided last time. The fallback you accepted with a difficult counterparty often lives in one lawyer’s head or an email thread. A chat session that resets afterwards does not build a clause library the partnership owns.
It sits outside Word. Most transactional drafting and redlining still happens in Microsoft Word. A separate chat tab becomes optional. An add-in on the live document gets used on real deals.
Supervision still sits with the lawyer. Suggested wording is useful. Releasing anything to the other side without a responsible lawyer’s approval is not, whatever tool produced the text.
Fluent drafting is not the same as drafting from your firm’s own positions.
How broader legal AI platforms fit
Harvey and Legora are capable platforms that go well beyond a general chat box. They cover research and many kinds of legal work, not only contracts. Legora now markets to small and mid-sized firms and to smaller in-house teams as well as to larger organisations. For a practice that needs that breadth, those products are a serious option.
Our view at Scriptum is that on broad per-seat legal AI, heavy research and litigation use can drive a lot of the underlying model cost, while contract work is often lighter and more repetitive. That is an argument, not a measured fact. A transactional firm or in-house team whose days are mostly contracts should still ask whether it is buying a platform for every kind of legal work, or a tool sized for the work it does most.
Where Scriptum fits
Scriptum is built only for transactional work. It runs as a Microsoft Word add-in, drafts from your firm’s own precedents and templates, and captures clause choices into a library your organisation owns. It is aimed at firms of twenty to a hundred fee earners and at in-house teams doing the same kind of contract work. Pricing is published on our website, starting at £79 per user per month on the Starter plan.
If you want to see how a specialist drafting tool compares with a general assistant on a live matter, book a demo and we will walk through one with you.