Skip to content

What to Use for Drafting Contracts: Copy-and-Pasting, GenAI, or Document Automation?

27 April 2025

ScrivenerFINALI've noticed people wondering on LinkedIn how law-firm lawyers stay relevant when potential clients have access to GenAI. I can't work up much enthusiasm for mulling that over—if someone who isn't a deal lawyer is willing to try their hand at contract drafting (of whatever sort), they probably wouldn't be much of a client.

So let's broaden the inquiry. Competent contract drafting requires that you be capable in three areas:

  • Being aware of the universe of deal points for a given kind of transaction 
  • Choosing deal points for a transaction of that sort
  • Coming up with verbiage to express that transaction

Let's consider how these three capabilities play out with three methodologies: traditional copy-and-pasting, generative artificial intelligence (GenAI), and document automation.

Copy-and-Pasting

Ideally, a drafter would be aware of a broad universe of deal points for a given kind of transaction. But in the traditional copy-and-paste world, a drafter's choices tend to be constrained by what's on offer in the one or more contracts that the drafter has elected to copy-and-paste from, or has been presented with. Escaping those constraints would require real work.

As regards coming up with contract verbiage, the whole point of copy-and-pasting is that you don't have to do that—instead, you copy-and-paste it! Unfortunately, one consequence of generation after generation copy-and-pasting, on faith, from precedent contracts and templates of questionable quality and relevance is that whatever verbiage you're copy-and-pasting is overwhelmingly likely to be dysfunctional, in terms of what it says and how it says it. Browse what's on Adams on Contract Drafting for copious evidence to that effect. The dysfunction is aggravated by addled conventional wisdom the legalistic mind dreams up to rationalize the dysfunction.

And although copy-and-pasting spares you drafting from scratch, copy-and-pasting from different contracts generally involves making plenty of adjustments. That's why creating a set of copy-and-paste deal documents has been a reliable source of billable hours for law firms.

So generally, you can expect traditional copy-and-pasting to offer a limited range of deal terms, generate dysfunctional text, and, if a law firm is doing the work, result in clients incurring real legal bills.

GenAI

Now we're in the glorious age of generative artificial intelligence, with the machines on the brink of taking over. Or so the techbro grifters and hype huffers tell us. What are the implications for contract drafting?

Well, whatever flavor of GenAI you use, you're going to have to tell it what you want, instead of just running with whatever the copy-and-paste gods give you. So in theory, GenAI could help return to contract drafting a sense of agency.

But I don't see that working out well, for three reasons. First, I suspect that many would find it hard to shrug off the passivity that comes with years of copy-and-pasting. Anything other than the shortest contract would require dozens of decisions; that might challenge even the most ardent "prompt engineer".

Second, it's routine for copy-and-pasters to work repeatedly with selected precedent contracts they've become familiar with. By contrast, whatever GenAI elects to throw at you would be an unknown quantity, so it would take time to vet it, assuming you're capable of doing so.

And third, GenAI has been trained on the dysfunction out there, so it can only offer users a grand souped-up copy-and-paste version of that dysfunction. (For more on that, see this 2024 blog post.)

These shortcomings perhaps explain why contract drafting appears to lag way behind contract review as the favorite GenAI contract-related task. For example, the legal tasks featured in the recent Vals Legal AI Report include redlining but don't include drafting.

Document Automation

What about document automation, also known as document assembly? With document automation, you present the user with an interview that allows them to consult guidance and select from among deal-term options. After completing the interview, the user downloads a contract that reflects the choices made.

But if you want to build your own document-automation system, the odds are against you. Document automation has been around for decades, but it has languished. That's because document automation is hard. (See this 2023 blog post for more about that.)

 The only way to make document automation more accessible for those who want to create contracts would be for a vendor to build a library of highly customizable document-automation templates. That's what Adams Contracts (and only Adams Contracts) is doing.

Instead of offering you a limited range of deal points (copy-and-paste) or relying on users to come up with their own deal points (GenAI), the interview for each Adams Contract template offers users a much broader range of deal-point choices—what I call the "decision tree". It also offers guidance, to help users make informed decisions. If you express interest in a given option, the dynamic interview might reshuffle to allow you to explore additional permutations of that option.

As for the verbiage, the contract you create using an Adams Contracts template would express the transaction you opted for, and it would express it clearly and concisely—it would comply with the guidelines in A Manual of Style for Contract Drafting. And it would incorporate the input of specialists under my editorial control.

So Adams Contracts turns into commodity tasks the first and third activities described at the top of this post—being aware of the universe of deal points and coming up with verbiage to express the transaction. It leaves to the user the task of figuring out what deal terms are appropriate for the transaction. Here's what this 2025 blog post says about that:

Our decision tree is a feature, not a bug. There’s the notion out there that you should keep as short as possible the time an organization spends drafting contracts. (See for example this LinkedIn post by Thomson Reuters.) For coming up with the verbiage, I say amen to that. But deciding the form of a deal is the highest function of those doing deals, and usually it takes time to think through those issues. Would you really rather reuse decisions made by others, for other deals? Adams Contracts templates are intended for those who want to make a fresh start. A fresh start requires a full range of choices.

How is it that Adams Contracts is so much better than copy-and-pasting and GenAI? Whereas copy-and-pasting and GenAI are erratic processes that serve up who knows what, Adams Contracts offers a product where every component is controlled. We're able to do that because I have the editorial chops, and the stomach, for the task. And LegalSifter (Adams Contracts is a division of LegalSifter) has been willing to invest in the resources required.

If you truly want high-quality contracts instead of paying lip service to quality, if you want contracts to address your needs instead of, say, expressing a cookie-cutter transaction, if you want them them fast and at a bargain price, and if you can cope with change, using Adams Contracts templates is your only option. Of course, currently your options are limited to our confidentiality agreement template, our boilerplate template, and our service agreement template. But one sure way of getting additional templates is to subscribe for what's currently on offer.

(To reflect that copy-and-pasting and GenAI are two ways of creating contracts that exhibit the dysfunction of traditional contract language, I've used for this post the illustration I commissioned from Russell Christian. I call it "The Traditionalist." You can get it on a T-shirt if you visit my store. All proceeds go to the Belmont Child Care Association.)