Software Leading the Blind, Based on Work Product of the Blind

Yesterday I saw this tweet by the estimable @ronfriedmann:

Every so often Ron says something glass-half-full about software for contracting; ever predictable, I respond glass-half-empty. Or I throw the glass over my shoulder. This is one of those occasions.

The link in Ron’s tweet takes you to @bobambrogi‘s article about LegalSifter, a software that reviews contracts prepared by the other side to a transaction and in a minute or two offers advice on negotiation points. Currently it works only with confidentiality agreements, but apparently in the next few days it will work with any kind of contract.

Skimming Bob’s article, I found what I had expected to find:

LegalSifter has trained the software using both its own library of contracts and others that are publicly available through sites such as the SEC’s EDGAR database.

So like a lot of other software for the contracts market, LegalSifter is a house built on sand. If you train software by having it crawl through the vast garage sale that is EDGAR, you’re doing garbage in, garbage out. No thanks. If the contract-tool ecosystem is blossoming, it’s an algal-bloom sort of blossoming, at least when it comes to contract language.

Look, I get it. Figuring out what to say in a contract and how to say it clearly is usually complicated and time-consuming. So is reviewing the other guy’s draft. And my work—for example, my most recent article, here—doesn’t make wrangling contract language any easier, it just explains the nature of the complexity.

But the fix isn’t surrendering to garbage in, garbage out. Instead, for purposes of contract drafting, you build content following comprehensive guidelines for the building blocks of contract language, with the help of subject-matter experts who do more than rely on conventional wisdom. You then make that content accessible and customizable using technology; I use Contract Express. Once you have an optimal contract, using technology to compare that contract to whatever the other side sends you becomes more meaningful.

What are the prospects for such a rational system? Most companies won’t have the resources or expertise to build this sort of content. It makes sense to have a publisher or trade group build a library of customizable templates. I continue to explore the possibilities.

About the author

Ken Adams is the leading authority on how to say clearly whatever you want to say in a contract. He’s author of A Manual of Style for Contract Drafting, and he offers online and in-person training around the world. He’s also chief content officer of LegalSifter, Inc., a company that combines artificial intelligence and expertise to assist with review of contracts.

4 thoughts on “Software Leading the Blind, Based on Work Product of the Blind”

  1. A few thoughts.

    1. You seem to be complaining about a product that reviews language in “garbage” contracts because you don’t like the garbage. And, they may be garbage. But they accurately represent the existing state of agreements.

    2. You also want to take the opportunity to knock the product – but per your description, it’s not a drafting tool. Rather, it’s an interpretation tool. If it can sift through the so-called garbage language and make good suggestions on how to negotiate for what you want, why do you care that it used the EDGAR to learn to be able to do this?

    3. And, as I’ve said before, I like what you represent as far as drafting clearer contract language goes. But your language isn’t field tested and much of the language in EDGAR-based contracts has been hashed through by numerous courts. So while you don’t like it (perhaps rightfully so), it is the industry standard. It might not be crystal clear, but it doesn’t seem to be bad enough that critical mass of people are willing to make a change. Which means that, like what I said in #2, products like LegalSifter must be tuned to the current state of usage, not just what you would desire.

    • 1. Who says we have to replicate dysfunction?

      2. If a product is making suggestions based on what it finds amid the dysfunction, necessarily those suggestions risk being unreliable.

      3. We are soooooo beyond the whole “tested” thing. Gawd help us. If you were serious about this, you’d point to specific recommendations and explain why they’re risky. You’re welcome to attempt that, but it would be a rather barren undertaking.

      • 1. The product isn’t replicating anything. It’s reviewing a document and telling you where there’s opportunity. You have the choice as the human in the relationship to decide how to resolve what it finds.

        2. See point 1 above. Don’t throw out the baby with the bath water. I like electric vehicles and think that the Leaf is a joke. That doesn’t mean that the concept of electric vehicles was bad… I just had to wait for the Tesla.

        3. No, we’re not. I appreciate that this is your gig. But show me a few thousand judicial decisions based on the language clarifications that you promulgate and I’ll back down. Again, I’m not saying that you’re wrong, I’m saying that you can’t turn this ship on a dime. It’s going to take a lot of time to get where you want the world to be. Poo pooing everything and anything that doesn’t use your suggestions is, as you like to say, your right. I just don’t think it’s a healthy way to be. :)

        At the end of the day, what would’ve your response to this product been if, instead of what it currently does, they added in the statement that the recommendations it would make would be based on Adams’ A Manual of Style for Contract Drafting? Would you be cool with the product then?


Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.