Can AI cut humans out of contract negotiations?

“Lawyers are tired. They’re bored a lot of the time,” says Jaeger Glucina. “Having something to do the grunt work for you and get you into a position where you can focus on strategy earlier: That’s key.”

The managing director and chief of a London-based company called, founded in 2015, is an artificial intelligence company specializing in the legal industry. She was admitted as a barrister and solicitor in New Zealand before she joined Luminance in 2017.

She notes saying that legal professionals of course have very high training. However, they waste most hours in going through (contracts). For instance, one may look at a document like non-disclosure agreement and require almost one hour to read. In a firm, people can see hundreds of such pieces on daily basis.

In addition, Luminance is in readiness to roll out a fully autonomous contract negotiation system known as Luminance Autopilot. In the next month, beta-testing with select users will be initiated before moving on to release it fully in the New Year.

The company has invited me to their London site so that I can see how it operates. Two laptops are sitting on the desk in front of me. For this demo, the one on the left is for Harry Botrovick, Luminance’s General Counsel. The second one to the right depicts Connagh McCormick, general counsel of prosapient, which is indeed a (real) Luminance client. A big screen at the back wall where the laptops show the change audit trail that each party performs on the contract.

Autopilot will be used by computers for an acceptable non-disclosure agreement. NDAs define the conditions in respect of which one organization will give other organization to their secret data.

The demo begins. Therefore, Mr Borovick’s machine has an e-mail containing the NDA, which prompts it to open up using MS Word. In other words, autopilot quickly goes through the contract and start making changes. Three years are more acceptable than six years, hence changing.The governing law for the contract is changed from Russia to England and Wales.

The next tricky part in the contract puts Luminance at risk with an unlimited liability, meaning they could be on the hook for an indefinite amount if the NDA terms are violated. Ms. Glucina notes that this could be a deal-breaker for Harry’s business.

To tackle this, the AI suggests capping the liability at £1 million and softening the clause. The other party had added some “holding harmless” language, absolving them of legal responsibility in certain situations. However, the AI recognized this as a problem and removed the clause to protect Harry from that risk.

This leads to a back-and-forth where both AIs work to improve the terms for their owners. Mr. Borovick’s AI automatically emails the revised NDA, and it opens on McCormick’s machine. McCormick’s AI notices the removal of “hold harmless” and introduces a liquidated damages provision, essentially turning the £1 million maximum liability into an agreed compensation if the agreement is breached.

Upon receiving the updated contract, Mr. Borovick’s AI strikes out the liquidated damages provision and inserts language so his firm is only liable for direct losses incurred.

Finally, version four of the contract becomes acceptable to both parties. McCormick’s AI accepts all the changes and sends it to Docusign, an online service for signing contracts.

“At that point, you get to decide whether we actually want the human to sign,” explains Ms. Glucina. “And that would be literally the only part that the human would have to do. We have a mutually agreed contract that’s been entirely negotiated by AI.”