Key Quotes:
- “The incentives aren’t aligned, in my opinion, with liability caps. It just begs the other side… to maybe engage in some shenanigans with the data because the penalty is capped.” ~ Sterling Miller, CEO & Senior Counsel, Hilgers PLLC
- “One of the most traumatic things you can impose on your business is litigation related to trade secrets. … The best way to keep something secret is to actually not share it.” ~ Olga Mack, CEO, TermSout
- “The non-standard NDAs are not quite ready for the AI world yet. But… an AI can still be your friend when it comes to brainstorming what are the issues… flagged for you what are the risks, how to produce suggested edits, what will the counterparty respond, and how you should respond to the counterparty.” ~ Jessica Nguyen, Head of Contract Nerds & Deputy General Counsel, Docusign
The modern legal landscape is witnessing a shift where routine non-disclosure agreements (NDAs) are increasingly automated, leaving practitioners to focus on high-stakes, non-standard provisions that require sound judgment and strategic redlining and negotiation tactics.
A party’s NDA terms and how they handle its negotiation are strong signals of how the relationship will be managed post-signature.
In a recent Contract Nerds webinar on “How to Draft, Review, and Negotiate Non-Standard NDA Terms,” I sat down with a panel of experts to dissect the “landmines” of NDA negotiation. The session featured Nada Alnajafi (Founder of Contract Nerds), Olga Mack (CEO of TermScout), and Sterling Miller (CEO and Senior Counsel at Hilgers PLLC). We discussed the specific NDA “landmine” terms that often unnecessarily slow down deals and reduce deal speed and efficiency, and how to address them.
First, understanding where deals get stuck is critical for any legal team aiming for efficiency. When we polled our audience of 1,155 live attendees on which clauses most often slow an NDA from sign to signature, the results were telling:
Poll 1: When you send an NDA, what are you really optimizing for?
This data highlights a significant tension: while we strive for velocity and operational efficiency, nearly half of us are bogged down by indemnity debates and another third are bogged down by privacy and security discussions for a contract that represents a “potential” business relationship that should be a straightforward confidentiality exchange.
This makes Clausey, the Contract Nerds logo, very sad.
Below are a few key takeaways from this educational and engaging webinar. I recommend watching the full 75-minute webinar recording to dive deeper into this topic.
Takeaway #1: Skip Non-Standard NDA Clauses to Preserve Deal Velocity
In the context of routine deals, including non-standard clauses like residuals, non-solicitation, or indemnity and limitation of liability language creates unnecessary friction and business risks (of slowing down or losing a deal) rather than offering necessary legal protection. Emphasis on necessary.
Ask yourself: What are the odds your organization would bring litigation over the information shared in these discussions? For the routine deal where the same presentation is shared with thousands of prospects, my guess is the odds are low.
I noted during our discussion, “standard NDAs can be handled by the business, oftentimes, because they require very little edits or negotiation.” Olga Mack, CEO of TermScout, emphasized that mutuality is the ultimate tool for velocity, stating that “enabling your sales team to sign NDAs as they are in conversation is the most powerful way to support business and become friends with everyone there.” Sterling Miller, CEO and Senior Counsel at Hilgers Graben PLLC, often the skeptic of these complex terms, warns that “it’s hard to justify giving someone an audit right at the NDA stage,” suggesting that practitioners should focus on the context of the transaction. Nada Alnajafi, Founder of Contract Nerds, reminds professionals that “a lot of these triggers are in NDAs because someone on the other side doesn’t understand that it’s too premature to have them at this stage,” advocating for more appropriate templates like POC agreements when necessary.
42.79% of legal and contracts professionals agreed. When an NDA is shared at this early stage of the relationship, we are optimizing for getting it signed without legal review or negotiation and including these non-standard terms run contradictory to this goal (Shout out to the 34.75% of folks who quietly worry while still appearing easy to work with – we see you).
Poll 2: Which clause most often slows your NDAs from sign to signature?
Takeaway #2: Don’t Overlook Term, Termination, and Trade Secret Language
Precise language is required to distinguish between the duration the NDA is in effect, the survival of general confidentiality obligations after the NDA expires or terminates, and the indefinite protection of trade secrets (if any are shared). Many standard templates mistakenly terminate all obligations after a set period with vaguely written terms and termination language, potentially leaving sensitive information unprotected.
Miller pointed out that “there’s nothing that says, after 3 years or 5 years, confidential information loses its confidentiality,” which is why he favors long survival periods for a receiving party’s confidentiality obligations after an NDA terminates or expires. To avoid long-term liability, Mack suggests that practitioners “balance velocity with protection” by considering whether certain survival terms belong in the NDA or the main services contract.
I wondered: Does AI understand this distinction in NDA termination language? We put AI to the test and I performed a live demo of Docusign’s AI Assisted Review on a sample NDA to test whether AI is capable of understanding and catching whether the NDA has sufficient survival of confidentiality obligations post-termination per my contract playbook – spoiler: it does! AI was able to catch non-conforming language and it even suggested redlines to the language to make the NDA conforming. I encourage you to use AI as your first pass NDA review buddy but, of course, ensure you check its output.
Takeaway #3: Define “Reasonable Standard of Care” in the Age of AI
As AI tools become ubiquitous for recording, summarizing, and analyzing meetings, emails, documents, and other content, we discussed what the “reasonable standard of care” meant for the receiving party’s obligation to protect confidential information. This language is commonly found in all NDAs and it’s also commonly undefined.
If appropriate given the nature of the information shared and associated risks, disclosing parties should now consider and potentially define the standard of care and also verify with the receiving party whether (1) their data will be used to train AI models, (2) the AI tools used by the receiving party meet the disclosing party’s policies and requirements, and (3) the disclosing party’s information will reside within secure enterprise environments.
I suggested that the most effective mitigation is ensuring receiving parties use enterprise versions of AI tools that do not train on inputted data. Miller noted the educational role of the lawyer in this new era, stating that “pointing that out specifically in these early days of AI, maybe, is smart” to ensure the other side realizes that running data through a public LLM could constitute a breach of the NDA. Alnajafi viewed these clauses as conversation starters, noting that “it’s a way to catch it, to force a discussion that maybe hasn’t happened yet… because all the processes of incorporating AI… is so brand new”.
Conclusion
The future of NDA review and negotiation practice lies in moving beyond the routine reviews to master the complex, non-standard clauses that require sophisticated legal and business judgment – a task that remains with humans. By leveraging AI as a “brainstorming buddy” while maintaining essential human oversight, legal and contracts professionals can protect their organizations from modern landmines like AI-driven data leakage and ambiguous NDA termination language while also achieving business goals to scale and operate efficiently. Ultimately, balancing speed with strategic protection ensures that the NDA remains a tool for business growth rather than a bottleneck to innovation.
Audience Q&A
What do you think about liquidated damages clauses in NDAs?
- Sterling Miller: On a scale of 1-10 with 10 rated as things to avoid, Miller rates liquidated damages a “7” on his scale of things to avoid, noting that “no one will find that” acceptable in a standard negotiation.
- Jessica Nguyen: I agreed, stating, “It’s a no for me… you’re basically putting a dollar value in the information that you will be disclosing, and then you’re also telling the receiving party how much you value that information.”
Why are AI provisions necessary if we already have standard confidentiality terms?
- Sterling Miller: Specific markers are needed because “they don’t realize that if they run it through an AI tool… they’ve just breached the NDA”.
- Nada Alnajafi: Use AI provisions strategically “to force a discussion that maybe hasn’t happened yet, that nobody’s asked about.”
How do you handle European prospects who insist on GDPR language in a simple NDA?
- Jessica Nguyen: Suggest a simple compromise: “a very simple sentence or two, where each party acknowledges that limited personal data will be exchanged… but, each party acknowledges that the information shared is for a legitimate business interest under GDPR.”
Continued Learning Opportunities
- You can access the full 75-minute webinar recording for no charge, thanks to the generous sponsorship of Docusign, Intelligent Agreement Management.
- Join our next free webinar live and get CLE or CPE credit! We host one webinar a month. Follow us on LinkedIn to stay updated on upcoming webinars.
- Join 23,000+ lawyers and contracts professionals who want to master contracting skills by subscribing to our weekly newsletter.
The post Mastering Non-Standard NDA Terms: Strategic Negotiation in the Age of AI appeared first on Contract Nerds.