Most lawyers get stuck on AI warranties without knowing the solution to provide a safe middle ground for both vendor and customers.

In this video, we explore one of the most important negotiation frameworks in technology contracting: moving from result-based warranties toward effort-based obligations.

Customers often ask vendors for absolute commitments in AI agreements, such as guarantees that the system will never hallucinate or generate inaccurate information.

The problem is that generative AI systems cannot realistically provide those guarantees.

This video explains how technology contracts have historically solved similar problems by focusing on risk mitigation rather than impossible outcomes. Using SaaS security obligations as a practical comparison, we examine how vendors routinely avoid guaranteeing “no security breaches,” while still committing to extensive security measures designed to reduce risk.

The same logic can be applied directly to AI.

The video walks through practical mitigation measures already being used in enterprise AI systems, including Retrieval-Augmented Generation (RAG), guardrails, prompt engineering, human review, monitoring systems, and feedback loops.

It also explains why this creates an unusual alignment between legal and R&D teams, because the same measures that improve contractual defensibility also improve product performance and customer trust.

This is where AI negotiations become much more practical and commercially realistic.

If you missed the first part of this 3-part training series - click here to watch.

Follow me on LinkedIn or join my newsletter to get updated on the next video in the series, dropping soon. 🚀