A Guide to Stress-Testing Limitation of Liability Clauses with AI
A limitation of liability clause caps what each party can recover if something goes wrong under a commercial contract. It is one of the highest-impact provisions in any deal — yet its real effect only becomes visible when you walk it through specific business scenarios.
What a limitation of liability clause does
Limitation of liability clauses typically combine two mechanisms: a monetary cap (often expressed as a multiple of fees paid in the prior 12 months) and an exclusion of certain damage types such as indirect, consequential, or lost-profit damages. Together they allocate commercial risk between supplier and customer.
Common carve-outs
Most negotiated contracts leave a number of obligations outside the cap. Typical carve-outs include:
- Breach of confidentiality.
- Indemnities for third-party IP infringement.
- Gross negligence, wilful misconduct or fraud.
- Personal injury and death.
- Payment obligations.
- Breach of data-protection or security obligations.
Scenarios that reveal hidden problems
Reading the clause in isolation is not enough. The same wording can be perfectly reasonable in one scenario and catastrophic in another. Walk every contract through scenarios like:
- Insolvency of the counterparty — does the cap still leave a meaningful claim after write-off risk?
- Data breach affecting end-customers — are regulatory fines and notification costs inside or outside the cap?
- Service outage during peak season — are lost profits excluded as "indirect" damages?
- IP infringement claim from a third party — does the indemnity actually sit above the cap?
- Repeated minor breaches — is the cap per-incident or aggregate across the contract term?
How to use AI to stress-test the clause
Contract Scenarios analyses your uploaded contract against each scenario you configure and reports, in plain English, how the contract as a whole — including the limitation of liability clause and its carve-outs — would play out. This makes it easy to spot the gap between what the clause looks like on the page and what it actually does when something goes wrong.