TL;DR
- Procurement teams renewing software contracts often miss critical AI-related changes to liability, warranties, and service levels, treating AI as "just another feature" rather than a contract risk.
- Providers are already updating terms without waiting for renewals, shifting fixed language to vague references like "bundled features" that can change without procurement's knowledge.
- Standard termination clauses (data return/deletion) weren't built for machine learning; it's unclear whether AI models can actually "un-learn" or fully purge a client's data.
- Legal teams tend to focus narrowly on protecting proprietary data from AI absorption, often overlooking bigger issues like liability allocation and SLAs for AI-generated outputs.
- AI isn't the enemy: contract analytics is a genuine win, helping teams flag non-standard terms, track renewal timelines, and spot compliance gaps (e.g., CCPA) across large contract portfolios at scale.
- Bottom line: procurement's job is to read the fine print now, negotiate the terms that matter, and make sure the contract has answers when problems surface later.
Most procurement teams are moving really fast right now. They’re moving as quickly as they can to adopt AI tools, get contracts renewed, and get to the next thing on their ever-expanding to-do list. Amid all that rush, it’s very likely something important that underpins everything procurement is doing is getting missed: the contracts governing the software that procurement teams use every day are changing quietly, and not always in ways that have been reviewed, negotiated, or even noticed.
Wendy Storino, a procurement leader with nearly two decades of experience in technology sourcing and supplier strategy, joined a recent episode of The Sourcing Hero podcast to walk through what is happening, why it matters, and what procurement needs to do about it.
The Gap Between Features and Contracts
The core problem, as Wendy sees it, is that most companies are still treating AI as though it were just another software feature. A provider adds AI capabilities to their platform, procurement renews the contract with a generalized addendum, and everyone moves on. The questions that should be asked at that moment about things like liability, warranty, service levels, data handling, and the accuracy of AI outputs, often go completely unnoticed.
“Procurement will come to a renewal and they'll see there's AI in there. And they didn't even consider, does the limitation of liability change? Does the warranty have to be modified? What happens with those outputs that the AI is doing?”
— Wendy Storino
The honest answer is that most practitioners know something needs attention. They just do not have the bandwidth to do anything more than get a standard addendum in place and move on. Unfortunately, standard addendums were not written with AI in mind, and the gaps getting papered over today may become expensive problems tomorrow.
How Contracts Are Already Changing Without You
Wendy also points out that many providers are not waiting for renewal conversations to update their terms. She reviewed contracts from a number of large software providers, including names that procurement teams deal with regularly, and found a consistent pattern. Language that once committed to fixed, agreed-upon terms had quietly shifted to references like “bundled features” or terms “listed on our website” that could be updated from time to time without procurement’s knowledge.
“They may give you a notice,” Wendy said, “but it goes to your general counsel, and no one ever really sees it.” The governing terms of a contract can shift without any meaningful review from the procurement or legal teams who originally negotiated it. Knowing that this is happening, as Wendy puts it, is half the battle.
Legal teams are paying attention, although perhaps not to everything. The primary concern she hears from legal is protecting proprietary processes and data from being absorbed into an AI model that could then surface that information more broadly. That is a legitimate concern, but it is a narrow one, and it tends to crowd out the harder conversations about liability allocation, SLA structures for AI outputs, and what happens when an AI-enabled system produces inaccurate results.
The Termination Problem Nobody Is Talking About
One of the most practical issues Wendy raises involves what happens when a contract ends. Most standard termination provisions in technology contracts require providers to return all data and certify that it has been deleted from their systems. But that language was written for a world without machine learning.
“Once they have an AI machine learning tool in their mix of software, how can they guarantee that all that data has been removed from the AI portion and returned and wiped?” Wendy asked. It is not a rhetorical question. It is an open one, and procurement teams negotiating renewals on AI-enabled platforms should be asking it directly.
Where AI Actually Helps
Wendy is not making the case against AI in procurement. Quite the opposite. She points to contract analytics as one of the more genuinely useful applications available, particularly for teams managing large portfolios of agreements with inconsistent language and varying terms.
AI can surface every termination provision across a contract portfolio, flag deviations from standard language, identify renewal timelines, and highlight gaps in regulatory compliance, such as which contracts still need to be updated for the California Consumer Privacy Act (CCPA). The analytical capacities that would require enormous manual effort from a team become something a well-configured tool can do at scale.
According to Wendy, procurement should not be afraid of AI in contracts, but they do have to be clear-eyed about what it changes and deliberate about how those changes get addressed.
Reading the Boring Parts
According to Wendy, true sourcing heroes read their contracts carefully when everyone else is in a hurry, they fight for the terms that matter most to the business, and, ultimately, they can still offer the final word a year later when the implementation team starts complaining about broken or unfulfilled promises. The contract either backs them up, or it does not. Either way, the answer should always be there.
In a moment when AI is changing software contracts faster than most procurement teams can track, that kind of ongoing transparency and governance is exactly the kind of work that protects the business.
For more on this conversation, listen to Wendy's full episode here:
